Noble Audio FoKus Artemis: two-minute review

In Ancient Greece, Artemis was a goddess of many things: vegetation, hunting, wild creatures, childbirth… the list goes on. It’s fair to say that she was considered abundant in talent. And her moniker is not wasted on the Noble Audio FoKus Artemis: a set of wireless headphones that have far-reaching, broad talents of their own.

The FoKus Artemis are, put simply, an exceptionally strong performing set of cans. In fact, they might be the best-sounding pair of wireless headphones I’ve used. They have a triple-driver configuration, with dynamic, balanced armature, and planar drivers, affording them all of the dynamism, precision, and speed required to showcase your favorite tracks in their best light.

Like the best headphones around, the Artemis supply palpable yet regimented bass, beautifully detailed and textured mids, and expressive, articulate treble. But the real shining star is their excellent soundstaging. The Artemis bring forth an immense sense of space and scale, making for a highly immersive listen time after time. With aptX Lossless, LDAC, and USB-C audio, you also get high-resolution, nuanced sound that competes with the greatest headphones on the market.

But something else that sets these headphones apart from the crowd is their granular sound control. With help from Audiodo, Noble has included personalized EQ on the Artemis, which can account for hearing loss and provide optimized audio for all users. There’s also parametric EQ and a number of presets that will suit any listener.

The FoKus Artemis boast quite the suite of features, actually, with an effective companion app, wear detection and multi-point connectivity to give you the everyday convenience that you’d hope for at this money. ANC is here too, with a number of modes, including a transparency option. Noise cancellation here isn’t the best of the best, and the Artemis’ looser fit also means a fair few external sounds creep in now and then. But it’s still relatively effective, and will still help dull the world around you.

Spatial audio modes, though functional, also felt a little underwhelming to me during testing, but the Artemis’ wide soundstage means you get a nice and expansive listen from the headphones anyway. Call quality is strong, though, owing to the headphones’ seven-mic array, and you get up to 55 hours of playback (35 hours with ANC on), as well as a replaceable battery, adding longevity to the headphones.

In my view, the Artemis look great as well. Their metallic detailing, fabric headband, and textured ear cups ensure they exude all of the style and luxury you’d look for in an ultra-premium pair of headphones. There aren’t any touch controls, but the physical buttons work well, and are highly responsive. It also helps that the headphones are extremely well-fitting, and they feel comfortable to wear throughout the day.

But here it comes: the price. The Noble Audio FoKus Artemis cost a whopping $899 / £799 (about AU$1,200). Is the price-tag justified? It really depends on your outlook. They’re a seriously advanced set of headphones, with the triple-driver set-up, personalized EQ and premium build adding a lot in terms of value. But you can get phenomenal sound for less — even the Bowers & Wilkins Px8 S2 are $100 / £170 less than these.

Still, if you want truly stellar audio and ultra-precise sound control, then it really doesn’t get much better than the Noble Audio FoKus Artemis. They’re a wonderful pair of headphones, and if you’ve got the budget, then they’re pretty easy for me to recommend. But for some slightly cheaper yet excellent alternatives, be sure to head down to the ‘also consider’ section.

Noble Audio FoKus Artemis review: price & release date

• Priced at $899 / £799 (about AU$1,200)
• Released in August 2026

The Noble Audio FoKus Artemis started shipping in August 2026, although I got to try them back in June, when the headphones were showcased at High End Vienna 2026. The Artemis are an ultra-premium pair of wireless headphones, and as a result, they carry a very hefty asking price. They’ll typically set you back $899 / £799 (about AU$1,200).

Sure, some rivals, like the Bang & Olufsen Beoplay H100 and Focal Bathys MG, will set you back even further, but I won’t pretend that the Artemis aren’t a huge investment. And you can get exceptional sounding headphones for considerably less — for example, the Bowers & Wilkins Px8 S2 are a good $100 / £170 less than the Artemis at $799 / £629 / AU$1,299.

Noble Audio FoKus Artemis review: specs

Noble Audio FoKus Artemis review: features

• Personal EQ can account for hearing loss
• Strong battery life of up to 55 hours
• ANC is acceptable, rather than great

Noble Audio has put a ton of thought and effort into the feature-set of the FoKus Artemis, and this results in a great user experience.

Where to start? Well, I think the most interesting function is the personalized EQ, which offers surprising depth. Noble Audio partnered with Swedish firm Audiodo to implement a hearing test on the Artemis, meaning you can unlock a sound that suits you — it even accounts for hearing loss. Using Noble’s app, you’ll be asked to listen to a set of prompts, which test your hearing across the frequency range in each ear, and you’ll get a custom EQ that provides bespoke, optimized audio.

This works really well, and can be toggled on or off. There are also a number of presets and a parametric equalizer if you want more granular control. Few headphones offer this attention to detail, and even though I loved the default tuning on the Artemis, they really can sound however you want them to.

There are also some spatial audio modes, with a subtle and immersive option. I didn’t love these so much — I found that tracks sounded less clear and concise, and given that the Artemis whip up an impressive sense of space anyway, I kept this off.

Let’s talk about one of the most important features next, ANC. The Artemis have a seven-microphone system with multiple noise-cancelling modes, including low, mid, and strong ANC and a transparency mode.

So, how good is noise-cancellation with the Artemis? Well, it’s decent. I used them across a huge range of environments: flights, trains, when out on walks, in the office, and at home. When using the strongest mode, I found that they eliminated sounds like typing and light chatter nicely, but more intensive stuff, like a plane’s engine running and passing-by traffic, tended to seep in quite easily. In part, this comes down to the more relaxed fit and clamping on the Artemis, but I did expect a bit more in terms of noise-crushing from these cans, given their price.

Something the mics handle very well, though, is calls. Whenever I made a video or voice call, people on the other end of the line reported that my speech was clear, and never tinny or obscured. This was also the case when I made some recordings and listened back.

The Artemis also impressed me a fair bit in terms of battery life. They offer up to 55 hours of playback with ANC off, and 35 hours. That’s slightly above the standard 30 hours that a lot of premium models offer, and means you’ll be able to listen for days on-end without having to charge the Artemis at all. The battery is also user-replaceable, which is becoming more common with EU legislation making this a requirement for portable devices.

Controlling the Artemis is also pretty straightforward. They feature traditional button controls, favored by a lot of audiophiles for their reliability. Personally, I’m a fan of touch controls (when done right), but this is omitted on the Artemis. Something that does feature, though, is wear detection and this is highly responsive and accurate — it can be toggled off if you’d prefer. Finally, the app is incredibly easy to use and well-laid out. I’ve seen some reports of iOS users having issues, but on my Android phone, the companion software worked perfectly.

• Features score: 4/5

Noble Audio FoKus Artemis review: sound quality

• Triple driver configurations delivers exceptional balance across the frequency range
• Wide soundstage creates a hypnotic listening experience
• Top-tier codec support and USB-C audio passthrough

It may not surprise you to find out that the Noble Audio FoKus Artemis are a truly spectacular-sounding pair of wireless headphones. In fact, they’re up there with some of the best performers I’ve tested here at TechRadar.

This model stands out from the crowd due to one specific design choice: its triple driver configuration. Dynamic drivers are great for handling the bass and mids, but balanced armature and planar drivers enable the Artemis to deliver enhanced precision, speed, and fidelity. But what’s the actual listening experience like?

Well, the most striking thing about the Artemis for me is their ability to conjure up a tremendously spacious, wide sound. For closed back, wireless headphones, the scale on offer is striking, and I was always treated to an almost hypnotic experience, no matter what genre I threw on.

In What We Got by Midas Field and Marjin Jensen, pumping bass hit with great impact and dynamism, without ever shrouding vocals in the mid-range or percussion up high. The speed and responsiveness of the low-end is excellent, and extends nice and low, meaning that even the darkest sounds in the frequency range sound full-bodied and controlled. Drums in the treble range were also precise and expressive without sounding overly bright, helping the Artemis to perfectly capture the energy of the track.

Switching over to Miracles by Change, and the Artemis continue to impress. Electric guitars were perfectly weighted in the mix, and sounded beautifully textured. They were also expertly separated from emotive vocals, which were in turn given plenty of room to express themselves. Subtle percussive elements were also perfectly audible, and the wide soundstage opened up a highly immersive listening experience.

As my phone supports both LDAC and aptX, I was able to get a highly detailed listen from the FoKus Artemis. USB-C audio passthrough also opens up genuine lossless playback, and with this, the Artemis offer up a phenomenally revealing, layered, and nuanced sound that helps to justify their high asking price.

• Sound quality score: 5/5

Noble Audio FoKus Artemis review: design

• Stunning design with fabric headband and metallic detailing
• Very comfortable in use, though looser fit has its downsides
• IP52 dust and waterproof rating

If I could use one word to describe the look of the Noble Audio FoKus Artemis, it would be ‘striking’. The silver metallic detailing on the earcups is bold and eye-catching, and truly exudes a sense of luxury. The fabric headband is also built to a great standard, and the plush earcups appear textured and stylish.

There are a few plastic elements on the Artemis, and you could argue that some parts — like the buttons and carry case — don’t exactly look the most premium. But overall I think that the Artemis are a great-looking pair of headphones. They’re also relatively flat and low-profile, which I really like — I’ve never been a fan of the bulkier look of headphones like the Sony WH-1000XM6.

On top of this, the Artemis are amazingly comfortable in-use. I happily kept them on for hours upon hours when making a long train journey, and never found them to be uncomfortable or overbearing. On my (relatively large) head, I’d say that the Artemis fall on the looser side, fit-wise. While this ensures great comfort, it could sometimes mean that external noises crept in a bit too easily — combined with the fairly middling ANC, this means that the Artemis fail to challenge the gods of the ANC domain, such as Bose and Sony.

Something that’s good to see, though, is that the Artemis have an IP rating. Sure, it’s nothing ground-breaking at IP52 — meaning they’re protected against dust ingress and water sprays — but very few over-ears actually display a rating, so kudos to Noble Audio for this.

• Design score: 4.5/5

Noble Audio FoKus Artemis review: value

• Undoubtedly expensive
• But the exemplary sound quality helps to justify this
• Still, exceptional performers are available for less

Look, the Noble Audio FoKus Artemis are a very expensive pair of wireless headphones. At $899 / £799 (about AU$1,200), they sit right up there with models like the Focal Bathys MG, and even exceed the excellent Bowers & Wilkins Px8 S2 in cost.

But are they really worth it? It honestly depends. If you want some of the best-sounding wireless cans on the market, then you’ll get just that with the Artemis. I’m not sure I’ve experienced any wireless headphones that outperform these. Their spacious, nuanced sound is a delight, and has made me never want to take them off. They also look great and offer a real sense of luxury, while features like the personalized EQ are rarely seen elsewhere.

Still, I can’t help but feel that their ANC could be better for the price, and you can still get incredible sound for a fair bit less with a luxury design from the Px8 S2. As a result, I’d consider the Artemis to be good value if you’re all in on sound, but perhaps consider the options in the ‘also consider’ section if you want to spend a little less.

• Value score: 3.5/5

Should I buy the Noble Audio FoKus Artemis?

Buy them if…

Don’t buy them if…

Noble Audio FoKus Artemis review: also consider

How I tested the Noble Audio FoKus Artemis

• Tested for one month
• Used on flights, trains, on walks, and in the office
• Mainly streamed tunes over Tidal

I spent one month testing the Noble Audio FoKus Artemis, using them as my primary set of headphones. I used them on flights, trains, and buses, but also when out on walks, working at the office, or relaxing at home. Using them in such a range of environments helped me to fully understand the effectiveness of their ANC capabilities.

When listening to music, I typically streamed tunes over Tidal on my Xiaomi 17 phone, which supports aptX and LDAC codecs, but I also dipped into Spotify from time to time. As always, I sifted through the TechRadar reference playlist when testing, but I also listened to countless hours of music from my personal library.

More generally, I’ve spent years testing audio gear here at TechRadar, including everything from wired IEMs like the Beyerdynamic DT 30 IE to party speakers such as the Marshall Bromley 750. I’ve also tested a whole host of premium headphones, including the Sony 1000X The Collexion and Bang & Olufsen H100, so I know exactly what a pair of headphones needs to stand out in a highly competitive market.

• Read TechRadar’s reviews guarantee
• First reviewed: October 2026

• OVHcloud and CGI are teaming with the European Space Agency to build its ESA Earth Observation Information System
• Christened "Digital EO," the project will amass an incredible 500PB of Earth observation data by 2035
• The decentralized cloud-based project will provide data and compute processing for scientists, researchers, businesses, and citizens across Europe and the UK

A huge new project to record Earth Observation data has been launched by the European Space Agency (ESA), with OVHcloud providing cloud storage and processing and project management courtesy of IT and business management firm CGI (via its France Défense et Spatial subsidiary).

The ESA Earth Observation Information System, or Digital EO, aims to collect data about our planet from satellites and other observation hardware for use in studies and research.

With initial deployment in Italy, France, and Germany, Digital EO will be able to extend to other European countries, as well as Canada. Despite leaving the European Union, the UK remains a member of the European Space Agency and as such has access to Digital EO.

Decentralized, scalable, resilient

Determining that the growth in data collection from satellites and other devices and hardware requires better management and wider access, the ESA has launched the Digital EO project as a decentralized system that relies on OVHcloud’s expertise.

“Digital EO must allow us to take a new step forward in the way Earth Observation data is managed and used,” said Nicolaus Hanowski, the ESA’s Head of EO Mission Management and Ground Segment Department.

“With CGI and OVHcloud, we will build and operate a European innovative and sovereign system capable of supporting this ambition and of providing lasting support to the ESA and European Commission programs like Copernicus and DestinE.”

OVHcloud will also support the evolution of the ESA Earth Observation Information System, potentially beyond the initial 30 month project, aiming to ensure performance and meet the needs of Digital EO’s users.

Digital EO backbone

The ESA, which is overseen by the European Commission, has a long history of relying on companies and technology across the continent, rather than buying in solutions from the USA, China, South Korea, or Japan.

French cloud technology firm OVHcloud is the ideal partner for the Digital EO project (OVHcloud state it is worth “multi-million” Euros). Its CEO and Founder, Octave Klaba, noted that, “Europe absolutely must have control over the infrastructure that runs its most strategic programs. With Digital EO, we are putting our sovereign cloud expertise at ESA’s service so that it can build innovative solutions to process, store and make the most of ever-growing volumes of Earth Observation data for the benefit of research, industry and operational user communities.”

OVHcloud aims to deliver “performance, resilience and security” says Klaba, adding that the ambition is “to build a robust, sovereign European Digital EO able to evolve over time, in step with ESA’s needs and ambitions.”

• The FCC has granted Starlink Mobile the right to operate internationally
• This opens the door both to a roaming service for US customers and a Starlink Mobile service in other countries
• The company already has partnerships abroad, and previous reports suggested Starlink Mobile was planning to launch a rival to conventional carriers

Starlink Mobile is barely off the ground in the US yet, with its service primarily offered through the dead zone-filling T-Satellite service in partnership with T-Mobile. But we know the brand has bigger mobile ambitions — and it now seems they won’t be limited to the US.

Following reports that SpaceX was planning to build a Starlink-branded mobile network in the US, the company — as spotted by NotebookCheck — has now received approval from the FCC (Federal Communications Commission) to offer international telecommunications services.

This, at a minimum, could allow Starlink Mobile to offer those dead zone-filling capabilities to US customers when they’re roaming abroad, but it also opens the door to Starlink Mobile services being offered directly to residents of other countries, rather than them only being available through partnerships with other networks.

And we know Starlink plans to level up its mobile service, with 5G-level capabilities and download speeds of up to 150Mbps, all beamed from satellites in space.

Those capabilities, coupled with its existing ability to deliver coverage where conventional mobile networks struggle, could make it either a major boon to carriers it partners with or a significant rival if SpaceX sets up a standalone alternative to the likes of T-Mobile and Verizon.

Starlink mobile services are already rolling out in numerous countries

Any rival service is probably still a way off, especially as the company might need to build some more traditional network infrastructure to compete with conventional carriers in densely populated areas.

But SpaceX is already starting to partner with carriers globally. For example, it recently announced a partnership with Japanese telecommunications company KDDI to offer “Starlink Mobile V2,” which will apparently offer “100x the data density, and will enable native calls, high-speed apps, streaming and more.”

And with this FCC green light, we might one day see Starlink Mobile services offered directly to customers both in the US and abroad, rather than through partnerships alone.

• IBM and NASA launch open source AI model to help further lunar research
• Researchers will be able to better analyze petabytes of Moon data from last five decades
• IBM and NASA also release dataset for public usage and analysis

IBM has launched a new open source AI model it hopes will help spur on NASA researchers in their push to get humanity back to the Moon.

The new NASA-IBM Lunar Foundation model, available on Hugging Face, will allow researchers to analyze decades of lunar observation data, and identify geological features that are critical to understand for NASA as it looks to build a sustained human presence on the Moon.

The model has been trained by IBM and NASA researchers on a huge, multimodal NASA dataset, which will also be released alongside the model, providing wider access to the latest advanced AI systems in a bid to push on wider progress in lunar exploration.

To the Moon (and beyond)

At its most obvious level, the model will allow a much easier way for researchers to study petabytes of data gathered on the Moon's surface for potentially hazardous locations such as ice deposits or craters.

Currently, scientists often rely on manual analysis or low-resolution, task-specific AI models, which can be not only computationally demanding, but also often lack the accuracy needed for detailed geographic analysis.

The new release will now mean that instead of needing to build a new AI model for every potential issue, scientists can now adapt a single foundation model to investigate a range of lunar geologic features.

The model has already proved useful, identifying craters and volcanic features far more accurately (see below) and significantly reducing errors in locating potential ice deposits.

IBM, which has worked with NASA for over five decades, including on the Apollo missions, believes the model could help future astronauts navigate safely and even find essential resources, as well as helping scientists better understand the Moon's geological history.

“Uncovering the mysteries of the Moon requires an ability to learn from an extraordinary volume of scientific data,” said Juan Bernabe-Moreno, Director of IBM Research Europe, UK and Ireland.

“The NASA-IBM Lunar Foundation Model gives scientists a foundation to explore the Moon at scale, connecting observations across instruments, revealing patterns that are difficult to see in isolation, and providing an open platform the global research community can build on.”

The release of the dataset will mark the first time a unified, publicly-available cache has been made available and ready for machine learning. It brings together over 30 spatially aligned layers from nine instruments across four missions, including tens of thousands of images and maps showing unique geophysical properties of the lunar surface from NASA’s Lunar Reconnaissance Orbiter (LRO) and NASA’s GRAIL mission.

Identifying lunar ice deposits could be particularly vital, as the presence of both water and oxygen will be crucial to establishing a human base on the Moon, and even creating rocket fuel for future Mars missions.

Scanning the Moon's volcanic features, known as Iregular Mare Patches, can allow scientists to better understand the Moon's volcanic history and thermal evolution, as well as helping identify potential sites for landing and other surface operations.

Finally, studying the Moon's craters can offer a wealth of information on its history, including the age of different terrains, their geology, and even the chemical composition of the early lunar interior - as well as again helping to identify safe landing sites without hazards such as steep slopes and boulders.

• The British government is using its UK Space Strategy to drive growth and reindustrialise the country
• £7.8 billion fund will back the project until 2030
• Funding will boost European rocket programs, space science and exploration, rail connectivity, astronomy and space science research, and space weather forecasting

The British government has announced a multi-faceted UK Space Strategy which seeks to “reindustrialise the country” with a £7.8 billion investment between now and 2030. Uniting a number of scientific and communications programs, the strategy is also part of the UK’s defense plans, which have been under fire in recent months.

While improved rail connectivity and satellite communications for rural communities will make a difference to everyday life, the program is largely focused on various scientific research projects, from pioneering orbital assembly and manufacturing to exploration and extraterrestrial weather.

Replacing 2021’s National Space Strategy, the UK Space Strategy is backed by two key ministers, and supported by various bodies with an interest in orbital research, science, and defene.

Research and ISAM

The £7.8 billion fund is intended for a variety of purposes, unified under the UK Space Strategy. These include a £149 million contribution to the European Space Agency’s (ESA) Vigil mission, and £880 million to improve space-based intelligence, surveillance, and reconnaissance, technologies that enable the UK to monitor military activity and protect assets in orbit.

A further £2.8 billion aims to strengthen the connectivity in Low Earth Orbit and SKYNET defense communications programs.

The UK government expects the UK’s space sector to grow beyond its current £18.6 billion value, with further funds allocated to rail connectivity (£57 million), developing SaxaVord Spaceport in Shetland (£30 million) and space science and exploration missions (£163 million). Projects like the Rosalind Franklin Mars Rover are cited in the announcement, with an indication that similar projects can be part of the reindustrialization program.

Representatives from UKspace (a space industry trade association), ADS Group (a similar body for aerospace, defense, security and space businesses), the Space Academic Network, and the Met Office have thrown their weight behind the UK Space Strategy. In addition, there is support from Joanne Wheeler, MBE, director of the Earth & Space Sustainability Initiative.

A new frontier

A pair of prominent politicians back the UK Space Strategy, with Business, Innovation, Science and Trade Secretary, Jonathan Reynolds, stating "As space becomes a new frontier of economic and military competition, this plan will help keep Britain secure by strengthening our ability to launch satellites, detect threats and protect the services people rely on every day.”

Meanwhile, Defence Secretary Wes Streeting, said: “As threats in orbit grow more sophisticated, so must our ability to detect and respond to them, ensuring we protect the interests of the UK and allies while fueling economic growth along the way.”

Streeting also referred to the Defence Investment Plan, a commitment to increase defense spending to 2.7% of GDP by 2029, as part of the UK’s move towards a NATO commitment of 5% GDP by 2035.

• Russia has launched 32 Rassvet satellites, though coverage remains limited
• Rassvet remains vastly smaller than Starlink’s expanding satellite constellation today
• Russia plans 156 satellite launches in 2026, 292 in 2027, 318 in 2028, and about 900 in 2035

Russia has launched 32 Rassvet communications satellites as Moscow works toward building a domestic network capable of competing with Starlink.

The current fleet, however, remains too small to provide continuous service, while reported orbital problems have raised further doubts about progress.

Russia still plans commercial operations from 2027, although reaching that stage depends on expanding the constellation and overcoming serious technical limitations.

Thirty-two satellites leave Russia far from continuous coverage

Ukraine's Defense Intelligence chief, Oleh Ivashchenko, said the existing constellation cannot yet deliver reliable communications across the areas Russia intends to serve.

“We know that today the Russians are trying to put satellites into orbit. But so far, there are not enough of them, and their effectiveness is low,” said Oleh Ivashchenko, chief of Ukraine’s Defense Intelligence.

Ivashchenko did not disclose the spacecraft’s orbital paths, operating frequency, or the Ukrainian regions where the system currently provides coverage.

However, he maintains that Rassvet remains far smaller than Starlink, which has more than 12,000 satellites in orbit and continues adding around 200 spacecraft each month.

The much smaller constellation means Rassvet cannot yet provide the same level of continuous coverage as Starlink.

Moscow has committed 102.8 billion rubles, or about $1.3 billion, to Rassvet as part of efforts to establish domestic broadband connectivity.

Russia plans to launch 156 satellites in 2026, followed by another 292 spacecraft by 2027, when authorities expect commercial operations to begin.

The programme ultimately aims for 318 satellites by 2028, while later plans envisage expanding the network from 250 spacecraft toward 900 by 2035.

Those figures reveal the scale of work remaining, since the existing fleet represents only a fraction of Russia's planned communications infrastructure.

Technical problems could make expansion considerably harder

Reports indicate that some Rassvet spacecraft faced difficulty gaining altitude, with certain satellites drifting downward after reaching orbit.

Defense Express reported on August 25 2026 that one Rassvet-3 communications satellite was lost, while two others faced possible atmospheric reentry.

Such incidents could complicate Russia's schedule because each failed spacecraft reduces the already limited constellation and creates additional pressure on future launches.

Building the planned network would require at least 18 Soyuz-2.1b missions, despite Russia currently carrying out only six to eight launches annually.

The launch requirement therefore creates another obstacle, especially if replacement spacecraft become necessary after further technical failures during orbital operations.

Rassvet could eventually provide communications links for unmanned aircraft and military units operating beyond the practical limits of conventional radio systems.

That possibility gained greater attention after Ukraine and SpaceX restricted unauthorized Starlink terminals reportedly used by Russian forces during February 2026.

“We are working on this. Russia’s ‘Rassvet svg’ must turn into a ‘zakat svg,’” said Oleh Ivashchenko, chief of Ukraine’s Defense Intelligence.

Russia continues developing Rassvet despite these setbacks, but the current constellation remains unable to provide continuous connectivity expected from a fully operating network.

Via United24media

• Some Rassvet satellites remain dangerously low after failing to raise their orbits
• One Rassvet spacecraft has already fallen from orbit and reentered Earth's atmosphere
• Two newer satellites have shown no observed orbital-raising maneuvers since July 2026

Moscow's push to field its own broadband satellite network is running into some technical and industrial issues, new reports have revealed.

Bureau 1440, the company developing the Rassvet-3 system, has sent up two batches of spacecraft in 2026, one in March and another in July.

According to reports, some of these systems did not climb smoothly toward their intended altitude; instead, they stalled, drifted, or fell back toward Earth.

A satellite already lost, others trending downward

One craft from the first group, known as Rassvet-3 No. 4, never managed to raise its orbit after separating from its rocket.

That satellite, catalogued as NORAD 68363, eventually reentered the atmosphere on June 6, marking an outright loss for the young constellation.

Two more spacecraft from the second launch group, numbers 17 and 23, appear to be following a similar pattern of inactivity.

Tracking data shows neither satellite has performed any observed maneuver to climb higher since it separated from its carrier rocket on July 19.

Both currently sit at an average altitude near 290 kilometers, with their lowest points measured at roughly 277 kilometers above the surface.

Without engine activity to counter this decline, the pair could eventually drop low enough that reentry becomes essentially unavoidable.

However, there are some positives, as some of the satellites have successfully begun climbing, though none has reached the network's target altitude of about 800 kilometers.

Spacecraft from the earliest operational batch launched just over five months ago, and at the moment, they hover between 500 and 550 kilometers.

A few more units from the group are not yet at that intermediate range, which means that the company is yet to hit its long-term altitude plan.

The Bureau 1440 has now paused raising some Rassvet satellites higher toward their planned 800-kilometer orbit, and the company hasn't revealed the reason for the pause.

Ground infrastructure adds to the program's troubles

Russia intends for the Rassvet network to rival SpaceX's Starlink by offering fast internet through hundreds of orbiting satellites.

Officials have said commercial service should begin in 2027, with plans calling for 292 satellites to be functioning in orbit by year's end.

Meeting that timeline would require Bureau 1440 to conduct a rapid sequence of additional launches over the coming months.

A separate complication has now emerged on the ground, far from the satellites already circling the planet.

Ukrainian forces allegedly used FP-5 Flamingo cruise missiles to strike Samara's Progress Rocket and Space Center.

That facility builds the Soyuz-2.1b rockets responsible for carrying every Rassvet-3 satellite launched so far this year.

The plant also supports numerous other Russian launches, including several military, reconnaissance, and communications missions unrelated to broadband service.

The rocket production capacity is one of the central bottlenecks limiting the constellation's overall deployment pace.

Should the damage at Progress persist, sustaining the launch tempo needed to reach nearly 300 satellites may prove genuinely difficult.

Via United24media

• Trump wants US to up aims to launch over 1,000 spacecraft every year by 2030
• Only 178 launches happened in 2025, mostly carried out by SpaceX
• Federal agencies must find new land for rocket launch sites

A new presidential memorandum signed by Donald Trump sets an ambitious annual benchmark for American commercial space activity within the next four years.

The new National Space Transportation Policy wants government agencies to work together so that the US hits over 1,000 spacecraft launches and returns per year by 2030.

In 2025, there were only 178 of these missions, though that number is already ten times higher than it was back in 2013, and SpaceX was behind most of them.

Land, permits, and spectrum take priority

The order replaces a national space policy that had been in place since 2013, when the previous framework was first adopted.

Within 90 days, several departments must scout available federal land that could host new sites for launching and landing spacecraft.

Approval processes for these facilities should move faster too, with environmental reviews and permits getting streamlined under the new rules.

Access to wireless spectrum, air traffic coordination and a stronger launch workforce are also named as priorities the government must address going forward.

Agencies have further been told to step back from any activity that might discourage or undercut private companies competing in the same space.

Export policies covering space technology will also be revisited periodically, aimed at protecting American intellectual property while expanding sales abroad.

The policy also calls for regular federal reviews of launch range schedules, intended to make better use of limited windows nationwide.

"This policy meets this moment of explosive growth and positions America to lead the next era of space," said Michael Kratsios, the White House's science and technology policy director.

Beyond Earth orbit, the memorandum pushes NASA toward landing American astronauts on the moon again before the end of 2028.

Early elements of a lunar base should follow two years later, alongside continued study of robotic and eventually human trips to Mars.

A dominant player and lingering doubts

Elon Musk's SpaceX remains the industry's clear leader, having carried out 170 missions last year that placed about 2,500 satellites into orbit.

Officials at the Federal Aviation Administration said back in May that the firm hopes to reach 10,000 flights annually within the next five years.

SpaceX has separately floated plans for roughly 1 million satellites designed to capture solar energy and feed power to AI data centers.

In addition to SpaceX, the U.S. will entertain requests from foreign companies that want to launch from American soil.

These firms will be screened individually, weighing security risk, investment value and site availability.

NASA, on the other hand, has been asked to build a lunar logistics network that private companies could eventually use to reach the moon commercially.

Lawmakers, meanwhile, have kept urging NASA to stay focused on Artemis, its long-running lunar program that has already consumed billions in funding.

That pressure suggests the administration's push toward 1,000 annual launches could still meet real budgetary and political resistance in Congress.

• NASA’s ground control software has a critical vulnerability that could allow third-party access to spacecraft
• The flaw has been found in a browser-based variant of NASA’s AMMOS Instrument Toolkit
• NASA has not publicly responded to the flaw’s disclosure

NASA’s open source ground control software has a critical vulnerability that could enable an unauthenticated attacker to gain access and take control of spacecraft. The AMMOS Instrument Toolkit’s browser-based interface is the source of the vulnerability, which has since been resolved.

The AIT-GUI (AMMOS Instrument Toolkit Graphical User Interface) tool up to version 2.5.1 has been identified as vulnerable, but the fault has been fixed in v2.5.2, according to researcher Yuval Elbar.

If the vulnerability had been found by a third party, it would have given access to issue commands to spacecraft, instruments, and execute server-side scripts. In addition, command sequences could potentially have been run by an attacker, leaving craft almost wholly beyond NASA’s control.

Weak API

Elbar, who works with the Cycode Agentic Development Security Platform, disclosed the vulnerability on August 18, 2026. The problem appears to start with AIT-GUI running as a web server with all network interfaces open, rather than the host’s setting. Incredibly, the API also has no authentication, or authorization protection. In addition, the CSRF (cross-site request forgery) protection on changing endpoints is also absent on affected versions of AIT-GUI.

Attackers can exploit basic access-control security failings in AIT-GUI, exposing the /cmd and /script/run and /seq prompts while also setting up filesystem paths. This enables files (potentially malware, or custom-built scripts) to be passed directly to NASA craft via a vulnerable AIT-GUI browser session, potentially escalating to full control.

To emphasize the scale of the weakness, Elbar introduced the disclosure with “A web GUI used to drive spacecraft and instrument commanding shipped a server that listens on every network interface, asks nobody for a password, and can be steered by any web page an operator happens to open.”

External access

As if this wasn’t concerning enough, the attacker doesn’t need to be on the same network. Direct access via an exposed port, or directing an operator to a web page hiding malicious code, or even simply a web page under live control by an attacker, can afford access to a third party.

Elbar adds “operational and ground-system software inherits the same web weaknesses as everything else, but with a far higher cost of failure. Auth, CSRF defense, and input confinement are not optional extras on a panel that commands hardware.”

Cycode advises administrators of AIT systems to upgrade AIT-GUI to v2.5.2 and run checks on the console port. They should also review command history.

In a few short hours, parts of the Earth will experience a rare and wonderful sight: a full solar eclipse. The moon will squeeze between us and the sun so that Spain, in particular, will, for a few minutes, fall into almost total darkness. What won't happen is that humans and other stationary objects not bolted to the ground will float up into the air. Gravity won't disappear.

But that's not the full story.

Listen, I ended up diving down a rabbit hole this morning, a journey triggered by one Weather Channel post on Threads:

"You all know we’re not actually going to lose gravity for 7 seconds today right? It’s important to me that you know that."

Turns out that there is a wild conspiracy theory going around that the eclipse will somehow trigger a gravity disruption, basically canceling it for 7 seconds.

The rumor was apparently set off, in part, by a TikTok claiming NASA had something called "Project Anchor" addressing the phenomenon.

NASA quickly quashed the rumor, telling Snopes:

"The Earth will not lose gravity on Aug. 12, 2026. Earth's gravity, or total gravitational force, is determined by its mass. The only way for the Earth to lose gravity would be for the Earth system, the combined mass of its core, mantle, crust, ocean, terrestrial water, and atmosphere, to lose mass. A total solar eclipse has no unusual impact on Earth's gravity. The gravitational attraction of the Sun and Moon on the Earth, which doesn't impact Earth's total gravity, but does impact tidal forces, is well understood and is predictable decades in advance."

Let's talk about the pendulum anomoly

So that settles it, and I think we can all rest easy that our feet will be firmly planted on the ground today, tomorrow, and beyond.

On the other hand, the concept of a solar eclipse having some nominal effects on gravity or the movement of objects affected by gravity is not totally out of the question, and the existence of such anomalies may explain why some people have extrapolated a wildly exaggerated solar eclipse gravitational impact.

There have been various reports over the decades of eclipses impacting, for instance, the swing of pendulums (according to New Scientist, the anomaly was first spotted in 1954). In fact, there have been some scholarly studies going back at least 25 years on the phenomenon.

In 1999, Chris P. Duif, then a physicist with the Delft University of Technology, The Netherlands, produced a scholarly paper examining, among other things, why certain pendulums reportedly altered their swings during a pair of solar eclipses.

Duif's paper notes that there were "weak conventional explanations" but added that the anomalies could still be "assigned to mundane causes such as temperature changes. "

In the end, though, Duif's research did not confirm the impact; equally, he seemed unable to rule it out. As he told New Scientist in 2009, "I'm not really convinced the anomaly exists, but it would be revolutionary if it turned out to be true."

In the paper, Duif appeared to hope for more study, writing, "If an anomalous eclipse effect does exist, a clever combination of experiments and input from theoretical models will be necessary to reveal its existence." One has to wonder if AI could assist in this pursuit. After all, it is quite good at finding meaning in the vast amounts of data typically found in such models.

The point is that there's this persistent idea that solar eclipses can affect gravity and that, it seems, has perpetuated and ballooned, with the help of conspiracy theorists, into this wild idea of a total gravitational pause. But, as The Weather Channel and NASA will quickly remind you, that's simply not the case.

• Chinese equipment fills critical roles, though it represents only a small share overall
• Russia plans to quadruple annual Angara rocket production within the next decade
• Internal documents expose extensive factory upgrades supporting future rocket manufacturing

Russia is pushing to significantly grow production of its Angara space rockets, to quadruple annual output within the next decade.

Internal planning documents obtained by researchers show the expansion depends heavily on imported machine tools, and most of them are Chinese-made.

That dependence stretches across shop floors that Russia will need to modernize before it can build four rockets a year instead of one.

Counting the Chinese machines

Of the 541 pieces of equipment listed in the procurement plan, 477 are Russian-made, and only 57 come from Chinese suppliers.

Two more machines came from Bulgaria, leaving the Chinese share of the total list at roughly 10.5%, a modest slice on paper.

That number changes sharply once each machine is weighed by its technological importance rather than simply counted one by one.

Measured that way, Chinese suppliers account for 28.2% of the entire project's technological core, based on the internal planning documents.

In the metalworking shop alone, over 47% of the equipment needed there comes from Chinese manufacturers instead of domestic ones.

The finance tells a similar story: over 84% of the entire 3.62-billion-rouble (~$44.71 million) equipment budget still flows into that same technological core.

In total, Khrunichev intends to put close to 900 million roubles (~ $11.29 million), toward equipment made in China.

Roughly 10 manufacturers account for most of that spending, and several of them also sell machinery to buyers in Europe and America.

Where the machines really come from

Some of those companies fit a wider trend that has taken hold since the conflict between Russia and Ukraine started in 2022.

Heavy forging and hydraulic press equipment fall under European Union export restrictions whenever the intended buyer is based inside Russia.

Washington has also placed sanctions on Russian makers of heavy hydraulic presses, one of them a firm called Pressmash, back in 2023.

Yet under that very same customs code, the EU's imports of Chinese-made machine tools still totalled 207 million euros (~$238 million) in 2025.

Those imports rose another 17.6% in the opening six months of 2026, measured against the equivalent stretch the previous year.

In effect, the same manufacturers profit from Western markets that back Ukraine while also helping modernize the military hardware Russia sends into orbit.

Sanctioning any single Chinese supplier is unlikely to stop this flow of equipment, since near-identical machines are made by several competitors.

Whether losing access to European and American customers would ever outweigh the profits of supplying Russia remains an open question.

The Angara rockets built with the Chinese equipment carry the Russian Ministry of Defense satellites that American experts claim could threaten assets in orbit and on the ground.

However, Russian experts described the spacecraft as operating "in the interests of the Russian Ministry of Defense."

That difference between Moscow's official position and Washington's assessment has persisted since the conflict in Ukraine began.

Via United24media

• Recycled Mars rover hardware may still cost NASA more than expected
• PROMISE faces mounting questions over unexpectedly high development and integration costs
• Existing rover hardware cannot eliminate costly engineering and certification requirements

NASA's long-term lunar exploration program includes robotic missions requiring specialized vehicles capable of carrying scientific instruments across the Moon.

One proposed mission involves sending the Polar Rover for Observation, Mapping, and In-Situ Exploration (PROMISE), a robotic Moon rover, to conduct scientific investigations on the lunar surface.

According to NASA, the proposal to build the rover from recycled Mars rover engineering hardware could become too expensive after estimates reached between $723 million and $1.33 billion.

Reusing existing rover hardware does not eliminate major mission costs

Instead of building the Moon rover from scratch, NASA plans to combine engineering models from the Curiosity and Perseverance rover programs.

Although the concept relies on existing hardware, NASA says the expected costs remain far beyond what the agency considers acceptable for such a mission.

The cost estimates were prepared by Casey Dreier, the Planetary Society's chief of space policy, after reviewing NASA's proposal for the planned lunar rover.

According to Dreier, the estimated total includes launch and landing costs ranging from $234 million to $320 million together with engineering expenses.

NASA Administrator Jared Isaacman did not challenge every figure within the estimate, but argued that launch and landing expenses would exist for any lunar mission.

Isaacman believes that it “would be crazy not to take advantage of hardware taxpayers have already invested hundreds of millions of dollars in."

But, "If PROMISE costs even 20 percent of your low-end estimate, we simply will not fly it," he concluded.

The proposal combines MAGGIE and OPTIMISM, engineering models developed to support testing for the Curiosity and Perseverance rover programs before their Mars missions.

Those vehicles were built to help engineers test hardware performance, software functions, and driving activities under controlled conditions before operational spacecraft left Earth.

NASA originally considered the concept because using hardware that already exists appeared capable of reducing development costs.

Recycled rover concept still requires major engineering work

Engineers would still need to certify the hardware for launch because equipment designed for ground testing cannot automatically survive spaceflight and lunar operations.

Additional work would also involve adapting communications equipment and integrating NASA's available plutonium radioisotope power system into the redesigned lunar vehicle.

The engineering models currently support existing activities, meaning removing them from those roles could affect testing work linked to other programs.

Dreier summarized the financial challenge by saying, "Regardless of whether one agrees with every estimate… one thing is clear: PROMISE is far from a 'freebie.'"

The engineering models currently support existing activities, meaning removing them from those roles could affect testing work linked to other programs.

NASA has not released its own detailed cost estimate for the proposed rover; therefore, the project's overall financial viability remains under wraps.

That said, its future will likely depend on whether engineers can reduce costs while overcoming technical challenges involved in converting Earth-based test hardware.

Via The Register

• Starlink Mobile could become a reality in 2027
• It will likely rely on the very new Starlink V3 satellite
• SpaceX's huge Starship will play a crucial role in the effort to compete with major carriers

SpaceX's first earnings call this week as a publicly traded company was full of promises and hyperbole, including the prediction from boaster-in-chief Elon Musk (also SpaceX CEO and Chief Engineer) that Starlink "will deliver a majority of the world's internet, at least in countries where we're allowed to operate, which is the vast majority of countries." As for a time scale, think not the distant future, but fewer than 10 years.

That's some bold talk, but SpaceX President, COO, and someone far less likely to wax bombastic, Gwynne Shotwell, backed up those claims. "As far as revenue, I'm just going to take a very high-level kind of top-down. The big three in the U.S., AT&T, Verizon, and T-Mobile, roughly between them, $600 billion a year. I anticipate us to be able to acquire quite a few of their customers because I think our service will be better." Shotwell told investors.

These promises all hinge, in part, on the development and launch of Starlink's V3 satellites, which will, on a per-satellite basis, drastically improve throughput. Starlink's current V2 satellite has a downlink throughput of roughly 80GBps. V3 can, the company claims, deliver 1 Terabit of downlink connectivity. Uplink jumps from roughly 7 Gbps to between 160 and 200 Gbps.

It's not just the speed; Starlink V3 satellites promise exponentially more beams, which is basically support for more concurrent users. V2 offers just under 200 beams while V3 promises 2048 per satellite.

Still, with just a reported 10,918 Starlink satellites in low Earth orbit (LEO), they will need a way of quickly delivering tens of thousands of satellites into orbit to achieve this lofty goal.

And so the other part of this hyperbolic plan hinges on SpaceX's still-in-testing Starship rocket. Granted, the 408ft tall spacecraft is rapidly transitioning to the operational phase and may have a full launch as soon as this month. It's already lifted some V3 satellites into orbit during one of its earlier orbital test flights. Now, to realize Musk's plan, it must shift to rapid launch deployment and then relaunch and deploy again.

Musk told investors that "Starlink V3 satellite is about an order of magnitude more capable than the Starlink V2 satellite." Then he added some color on the scale of what SpaceX and Starlink are planning: "We expect to launch about an order of magnitude more Starlink V3 satellites. That would mean a roughly two-order-of-magnitude increase in the delivered bandwidth."

Musk believes we'll need all that bandwidth, not just for potential consumer mobile broadband users but for the vast amounts of data and connectivity required by AI and robotics (both here and possibly on a Moon base), two other sectors where Musk has a vested interest.

"With the advent of AI, and humanoid robotics and vehicle robotics and just a massive number of robots, the appetite for bandwidth will be much greater than it has been in the past," he explained during the earnings call.

Starlink Mobile reality check

Starlink, though, is quite a distance away from competing directly with Verizon, T-Mobile, or AT&T in the US. Across them, they have more than 300M connections (or customers). Starlink has a reported 12M global customers.

As for those V3 satellites that will help propel Starlink Mobile into the mobile broadband competitive space, they do not officially launch until next year. In the meantime, Starlink's V2 Satellites will fill the gaps and help provide partners like EchoStar with better, faster, and maybe more reliable satellite connectivity.

We do already have some proof of the efficacy of Starlink WiFi connectivity and throughput. Earlier this year, TechRadar US Managing Editor for News Jake Krol took a flight on a Starlink-equipped United Airlines aircraft where he enjoyed incredible connectivity and speed. Krol was a fan: "I could livestream a show or movie, or even select a 'live channel' via Disney+ on an iPad, watch any video on YouTube, and scroll through TikTok or Instagram Reels without the slowdowns or loading issues typical of non-Starlink-equipped United flights."

Of course, that connectivity goes from the Starlink satellites to a responder on the jet, which then feeds the signal to onboard Wi-Fi and connected devices. In the case of Starlink Mobile, there'll potentially be no middleman, just your phone, a satellite, and instant, fast connectivity.

The benefit of the low-Earth-orbit V3 satellites, when they do finally launch and fill the skies, is that they should allow consumers to connect with them just as they would with towers assigned to their favorite carrier.

There is also one major hurdle that SpaceX and Starlink already cleared when the FCC granted Starlink approval in January 2026 for the company's "Next-gen satellite constellation."

Even with all this in place, the stars must align with Starship being fully operational, all those new satellites launched, and customers lined up by the millions. The last bit is surely the biggest question mark. Even if most of their modern cellphones support these direct-to-satellite connections, how likely are consumers to leave their trusted carrier behind?

Shotwell, for one, is bullish: "I'm quite excited about Starlink Mobile going forward," she told investors, "We will start to fly the satellites next year, and we will start providing service end of next year."

• A spent Falcon 9 upper-stage rocket will hit the Moon near the Einstein Crater on 5 August at 5,400 mph, releasing roughly the energy equivalent of three tons of TNT
• The impact, though unintentional, was predicted since September 2025, giving NASA's LRO and South Korea's Danuri time to plan before-and-after imaging of the new crater
• Its scientific value is a rehearsal rather than a measurement: no seismometers have operated on the moon since 1977, so the event mainly tests impact-localization methods for future experiments

At roughly 06:35 UTC on August 5 2026, a spent Falcon 9 upper stage will stop being space junk and start being a crater.

The object, cataloged as 2025-010D, is projected to strike the lunar surface near the crater Einstein at 2.43 kilometers per second, which converts to about 8,700 kph or 5,400 mph.

While this was not intentionally aimed at the moon, almost a year of advance notice separates it from the usual category of orbital accident, even as NASA's Lunar Reconnaissance Orbiter and South Korea's Danuri are being prepared to take before-and-after images of the area on the moon where it is expected to make its impact.

An increasingly crowded atmosphere

The object itself is the spent upper stage of the Falcon 9 which launched from Kennedy Space Center on January 15, 2025, carrying two commercial landers: Firefly Aerospace's Blue Ghost Mission 1, which went on to make the first fully successful private lunar touchdown that March, and ispace's Hakuto-R Mission 2, which was lost shortly before landing in June.

Independent astronomer Bill Gray, whose Project Pluto software produced the original prediction, puts the stage at roughly 12 meters long, about the height of a five-story building, with a mass near 4,900 kg.

After separation, it stayed in an elongated Earth orbit with a 26-day period, swinging between roughly 220,000 and 510,000 km. Its path and the moon's path cross, but for 18 months the two objects have never met at the crossing point.

Gray expects a single crater around 17 meters across, reasoning by analogy with the double crater left by a Chang'e-5 T1 upper stage in 2022. The observation-planning paper led by Benjamin Fernando points higher, at roughly 27 meters across and 5 meters deep. Either way, the impact should release energy equivalent to about three tons of TNT.

There are limits to what can be measured from it. When NASA deliberately crashed Saturn V third stages and lunar modules into the moon between 1969 and 1972, those impacts were seismic sources fired into an emplaced network of four seismometers. That network was switched off in 1977, and nothing has replaced it yet. So while orbiters will photograph the resulting crater, the impact cannot be recorded seismically at all.

This will be only the second dead rocket known to have hit the moon by accident, after the 2022 Chang'e-5 T1 event, which China has never publicly acknowledged. Gray, for his part, says the episode points to a certain carelessness about how leftover space hardware gets disposed of. Objects in low Earth orbit eventually meet the atmosphere and burn up. The moon has no such cleaning mechanism, so lunar impacts are permanent.

• SpaceXAI just launched a Grok add-in for Excel
• However OpenAI, Anthropic and others have already beaten Grok to it
• The latest Grok 4.5 model is said to be an Excel genius

SpaceXAI (formerly xAI) has launched Grok for Excel as an add-in which looks to put Grok's conversational AI abilities within a side panel to help with formulas, charts, analysis and more.

Although the add-in is technically free, it requires a paid account to use the features, with only SuperGrok, SuperGrok Heavy and Business/Enterprise plans capable of handling tasks within the add-in.

SpaceXAI also refers to the add-in as a 'free' Microsoft Marketplace tool, loosely implying its use is free despite current account limitations.

Grok now gets an Excel add-in, too

"Describe the outcome and Grok handles the syntax – averages, sorts, and fills land as normal edits in the formula bar," the company wrote in a post announcing the new tool.

Grok for Excel supports Excel 2019 or later on Mac as well as M365 versions and web versions, and joins similar Word, PowerPoint and Outlook add-ins. All the Office add-ins promise to interact with documents, understand data and draft new content, but there are no autonomous agentic features such as sending an email in Outlook.

Realistically, SpaceXAI is just the latest company to be acknowledging that individual sectors need role-specific support, with finance emerging as one of the priorities. OpenAI already has ChatGPT for Excel, Anthropic has Claude for Excel, and of course Microsoft has its own Copilot add-in for paid plans.

The news comes days after the company launched its latest Grok 4.5 model. "Grok Build is capable of building complex Excel models that involve research from the web, multi-sheet formula use, and even leaves stickies or notes behind for future reference," the company wrote.

SpaceXAI also praised the model's capabilities across Word and PowerPoint.

• FCC approves testing of satellites designed to reflect sunlight toward Earth
• Startup plans thousands of orbital mirrors capable of extending daylight periods
• Scientists warn artificial sunlight could interfere with astronomical observations worldwide

A California startup has received approval from the Federal Communications Commission (FCC) to test satellites designed to reflect sunlight toward Earth.

Reflect Orbital plans to launch its Eärendil-1 spacecraft with large orbital mirrors capable of extending daylight in selected areas.

The concept promises new applications for energy and emergency services, although scientists have raised concerns about its broader consequences.

Reflect Orbital to begin testing of ambitious sunlight reflection technology

The Eärendil-1 satellite will operate in low Earth orbit with four thin film reflectors measuring about 18 meters, or roughly 60 feet, across.

The reflected sunlight will be directed toward specific locations for short periods, supporting activities such as construction work, search and rescue operations, and solar power generation.

Reflect Orbital says its mirrors could eventually deliver up to 36,000 lux, roughly comparable to genuine daylight conditions found outdoors.

The system could also provide around 100 lux continuously, similar to lighting typically found inside standard indoor workspaces around the clock.

“We’re grateful to the FCC for recognizing the importance of testing novel technologies in space,” Ben Nowack, chief executive of Reflect Orbital, said in a statement.

“This license is the first step toward rigorously testing our technology’s efficacy and the safeguards we have developed.”

The idea has drawn comparisons with the fictional Icarus satellite from the James Bond film Die Another Day, where an orbital reflector was designed to bring sunlight to dark regions.

While the real technology differs significantly from the movie concept, critics argue similar questions about unintended effects remain relevant as deployment plans expand.

Astronomers and environmental groups warn about possible impacts

Astronomers have expressed concerns that reflected sunlight from thousands of satellites could interfere with observations of faint objects across the night sky.

Tony Tyson, chief scientist at the Vera C. Rubin Observatory, described the proposal as “even crazier” than previous satellite constellation challenges affecting astronomical research.

The European Southern Observatory warned that a full deployment could increase sky brightness at its facilities by three to four times, reducing the ability of telescopes to detect distant and faint astronomical sources.

Researchers also argue that artificial nighttime illumination could affect biological systems that depend on natural day and night cycles.

These concerns involve plants, animals and humans that rely upon predictable environmental patterns for normal behaviour.

The FCC stated that scientific and environmental issues fall outside its direct regulatory authority, which focuses mainly on communications services involving radio, television, satellite and related systems.

The agency approved testing because it believes allowing new space technologies supports innovation and economic development.

Critics may instead seek reviews from agencies such as NASA and the Environmental Protection Agency, which handle broader questions involving space activity and environmental effects.

Whether orbital mirrors become a useful technology or create unexpected challenges will depend on future testing, oversight, and independent scientific assessments.

Via Tom's Hardware

• 32 “Meadowlands” Counter Communication Systems have been commissioned by U.S. Space Force Combat Forces Command
• Rather than laser, microwave or plasma, the weapon is an electromagnetic beam
• Enemy satellites can be disabled and signals jammed using the weapon

A fleet of 32 “Meadowlands” Counter Communication Systems (CCS) is being assembled by US Space Force Combat Forces Command, which has just taken receipt of the first mobile electromagnetic warfare system.

While recent projectile-free weapons systems have employed lasers, microwaves, and ionized plasma gas, a Meadowlands unit – portable enough to be transported by air – uses targeted electromagnetic fields.

Electromagnetic warfare has proved successful in recent years, with Operation Midnight Hammer – the June 2025 US airstrikes on Iran – using the technology to create a “silence zone” over the country to facilitate the attack. By disrupting communications with the electromagnetic weapon on that occasion, targets were struck with minimal countermeasures deployed.

What does an Electromagnetic Weapon (EW) do?

Systems like Meadowlands CCS have the capability to interfere with enemy satellites and jam their signals (affecting uplink and downlink transfers). Described as being able to “detect, deny, disrupt, and degrade adversary capabilities” the technology can also change or disrupt target data, and generate confusion.

US Space Force Col. Angelo Fernandez is the commander of Mission Delta 3 – Space Electromagnetic Warfare, a unit dedicated to “training and equipping electromagnetic warfare professionals.”

He states: “Our Guardians are at the forefront of joint operations, so we are fielding capability that best enables their success and the success of the joint force. Every day, Space Force electromagnetic warfare forces become better organized, trained, equipped, and mentally prepared to operate in hostile environments.

The Meadowlands units are mounted on a six-wheel chassis, and appear to be intended for ground deployment using a separate tractor. They’re small enough to be deployed in harsh environments or behind front lines, and can be controlled remotely.

What is Space Force?

It was reported in February 2026 that Space Force was involved in Operation Midnight Hammer, which included gaining control of the electromagnet spectrum.

While sounding like a fictional arm of the military, Space Force was created in 2019 to protect U.S. interests in space. Operating under the Department of Air Defense, it is concerned with putting satellites into space, managing GPS, and tracking space debris. Its operatives are referred to as “Guardians.” These are recruited from the USAF, other branches of the military, or from civilian application pools.

(Conversely, NASA is concerned with exploration and scientific research, and is a civilian agency.)

Meadowlands CCS join existing high-powered microwaves, precision lasers within the arsenal of Space Force, which holds no kinetic weapons for deployment.

“This upgraded system enables us to more effectively and efficiently support the joint scheme of maneuver across the continuum of conflict,” said U.S. Space Force Lt. Col. Ryan Skilling, 4th Electromagnetic Warfare Squadron commander.

• 25 million viewers watch the first-ever 4K livestream beamed from the Moon
• 260Mbps transfer speeds send video more than 250,000 miles
• AWS GovCloud has an important role in NASA simulations

Amazon Web Services – the biggest cloud hyperscaler with a market share equaling the combined share of both Microsoft and Google – just helped to enable the first-ever like 4K video stream from the Moon.

The company transmitted video footage from the Orion spacecraft by laser, with an estimated 25 million people watching the coverage across NASA+, YouTube and Amazon’s own Prime Video platform.

Not only is this the first time 4K video has ever been transmitted from the Moon to Earth, but it’s also major news for video transmission by laser optics instead of conventional radio.

AWS enabled 260Mbps transfer speeds via laser

Key to supporting the video transmission was NASA’s Orion Artemis II Optical Communications System (O2O) – a laser-based terminal more than 20 years in the making that supports up to 260Mbps transfer speeds. In other words, fast enough for real-time 4K video and other crucial mission data.

Generally, streaming platforms like Amazon Prime Video would recommend having internet speeds of 15-20Mbps to stream 4K video. With NASA and AWS opening up 260Mbps, more than enough buffer was available to continue transmitting other key telemetry, voice communications, mission data and other files.

Unlike traditional radio systems, optical communications promise much higher bandwidth and are better suited to transmitting much larger datasets, thus they are expected to become increasingly important as space exploration pushes new boundaries.

In this case, laser transmissions were received by Mount Stromlo Observatory near Canberra, Australia, and NASA’s White Sands Complex in New Mexico was responsible for processing and distributing signals. “AWS, NASA, and ANU partnered and stood up the connection in a matter of weeks, for the cost of a laptop,” the company declared.

All in, the video was transmitted an estimated quarter of a million miles, “connecting viewers to the farthest humans ever to travel from Earth.”

But Amazon’s partnership with NASA runs even deeper than this, with the company hosting the official NASA+ streaming platform using AWS Elemental services.

Amazon plays a crucial role in NASA’s trajectory

Besides video, Amazon’s cloud infrastructure also proves instrumental to other core NASA operations. Its flight sciences team at the Johnson Space Center runs tens of thousands of trajectory simulations for every launch opportunity, producing as much as 2-5TB of data per launch window.

Amazon proudly proclaimed that these simulations run on AWS GovCloud (US), enabling maximum security for sensitive data. Using a technique known as cloud bursting, NASA is able to “scale into hundreds of additional Intel-based cloud instances on demand,” allowing it to retarget and optimize trajectory in near real-time.

Mission control aside, Amazon’s huge video streaming success lays out the framework for future space missions, including the upcoming Artemis IV lunar landing which NASA hopes to get into the hands of 250 million live viewers.

• Researchers have spotted problems with AI pattern matching in science data
• It could mean false flags for signatures of life on other planets
• AI can still be useful, but checks need to be built in

One of the ways AI can be most helpful is in trawling through masses of scientific data that human researchers don't have time to analyze, looking for patterns — but this use-case is now proving problematic when it comes to the search for life beyond our planet.

A new study from researchers at Michigan State University suggests that AI systems can be too easily fooled into identifying signatures of life out in the universe where none exist. We need these flags to be accurate to know where to point our telescopes next, so it's important that the detection processes work.

The researchers set up a digital simulation including a key sign of life: the ability for molecules to replicate and mutate. Software was used to generate tens of thousands of digital organisms with and without this ability, which where then used to trail a neural network to spot the difference with an accuracy rating of 99.7%.

When the neural network was pointed towards data it hadn't previously seen, however, the AI's life-spotting skills fell apart. The researchers started with a digital organism that couldn't copy itself, which the AI correctly identified, then began making small edits and asking the AI to check again.

Essentially, as the AI was nudged out of its comfort zone of training data, it started seeing life where there wasn't any. "No matter what sequence of commands we started with, we were able to fool the AI 100% of the time," said Ankit Gupta, one of the researchers.

Space and beyond

It's worth bearing in mind the limitations of this research: these tests were carried out in an artificial, digital simulation, and so didn't rely on any real data. The researchers were deliberately searching for errors too, rather than letting them happen by chance.

However, the study methods are solid enough to be concerning. The worry is that a Mars rover or a deep-space telescope could identify a life signature with a high degree of confidence, without necessarily having a human in the loop to check.

The researchers found there were a vast number of sequences that could trip up the AI too, meaning the risk of a mistake is more likely. While the digital organisms incorrectly identified by the neural network were close to what it had been trained to spot, they weren't full matches — despite the AI thinking they were.

These issues could crop up outside of space exploration too. The same errors might appear when looking for patterns in medical scans, security camera footage, and everywhere else the technology is used.

That said, the researchers are keen to emphasize that AI can still be useful in these scenarios — it just needs careful checks and supervision. "AI has an Achilles' heel: it can see a pattern and completely misclassify it," said Christoph Adami, one of the team. "There needs to be a human in the loop."

• Supercritical CO₂ could eliminate some of rocketry's most expensive infrastructure requirements
• Cold launches avoid exposing launch pads to destructive exhaust temperatures exceeding 3,000°C
• Engine ignition occurs only after the rocket clears the launch platform safely

Chinese aerospace startup Z-Trak Space is exploring an unusual launch system using carbon dioxide (CO₂) commonly associated with fizzy drinks rather than rocket exhaust.

The proposal centres on supercritical CO₂, a state achieved when the gas remains above specific temperature and pressure thresholds simultaneously.

Developers believe the approach could lower infrastructure requirements while supporting faster launch schedules for China's expanding commercial space sector.

A cold launch system could reduce dependence on expensive ground facilities

The technology ejects rockets using rapidly expanding high-pressure gas before igniting engines only after reaching a predetermined altitude overhead.

That sequence differs sharply from conventional hot launches where engines ignite directly on launch pads before the vehicle begins climbing skyward.

Because combustion occurs later in flight, scorching exhaust gases never strike the launch platform or surrounding support infrastructure below.

Traditional launch facilities require extensive protection systems because exhaust streams exceeding 3,000°C can damage structures rapidly during operations.

Launch towers and deep flame trenches therefore require substantial repairs and maintenance after missions involving repeated exposure to extreme temperatures.

According to the company, those requirements increase costs and slow launch frequencies whenever operators pursue swift response commercial missions.

“In commercial space launch scenarios requiring high frequency and rapid response, this incurs high time and financial costs, directly affecting launch pacing and response speed,” the company said.

Z-Trak Space believes eliminating those burdens could increase launch pacing while reducing dependence upon permanent launch infrastructure considerably over time.

China's commercial sector sees cold launch technology as a possible advantage

The initiative combines efforts from Hunan based Zhiyu Aerospace Technology and Chiyang Space Power Technology Company through a recent collaboration agreement.

Z-Trak Space founder Zhang Zihan, who studied aerospace engineering at Hong Kong Polytechnic University, described the concept as potentially transformative.

Zhang said the method would "fundamentally change the model for small liquid-fuel launch vehicles" if implemented successfully at commercial scale.

The company also argued that supercritical CO₂ launches could reduce fuel use during lower altitude portions of flight operations.

Because supercritical CO₂ is non-toxic, developers also claim launches would avoid generating harmful emissions near ground facilities locally.

Z-Trak described the system as offering "high safety, low costs, clean environmental performance and low ablation" during launch procedures.

China's commercial space industry recorded 50 launches during the previous year, representing more than half of national missions completed overall.

Whether CO₂ launches become commercially practical remains uncertain, although the proposal reflects experimentation occurring throughout China's rapidly changing space industry.

Via SCMP

• NASA's AI medical tool works where Earth-based doctors simply cannot reach
• Deep space has no signal — so NASA built its own offline doctor
• RamaLama runs AI models the same way containers run software — predictably

Astronauts aboard the International Space Station (ISS) currently rely heavily on Earth-based doctors whenever medical issues emerge hundreds of kilometres overhead.

That arrangement works reasonably well in low Earth orbit, where communication delays remain short enough for near real-time consultation sessions.

It becomes far less practical once crews travel beyond Earth orbit, where signals can take minutes rather than seconds to arrive.

An AI doctor built to work without an internet connection

Researchers at NASA's Johnson Space Center in Houston are now testing a clinical decision support system called Crew Medical Officer Digital Assistant, or CMO-DA.

The system is designed to help astronauts diagnose and treat medical conditions during deep space missions, where real-time communication with Earth-based doctors could be limited or entirely impossible.

Powering it is RamaLama, a Red Hat-backed open source tool built to simplify how developers run and serve AI models across diverse hardware environments.

According to Red Hat, RamaLama treats AI models like container images, running them in isolated, security-first environments using Open Container Initiative-compliant containers that are portable and predictable across hardware.

This approach allows the CMO-DA to perform what the team calls multimodal inference — processing both large language models for complex medical reasoning and Vision Language Models for image-based symptom analysis.

The system can therefore evaluate both written symptom descriptions and visual data without requiring any connection to a terrestrial cloud server.

That offline capability is not a convenience feature but a mission-critical requirement, since deep space communication delays make cloud dependency genuinely dangerous for crew health outcomes.

Testing is currently running on HPE hardware — specifically the terrestrial twin of the Spaceborne Computer already aboard the ISS — giving researchers a reliable Earth-based replica of the actual deployment environment.

By using open-source tools throughout, NASA researchers have built a system that is reproducible and auditable, which the team describes as essential for human safety in mission-critical environments.

From Earth testing to the ISS and beyond

Once the terrestrial testing phase is complete, the CMO-DA will be demonstrated to NASA leadership for evaluation of further deployment aboard the International Space Station.

The next iteration of the system will integrate Red Hat Enterprise Linux AI, known as RHEL AI.

This is to provide a stable, hardened foundation for scaling and managing containerized AI applications in remote and extreme environments.

RamaLama itself was built with a stated goal of making AI "boring" — meaning reliable, predictable, and unglamorous in the best possible sense for mission-critical applications.

The same architecture being tested for astronaut health could eventually serve as a blueprint for delivering medical support in the most remote areas on Earth.

Whether the CMO-DA eventually evolves into something resembling Star Trek's handheld Tricorder remains unknown.

What is known is that an open source AI tool is already diagnosing symptoms aboard a replica of hardware currently orbiting Earth.

Elon Musk's long-term ambition of colonizing Mars is an idea you'd normally find in science fiction novels rather than something actually on the cards for humanity today. Yet that's what the tech entrepreneur is hoping we can achieve within his lifetime. As for the first humans on Mars? It's likely to be much sooner – and even as soon as 2030.

Life on Mars

Musk bluntly outlined the possible risks that any manned mission to Mars will involve when speaking at the International Astronautical Congress (IAC) a decade ago.

His comments underlined the fact that the engineers and astronauts involved in the first flights will have to accept extreme risk – and serve as a means of psychological filtering for anybody considering a role in such a mission as and when it may happen in the future.

During his talk, Musk also launched a concept design for an Interplanetary Transport System spaceship, which is designed to carry between 100 and 200 settlers to Mars, with roughly 1,000 of these vessels leaving Earth in unison.

Ashes to Ashes

Despite Musk's longstanding ambition for SpaceX to get people onto Mars, developments in recent months and years have stalled progress – especially a pivot from the company to focus on the Moon.

Musk formally delayed the Mars infrastructure timeline by up to seven years in February, citing the need to protect humanity's future by building a "self-growing city" on the lunar surface. This is despite saying in January last year that "we're going straight to Mars" while describing the Moon as "a distraction".

Musk's comments over the last few years about getting to Mars have come across as overly ambitious, putting it mildly. Critics suggest that SpaceX may never actually achieve this – including Britain's chief astrophysicist Lord Martin Rees, who described the plans as "dangerous delusion".

• SpaceX is offering half-price Starlink for some users in Memphis
• Reisdents near an xAI data center are eligible for the offer
• The company is being sued for pollution caused by its infrastructure

Here's one way to try and overcome local opposition to your energy-hungry, water-sucking, AI-feeding data centers: offer residents some sweeteners in return for having the infrastructure built in their backyard.

As Bloomberg reports, SpaceX is now giving out half-price Starlink subscriptions to those around the Memphis, Tennessee area, where the company operates existing Colossus data centers and "continues to invest in the Memphis area".

There will be no upfront hardware costs for new customers, and the discounts can be shared with "nearby" friends and family through a referral. However, if you change address and move out of the area, you lose your discount.

While 'appeasing local residents' isn't explicitly mentioned in the SpaceX statement, it doesn't take much reading between the lines. If you live around Memphis or neighboring Southaven, you can see if your address is in an eligible part of the area by looking up the location online.

Pollution problems

The data centers in this region are primarily for SpaceX subsidiary xAI, and the company is currently being sued for violating the Clean Air Act: it's claimed that the methane gas turbines powering the Colossus data centers don't have the required permissions.

"As we shared since xAI started operating in Memphis, our homes, churches, and playgrounds will not be sacrifice zones for Big Tech’s convenience," says Abre’ Conner, a director at the National Association for the Advancement of Colored People (NAACP).

It remains to be seen whether half-price internet changes the mood of those affected: the discount might also be known "as the please don't stop us from running illegal natural gas turbines discount" quips one user on Reddit,

Discounted Starlink packages aren't the only commitment xAI is making though. The company has also said it will resume work on a waste water recycling plant from next year, which will provide clean water for the data centers and local authorities, though it isn't clear why work was halted to begin with.

• SpaceX is reportedly keen on a Starlink mobile network
• The company might try and build out its own infrastructure
• Analysts believe a carrier deal is going to be more likely

SpaceX is reportedly planning to launch a Starlink-branded mobile network to compete with Verizon Wireless, AT&T, and T-Mobile in the US — though not everyone is convinced that the company will be able to make much of a dent in the market.

According to the Financial Times, SpaceX has recently shared its ambitions with investors, and intends to offer contracts to individual consumers. It already partners with T-Mobile in the US to provide emergency satellite messaging services.

This wouldn't be a satellite-based network though — that wouldn't be able to offer enough capacity on its own. The thinking is that SpaceX plans to build out its own terrestrial infrastructure, and indeed has already paid $17 billion to EchoStar for a chunk of wireless spectrum licences.

And the big three carriers have so far refused to do Mobile Virtual Network Operator (MVNO) deals with SpaceX, though some analysts are of the opinion that agreements may be worked out eventually — possibly with the EchoStar spectrum as leverage.

Satellites and broadband

Setting up a Starlink Mobile MVNO, purchasing an existing MVNO, or even acquiring an existing carrier are all options on the table for SpaceX, according to current speculation. All of those paths would seem to be easier than building new cell towers and base stations.

Analyst David Barden, from New Street Research, told the FT that it would be "incredibly hard" for SpaceX to build its own networks in a "saturated" market like the US. However, he also agrees that this might be a negotiating tactic SpaceX is using to get the existing networks to cut a deal.

SpaceX and Starlink are clearly keen on expanding their communications offerings, with next-gen satellites in the pipeline, and all of the top-end smartphones available to buy can now connect to these satellites without any extra hardware.

Add in the growing Starlink broadband business and the reported terrestrial network capacity as well, and we could be looking at a very comprehensive offering from Starlink in the not-too-distant future.

Although the technology powering them has evolved in performance and efficiency, the core principle of data centers hasn’t changed much in the past two decades.

Every new generation of infrastructure has been designed to shorten the distance to users, reducing latency and increasing reliability. However AI is putting serious strain on that model, and unprecedented demand for data centers is making companies think outside the box.

Terrestrial projects now face challenges from all directions, including land availability, cooling, power and water consumption, and local opposition.

Step in orbital data centers. Some pioneers are now looking to rocket data centers up into space, and the model keeps shifting further from concept and closer to reality.

While enormous technical hurdles remain, advocates argue orbital data centers could eventually complement terrestrial campuses by shifting some of the most energy-intensive AI workloads away from land-based constraints.

Networking is what will define good orbital compute

Some of the benefits include virtually unlimited space to expand, uninterrupted access to green solar energy and cooler environments that don’t require so much cooling.

But while sending compute into space tackles some of the problems we face today, it reverses one of the biggest fixes we’ve spent years improving – latency.

In today’s instant economy, a data center has no use if it can’t exchange information quickly and reliably. Terrestrial networks lean heavily on fiber networks and physical connections, but orbital data centers would need to improve wireless transmission across hundreds of miles.

As work on this progresses, Internet exchange operator DE-CIX CEO Ivo Ivanov says orbital compute should be viewed as another layer in our increasingly distributed digital ecosystem – not an outright replacement for terrestrial compute.

To better understand whether orbital data centers are just another ambitious experiment or whether they could be the next evolution of digital infrastructure, I spoke with Ivanov about the industry’s biggest challenges.

• You noted that SpaceX's IPO has brought the idea of AI data centers in space into the mainstream. Setting the hype aside, how close is this to real, and what's actually driving the biggest companies in the world to look at space?

The idea of placing data centers in orbit is the kind of thing that immediately captures people's imagination. It belongs in the same category of technological ambition as reusable rockets, commercial spaceflight, or landing on the Moon.

For decades it felt like something that belonged in science fiction. Today, for the first time, it's being discussed as a serious infrastructure proposition. But I don’t think the real story here is about space – it’s about AI, and how it’s altering our global trajectory.

Over the last few years, we've seen an extraordinary increase in demand for compute capacity. AI training clusters are growing larger, power requirements are rising, and in many regions the availability of energy, land, and cooling has become a genuine constraint.

These constraints are what drive innovation, and that’s what we’re seeing play out here.

In theory, space offers access to continuous solar energy and room to scale without the physical limitations we experience here on Earth. Having said that, we're still at a very early stage.

I don't think we're talking about replacing terrestrial infrastructure anytime soon. What we're seeing today is the beginning of a serious exploration into how orbital resources could complement the digital infrastructure ecosystem that already exists.

There is still work to be done, particularly around networking – putting an AI data center into orbit would be a milestone achievement, but it won’t count for much if we have to truck hard drives into space to do any training. For now, all anybody can say with confidence is that it's becoming part of a broader conversation about how society will support the next generation of AI workloads.

• What problems could orbital data centers potentially solve that terrestrial infrastructure struggles with today?

I think the broader answer is that orbital infrastructure has the potential to ease some of the physical constraints that have shaped and limited digital infrastructure for decades. We can’t simply keep building more data centers – they need access to land, power, cooling, transport links, and of course, connectivity.

As AI continues to grow, some of those resources will become harder to secure, especially at the pace the industry wants to move.

That's why the industry is starting to look at non-terrestrial alternatives. Suddenly, you're talking about access to potentially unlimited solar energy, plenty of space, and the opportunity to scale at a pace that matches our ambitions for AI.

But putting data centers into orbit, as impressive as it would be, is only the first step. Google’s Project Suncatcher is already exploring how power might be harvested from the sun, while the European Space Agency’s OFELIAS project is exploring how optical feeder links between Earth and orbit can be optimized to reduce latency and disruption. There are lots of moving parts.

• From a networking perspective, what challenges will emerge once data centers move into orbit?

The moment you put a data center into orbit, networking becomes the make-or-break factor. A data center on Earth can usually rely on a dense ecosystem of fiber routes, Internet Exchanges, cloud on-ramps, and interconnection facilities. In orbit, that ecosystem doesn't exist yet.

Every workload, every application, and every AI model depends on data moving between different locations, and suddenly those journeys become much more complex.

A lot of the discussion I’m hearing focuses on bandwidth or speed, but I think predictability is the more interesting challenge. AI systems need data to arrive quickly, but they also need it to arrive consistently.

Optical and laser-based communications have enormous potential, but they also introduce new variables like cloud cover, atmospheric turbulence, satellite handovers, and changing orbital positions that can all affect how data moves between Earth and space.

That's why I believe the real challenge isn't simply connecting an orbital data center to the ground, but creating an interconnection layer that makes orbital, terrestrial, cloud, and edge infrastructure behave as though they're part of the same ecosystem.

That's a much bigger networking challenge than people realize.

• Elon Musk said orbital infrastructure is "only a few milliseconds away." How much could those milliseconds matter, and is latency the biggest challenge?

Those milliseconds absolutely matter. AI systems are fundamentally data-driven, and every additional moment spent moving information between users, applications, models, and infrastructure affects performance and productivity.

That's one of the reasons the industry is investing so heavily in edge computing – the closer you can bring compute resources to the data, applications, and users they serve, the better the experience tends to be.

Certainly, low-Earth orbit is only milliseconds away, but for latency-sensitive use cases, the 20-40 ms it takes to reach the stratosphere will provide a significant challenge for some AI inference use cases. So not all applications can possibly be served from hundreds of kilometers above the ground.

Having said that, performance isn't just a question of latency. A network that consistently delivers twenty-five milliseconds of latency is often more valuable for many enterprise use-cases than one that fluctuates between twenty and a hundred.

Physics determines how quickly data can travel, but factors such as atmospheric conditions, satellite handovers, routing decisions, and network architecture all influence how predictable that experience is.

• If a data center is in orbit and the AI workloads it serves are on Earth, what actually has to happen for those two worlds to behave as one network?

The way I like to think about it is that users shouldn't have to think about where a workload is running. Whether an application is being served from a terrestrial data center, an edge location, a cloud region, or one day an orbital platform, the experience should feel seamless.

Achieving that is much harder than it sounds, because data has to move continuously between different environments, and those environments need to behave as though they're part of the same network, even when they're separated by hundreds or thousands of miles and, increasingly, by the boundary between Earth and space.

That's one of the reasons projects like the ESA's OFELIAS initiative, which I mentioned earlier, are so important. As part of that initiative, DE-CIX is working alongside partners including the German Aerospace Center to explore how optical feeder links between satellites and the ground can become more stable, efficient, and predictable.

Rather than simply building more connections, we need to create an infrastructure foundation that allows terrestrial, satellite, and orbital resources to work together as a single ecosystem.

We've spent decades building that kind of seamless interconnection on Earth, and now we need to take what we’ve learned and combine it with new technologies to bring non-terrestrial infrastructure into the picture.

• Where does this leave terrestrial data centers and interconnection? Will orbital compute compete with what exists on the ground, or depend on it?

I don't see this as a competition at all. Digital infrastructure has a long history of adding new layers rather than replacing existing ones. Cloud computing didn't eliminate enterprise data centers, edge computing hasn't eliminated centralized clouds, and satellite connectivity hasn't replaced terrestrial networks.

Instead, each technology tends to find the role it's best suited to, and the overall ecosystem becomes more capable as a result.

I think orbital compute will follow a similar path. Some AI workloads may benefit from being processed closer to abundant energy sources in orbit, while others will remain in terrestrial edge data centers because they require ultra-low latency inference, regulatory compliance, or proximity to users and applications.

It’s not about choosing one over the other, but creating an architecture where workloads can be placed wherever they make the most sense and where data can move seamlessly between them.

• What should the industry be doing now to prepare networks for orbital compute?

I think the first step is recognizing that orbital compute is more than a compute challenge. It's a networking one. The industry has spent decades building highly interconnected ecosystems on Earth, where data can move efficiently between enterprises, cloud providers, networks, content platforms, and users.

If compute begins expanding into orbit, we'll need to ensure the same level of interconnection exists between terrestrial, satellite, and orbital environments. Advances in satellite communications, laser-based networking, and distributed infrastructure are creating new possibilities that didn't exist a decade ago.

What I'd like to see now is greater collaboration between network operators, cloud providers, infrastructure companies, and space organizations.

The technologies themselves are advancing rapidly, but the challenge is ensuring they evolve as part of a coherent ecosystem rather than a collection of isolated systems.

• Looking ten years ahead, what does a fully interconnected ecosystem spanning terrestrial, edge, cloud, satellite, and orbital infrastructure actually look like? What technologies will make it work?

The most interesting thing about the future is that people probably won't think about infrastructure at all. Today, we spend a lot of time talking about whether something runs in a cloud region, an edge location, a data center, or on a particular network.

Ten years from now, if we succeed with orbital infrastructure, workloads will simply move to wherever they can be processed most efficiently, and data will flow across terrestrial, satellite, and orbital environments as naturally as it moves across continents today.

When we look back at the great infrastructure achievements of the past, whether it's railways, electricity grids, undersea cables, or the Internet itself, their success wasn't measured by the technology, but by how completely they disappeared into everyday life.

People stopped thinking about the infrastructure and started focusing on what it enabled. I think we'll see something similar here. The future of digital infrastructure will be defined by our ability to connect every layer of the ecosystem into something that feels seamless.

That's when we'll know we've succeeded.

Demand for AI-enhanced products and services is so high that some of the world’s largest tech companies are now looking beyond Earth for solutions. As hyperscalers struggle with power shortages, grid connection delays and land constraints, attention has recently turned to orbital infrastructure as a potential solution.

Recently, Meta reportedly reserved around 1GW of solar capacity as part of its longer-term plans to support future AI data centers, while SpaceX’s Elon Musk has repeatedly discussed ambitions to make orbital infrastructure more accessible.

On paper, the logic is easy to understand, with abundant solar energy to meet power needs and none of the land constraints imposed on Earth. However, engineering for space-bound infrastructure is only part of the worry, and some security and infrastructure experts warn the industry could be dramatically underestimating the operational risks involved.

According to Acre Security CEO Kumar Sokka, the biggest challenge might not be compute, launch economics or cooling, but rather resilience.

Who’s going to look after orbital infrastructure? And how?

Terrestrial data centers are built around one critical assumption – that if something were to go wrong, someone can physically read the hardware to make the necessary changes. Technicians can’t exactly hop on a rocket to replace failed components, swap power systems or restore infrastructure as easily.

As a result, a hardware failure that might take hours to resolve on Earth could become a months-long problem in space, dependent on launch schedules, robotic repair systems and even complete satellite replacement.

Operators would also have to consider space debris, radiation exposure and thermal extremes on top of hostile maintenance environments.

To better understand the resilience and security implications of moving AI infrastructure into orbit, I spoke with Acre Security’s Kumar Sokka about why orbital compute may fundamentally challenge what we know about AI compute today, how outage recovery changes when hardware becomes inaccessible, and why the industry could be trading one type of constraint or risk for another.

• Hyperscalers have deep pockets and the brightest minds. Surely they've considered resilience? What could go wrong?

These are genuinely excellent engineering teams, and the rigour they've brought to terrestrial infrastructure over the last two decades is real. We work alongside that infrastructure every day.

But the resilience frameworks they've built are all grounded in one assumption: someone can physically reach the hardware. That's what makes layered access control, physical redundancy, and rapid component replacement possible. The moment infrastructure moves to orbit, that assumption disappears.

A routine fix that takes four hours on the ground becomes a three-to-six month problem involving a launch window. What we're seeing with recent announcements about compute demand outpacing what terrestrial power and land can deliver tells us this isn't a theoretical conversation anymore.

The engineering ambition is real, and the physical security frameworks need to keep pace with it.

• How would data centres in space differ from existing satellites when it comes to risk mitigation and redundancy?

The architecture is fundamentally different. Satellites are purpose-built, designed to operate independently, and constellations are built so that if one node fails, traffic routes around it. Data centres don't work that way.

You're running interdependent workloads where a single compute job may span thousands of processors, and a partial failure can bring the whole job down.

The fault tolerance models that work well for independent satellite nodes don't translate cleanly to complex, tightly coupled compute infrastructure. That gap hasn't been fully addressed yet, and it's where we think the industry needs to invest serious thinking.

• Could hot-swappable redundancy or 50% extra capacity solve the maintenance problem?

In principle, yes, but economics can create real tension. On the ground, spare capacity is relatively inexpensive. You run N+1 configurations, you keep hardware in a warehouse, the cost is manageable.

In orbit, every kilogram of redundant hardware carries a launch cost. And hot-swapping in a vacuum, in microgravity, with active thermal management requirements, requires automated repair capability we don't yet have at scale.

The ISS was specifically designed for human maintenance and still requires EVAs for hardware work. If you're talking about the scale some operators are projecting, the maintenance model needs to be fundamentally reimagined before the infrastructure gets there.

• How should the industry evolve its strategy to protect launch-and-forget, unreachable hardware?

This is the question we find most interesting, because orbital infrastructure inverts a lot of what physical security is built around. Our discipline is fundamentally about controlling access, who gets in, what they can reach, how you monitor the perimeter.

In orbit, that model flips: no one can get in, including the operators. So the strategy has to shift from protect-and-respond to predict-and-pre-empt. That means self-diagnosing systems, AI-driven anomaly detection that identifies component degradation before it becomes failure, and hardware designed from day one for graceful degradation rather than hard failure.

You also have the debris environment to account for, tens of thousands of tracked objects, millions of smaller fragments, all at extraordinary velocity. You can't fence that perimeter. You can only design around it, and you have to do that thinking before the hardware launches.

• Isn't colocation on the same orbital platform acceptable in a hybrid system with Earth-based capacity running concurrently?

If you're genuinely treating orbital compute as supplementary, with full failover capacity maintained on the ground, then the single-point-of-failure risk is managed. But the underlying business logic creates pressure in the other direction.

The argument for space infrastructure in the first place is that terrestrial capacity can't meet demand. Once that becomes true, the economics push operators toward dependency rather than redundancy.

If you're running production workloads in orbit that genuinely can't run elsewhere, the hybrid framing starts to break down. What was designed as a backup becomes a critical dependency, and that shift can happen gradually without anyone making an explicit decision.

• Acre protects Google, AWS and other hyperscalers. Isn't space an opportunity for your business?

It's something we're actively thinking about, but I'd be honest that our focus right now is on getting the unified platform vision right here on the ground.

We're building infrastructure that brings access control, intrusion detection, video, and visitor management together in a way that works across enormously complex, distributed facilities, and that work isn't done. What space does is sharpen the thinking.

The principles we're developing, unified monitoring, anomaly detection across interdependent systems, designing for environments where you can't always send someone in to fix a problem, those translate.

When the industry is ready to have a serious conversation about orbital security strategy, we want to be the organisation that's already thought it through. But the most valuable thing we can do right now is build the platform foundation that makes that possible.

• SpaceX admits global GPU shortages threaten its orbital AI computing ambitions
• Orbital data centers cannot operate without enormous volumes of advanced semiconductor hardware
• TeraFab remains uncertain despite massive investment and ambitious semiconductor production goals

SpaceX has filed paperwork for its upcoming IPO that reveals a troubling dependency on a handful of GPU suppliers.

The company's Form S-1 document openly states that orbital AI ambitions require more chips than the market currently provides.

Elon Musk's enterprise procures all its GPUs on a purchase-order basis without any long-term contractual arrangements with direct chip suppliers.

Space ambitions collide with a brutal global GPU shortage

This purchase approach leaves SpaceX vulnerable to every disruption that hits semiconductor manufacturing, from natural disasters to geopolitical conflicts.

The company envisions putting computing infrastructure in space, but that vision requires a volume of GPUs far beyond what any supplier can currently deliver.

Major AI chip buyers like Nvidia have already locked down massive supply commitments totaling $145 billion, which pushes smaller purchasers further back in the queue.

SpaceX admits that "manufacturing and supply of servers and network equipment for our technical infrastructure, particularly for GPUs and other specialized components, is limited to a small number of qualified suppliers."

"Our ability to achieve orbital AI at scale depends on our ability to access a sufficient number of AI chips, significantly more than are currently available to us," the filing SpaceX reads.

Orbital data centers cannot launch without these components, and the current procurement strategy offers no protection against shortages.

To escape this dependency, SpaceX, together with Tesla and xAI, plans to build TeraFab, a dedicated semiconductor facility in Texas using Intel's 14A process technology.

The project requires tens of billions of dollars in investment, yet the S-1 filing explicitly warns that TeraFab may fail.

"While we expect to construct Terafab to address such supply constraints, Terafab may not be successful, in which case we may not have other sources of sufficient AI chips to meet our orbital AI compute demands," the document states.

Even if construction succeeds, the company expects to continue sourcing most hardware from third-party suppliers

This means that the orbital data center plan remains tethered to the same unreliable market.

TeraFab still depends heavily upon unstable partnerships and uncertain execution

At the moment, the partnership between SpaceX, Tesla, and Intel on TeraFab remains unstable, and there is no legal obligation for any of the parties to remain committed.

"While we have a framework agreement with Tesla, neither Tesla nor Intel are obligated to remain a part of the project, and we may not enter into any such definitive agreements," the Form S-1 reads.

If either partner departs, TeraFab loses both a significant customer and the process technology developer needed to make the chips.

IPO risk factors often include improbable disasters, so these admissions require measured interpretation.

Yet SpaceX has identified a concrete bottleneck: orbital data centers demand chips that do not exist in sufficient quantities anywhere on Earth.

No amount of rocket reusability or satellite engineering can bypass a foundry’s limited wafer output, and TeraFab remains a gamble rather than a guarantee.

Via Tomshardware

• Voyager still operates using assembly code written nearly half a century ago
• NASA maintains interstellar spacecraft with less memory than a smartphone image today
• The engineers who built Voyager are disappearing faster than the spacecraft itself

Launches in 1977, NASA's twin Voyager spacecraft continues operating with onboard computers which run assembly language written for custom General Electric processors.

Each spacecraft carries three separate computer systems, with a total memory of roughly 64 to 70 kilobytes across all three - less storage than a single small image file on a modern smartphone today.

NASA's Suzy Dodd has compared operating Voyager to flying an Apple II, capturing how primitive the computing resources have become by modern standards.

What the spacecraft actually runs and why the language matters

The popular shorthand often says Voyager runs on Fortran, but that description blurs two different things together.

The spacecraft's low-level flight work depends on assembly language programming on highly specialised hardware designed in the early 1970s.

Fortran has been associated with ground systems and older mission tooling, not with the onboard flight software itself.

When NASA went looking for a replacement engineer in 2015, the job posting covered both assembly language skills and a deep understanding of the spacecraft's unique hardware architecture.

Forty-nine years of continuous operations have produced knowledge gaps that matter far more than the programming language itself.

Around the start of the interstellar mission after Voyager 2's flyby of Neptune in August 1989, the flight software was updated to make each spacecraft more autonomous.

That version, augmented by command sequences the team uploads every few months, is the basis of what is running now on both probes.

However, the team has shrunk and aged dramatically across decades, and much of the original paper documentation has been lost or fragmented over time.

Original engineers are no longer available to help

Larry Zottarelli was the last original Voyager engineer still working on the project when he retired in 2016 at the age of 80.

All of the other original engineers are dead or over 90, like Dr. Gary Flandro, an aerospace/trajectory engineer who is now living in retirement.

Dodd told Live Science in early 2024 that the people who built the spacecraft are not alive anymore, leaving a shrinking team to maintain code that no one fully understands.

The Voyager signal now takes more than 23 hours to reach Earth, and by the time NASA receives the next status check, the spacecraft will already be 1.5 million kilometres further into interstellar space.

The mission continues, but the institutional memory that built it is fading faster than the plutonium power sources that keep the probes alive.

Each passing year takes more of that knowledge with it, and when the last engineer who understands the assembly code retires or passes away, NASA will be left with paper documentation, a diminishing signal, and a spacecraft that no one alive can truly repair.

Via SpaceDaily

• NASA developed autonomous spacecraft processors with dramatically higher computing performance levels
• New radiation-hardened chip delivers hundreds of times greater processing capability
• Deep-space communication delays are driving demand for autonomous onboard decision systems

A famous warning about autonomous machines from the 1968 film 2001: A Space Odyssey appears to have faded from NASA's memory, if it was ever taken seriously at all.

The space agency is now developing a powerful new processor that could allow spacecraft to make independent decisions during deep space missions.

As part of the High-Performance Spaceflight Computing (HPSC) project, this technology aims to reduce reliance on Earth-based controllers, which currently face long communication delays.

A leap in space computing performance

NASA claims its new radiation-hardened chip delivers up to 100 times more computing power than current spaceflight hardware, and early test results have even shown performance levels roughly 500 times greater than existing radiation-protected processors.

“Building on the legacy of previous space processors, this new multicore system is fault-tolerant, flexible, and extremely high-performing,” said Eugene Schwanbeck, a program manager at NASA Langley Research Center.

Any processor destined for deep space must endure extreme electromagnetic radiation and dramatic temperature swings.

High-energy particles from the Sun can easily trigger computer errors that force conventional spacecraft into a protective “safe mode.”

Engineers at NASA’s Jet Propulsion Laboratory are subjecting the prototype to punishing simulations of these conditions.

“We are putting these new chips through the wringer by carrying out radiation, thermal, and shock tests,” explained Jim Butler, the project manager for High Performance Space Computing at JPL.

The chip must also handle the unique challenges of planetary landings without human intervention.

These dramatic improvements raise a legitimate question about whether engineers have considered the potential risks of truly autonomous machines.

Autonomy versus the ghost of HAL

Science fiction enthusiasts reading this might immediately recall 2001: A Space Odyssey, where a thinking computer named HAL tragically malfunctions in a story serving as a cautionary tale about granting machines too much independent authority over human lives.

NASA now envisions spacecraft that can process scientific data instantly and respond to unexpected hazards without waiting for instructions from Earth.

The agency is testing how the chip handles high-fidelity landing scenarios that would normally require power-intensive hardware to process massive sensor data volumes.

One must acknowledge that modern spacecraft already rely on automated systems for many routine functions.

The difference here lies in the scale of autonomy and the use of onboard artificial intelligence for mission-critical decisions.

NASA’s collaboration with Microchip Technology has already produced sample chips for defense and commercial aerospace partners.

The finished processor could eventually support crewed missions to the Moon and Mars, where communication delays of several seconds would make real-time human control impractical.

Whether this technological leap invites unforeseen risks remains an open question - after all, HAL’s famous line, “I’m sorry, Dave, I’m afraid I can’t do that,” began with the best engineering intentions.

Via ScienceDaily

• Google and SpaceX are reportedly in talks over data centers in orbit
• It's the infrastructure that many big tech companies are investing in
• However, there are concerns over how feasible the technology actually is

There's a lot of effort and money going into plans to put data centers in space — mostly to meet the growing demands of AI processing — and Google and SpaceX are now reportedly in talks to clutter up Earth's orbit even further.

As per The Wall Street Journal and "people familiar with the discussions", talks are underway between the two giant tech companies. Google would provide the data centers, and SpaceX would provide the rockets to get them where they need to be.

There's not a lot of other detail here, although the report says Google is discussing its options with "other rocket-launch companies". It's a sensitive time for SpaceX, with an IPO expected in just a few months, and deals like this would certainly do no harm as far as the company's valuation goes.

It's also worth noting that Google owns 6.1% of SpaceX. Google has made no secret of its plans to put data centers up in space: "There's no doubt to me that a decade or so away, we'll be viewing it as a more normal way to build data centers," Google CEO Sundar Pichai told Fox News in an interview last November.

But can it work?

Given that so many big companies are investing in this technology, you'd be forgiven for thinking that it's all ready to go — but that's not quite the case.

As it stands, the tech isn't currently available due to various engineering challenges, and people like OpenAI's Sam Altman don't think this is a feasible idea in the short-term.

On the plus side: there's a lot of room up in space (though not quite as much as there used to be), and a lot of power, thanks to the constant availability of solar energy. It wouldn't be necessary to disrupt neighborhoods or overload power grids every time a giant new data center needed to be built.

However, this is also incredibly costly and incredibly difficult from a technical perspective. Current data center technology isn't built to withstand the extremes of space, and there's also the question of how repairs can be carried out in orbit.

For those who fully believe in the future of the technology, those problems can be overcome given time, funding, and further research. However, it's going to be a while before we see this working at scale.

• AMD frames openness as essential for multi-vendor space mission architectures
• Vendor lock-in risks increase significantly in long-duration orbital deployments
• Modular systems improve flexibility across complex, multi-supplier mission environments

AMD has outlined its vision for AI in space, claiming to be a champion of open platforms and modular design in a market where monolithic solutions dominate.

The company argues no single vendor can or should dictate the full solution for space missions, which are often the work of multiple firms.

"Space missions are assembled from many specialized suppliers, and no single vendor can (or should) dictate the full solution," AMD stated in a recent announcement.

AMD pushes open platforms for space

AMD’s argument rests on the structure of the space industry itself. Missions typically combine hardware, software, and subsystems from multiple contractors.

This makes interoperability a requirement rather than a preference, as components must work together across different vendors.

In that context, proprietary platforms risk introducing dependencies that can limit flexibility or complicate long-term operations.

The company is leaning into open standards and modular design to reduce that friction.

Its strategy centers on enabling partners to integrate and validate systems across vendors without being tied to a single ecosystem.

That includes investment in open approaches to security, interconnects, and infrastructure, alongside its ROCm software stack for AI and high-performance computing.

ROCm is intended to give developers a pathway from low-level kernels to full applications on AMD accelerators.

More importantly, it represents an alternative to tightly controlled software ecosystems that dominate AI development today.

Why space amplifies the need for modular systems

AMD ties its openness strategy directly to the realities of operating in orbit. Space systems face strict power and thermal limits, intermittent communication with Earth, and long mission lifecycles.

These constraints make adaptability and resilience more critical than in most terrestrial deployments.

In such environments, reliance on a single vendor can introduce risk. If a component becomes obsolete or unsupported, replacing or upgrading it is far more complex than in ground-based systems.

AMD’s position is that modular, interoperable architectures allow mission designers to swap, upgrade, or validate components more easily over time.

The same logic extends to onboard AI. With limited bandwidth and communication windows, spacecraft increasingly need to process data locally.

AMD argues that open platforms make it easier to deploy and evolve these capabilities across heterogeneous hardware, rather than locking missions into a fixed stack from launch.

Openness alone may not be enough

The challenge for AMD is that the space market has historically rewarded proven reliability over architectural philosophy.

Competitors already have deep relationships with space agencies and, in some cases, purpose-built hardware designed specifically for radiation-heavy environments.

AMD points to its existing track record, including contributions to image processing for NASA missions.

However, extending that experience into large-scale AI infrastructure in orbit is a different step.

For now, AMD is making its case early, framing openness as not just a design preference but a requirement for resilience in space.

Whether that argument translates into contracts will depend less on philosophy and more on execution in an environment where failure is not easily tolerated.

• The FCC has relaxed low-orbit satellite restrictions
• It works out to a 7x capacity increase for internet satellites
• The move is welcomed by SpaceX and Amazon

The US Federal Communications Commission (FCC) has delivered some good news to Elon Musk's Starlink and other satellite internet providers, by removing several key restrictions on how much energy satellites can transmit to and receive from ground stations.

Technically, these are the Equivalent Power Flux Density (EPFD) rules set in the 1990s, as PC Mag reports. In simple terms, it means Starlink and others will be able to operate a much denser array of satellites for providing services to customers.

We're talking as much as a sevenfold increase in satellite network capacity, which should mean a connection that's more reliable and faster. However, it's going to take a while before satellite companies will be able to take advantage.

"This is a major step toward enhancing the satellite broadband experience for millions of Americans by enabling faster speeds, lower costs, and greater reliability," says the FCC. "This change could also unlock more than $2 billion in economic benefits for the American people and up to seven-fold more capacity for space-based broadband services."

Sharing spectrum

The rules were introduced to prevent radio signal interference between satellites in non-geostationary orbit (NGSO) — also known as low-Earth orbit — from affecting geostationary satellites (GSO) higher in orbit. Now, both NGSO and GSO satellites will be able to share more of the communications spectrum.

It's a change that's been welcomed by Starlink's David Goldman, as well as Brian Huseman, a Vice President at Amazon — another company looking to build out a network of internet-providing satellites with Amazon Leo.

As Gizmodo reports, Viasat is one of the GSO satellite companies that has warned that the move could potentially lead to more interference without proper regulation. Viasat has also raised concerns about Starlink getting a monopoly on this particular market.

The view of the FCC is that modern satellites are designed to share spectrum more effectively than they did in the past, and the 1990s rules no longer apply. Starlink, owned by SpaceX, will no doubt be keen to get its next batch of satellites launched.

• SpaceX admits orbital AI data centers may never become commercially viable
• SpaceX S-1 filing reveals unproven technologies behind space-based computing infrastructure
• Harsh space conditions threaten the reliability of sensitive AI hardware systems

SpaceX has warned potential investors its ambitious plans to build AI data centers in orbit may never become commercially viable due to unproven technologies and the harsh realities of space.

The company disclosed these risks in its pre-IPO S-1 filing, which US securities law requires to inform investors of potential pitfalls while shielding the company from future legal liability.

"Our initiatives to develop orbital AI compute and in-orbit, lunar, and interplanetary industrialization are in early stages, involve technical complexity and unproven technologies, and may not achieve commercial viability," SpaceX said in an excerpt from the filing seen by Reuters.

A reality check behind the hype

Any future orbital data centers will operate "in the harsh and unpredictable environment of space, exposing them to a wide and unique range of space-related risks that could cause them to malfunction or fail," the document added.

Elon Musk has been characteristically bullish about space-based AI in recent public appearances.

He said at the World Economic Forum in January 2026 that building AI data centers in space was "a no-brainer" and that it would be the cheapest place to put AI within two to three years.

In February, after announcing a merger between SpaceX and xAI, he stated that "space-based AI is obviously the only way to scale."

The S-1 filing, however, presents a far more cautious assessment, acknowledging that the necessary technologies are still unproven and may never work reliably in orbit.

AI tools which function perfectly on Earth would need to withstand the environmental conditions of space without any possibility of on-site repairs.

To deploy data centers in space, SpaceX relies on Starship, its next-generation fully reusable rocket, but it has suffered several delays and testing failures.

"Any failure or delay in the development of Starship at scale or in achieving the required launch cadence, reusability, and capabilities thereof would delay or limit our ability to execute our growth strategy," the filing said.

If Starship does not achieve its promised launch cadence and reusability, the economics of placing a data center in orbit collapse entirely.

What must be solved before space data centers can work

The filing's warnings boil down to one fundamental problem: no one has ever built and operated a data center in space before.

Radiation can corrupt memory and damage electronics beyond what Earth-based shielding can easily prevent.

Temperature swings between sunlight and shadow can stress components beyond their design limits.

There is no way to repair or upgrade hardware once it is in orbit, meaning every component must work perfectly for its entire intended lifespan.

SpaceX would need to solve all of these problems while also making the economics work against Earth-based alternatives that improve every year.

A data center on the ground costs less to build, less to maintain, and technicians can fix it with a spare part and a screwdriver.

Until SpaceX shows that its orbital infrastructure can operate reliably and affordably in space, the warnings in the filing are not just legal boilerplate but a genuine assessment of commercial reality.

• A $6.27 billion GPS overhaul collapsed under persistent technical failures
• Testing exposed system flaws that risked global GPS reliability
• US Space Force abandoned the project after the timeline became operationally irrelevant

The United States Space Force has terminated the Next Generation Operational Control System (OCX) program after it discovered extensive system issues which made timely delivery impossible.

The target of this program was to modernize the command and control of the GPS satellite constellation.

However after multiple attempts and collaborations, the Defense Acquisition Executive could no longer ignore its inefficiencies and had to let it go.

What went wrong with the $6.27 billion program?

As of January 2026, the OCX program had cost approximately $6.27 billion, including Raytheon funding and government testing costs.

The Space Force contractually accepted OCX from Raytheon in July 2025 after factory testing, then began extensive integrated systems testing.

During this testing phase, officials discovered problems across a broad range of capability areas that would put current GPS military and civilian capabilities at risk.

The system was intended to replace the current Architecture Evolution Plan (AEP) and the Launch, Anomaly, and Disposal Operations system.

Instead, the Space Force determined that additional investment in OCX was no longer the best solution for protecting and advancing GPS capabilities.

Space Force walked away

In an attempt to make the system work, the government and contractor team adopted a collaborative approach.

However, it became evident that onboarding OCX within an operationally relevant timeline is practically impossible.

"Regrettably, extensive system issues arose during the integrated testing of OCX with the broader GPS enterprise," said Mission Delta 31 Commander Col. Stephen Hobbs.

"The challenges of onboarding the system within an operationally relevant timeline proved insurmountable."

The cancellation of a $6.27 billion program after years of development is a clear admission of failure, however, the Space Force had to decide whether to walk away or sink more money into an insurmountable timeline, and it chose the former.

Going forward, instead of completing OCX, the Space Force will continue enhancing the current control system to operate the GPS satellite constellation.

It has made incremental improvements to the existing AEP system over the last ten years due to past OCX delays.

Those successful upgrades provide confidence that further enhancements to GPS ground systems will continue to support the enterprise.

“It’s important we refine and update acquisition processes to prioritize rapid, incremental capability delivery versus complex ‘all or nothing’ system deliveries,” said Acting Service Acquisition Executive Tom Ainsworth.

“The Department of War has made it clear that we need to deliver warfighting capability at a faster rate. We must continue to work with industry to meet the needs of our warfighters as we focus on delivering the right technology on the right timeline to enhance our capabilities and maintain space superiority.”

By now, we've seen dozens of incredible and inspiring Artemis II mission images, all taken with a range of cameras, including something similar to the iPhone you have in your pocket. Some of the best, though, were captured with pro-level, mirrorless cameras like the Nikon Z9, a powerful, behemoth of a 45.7MP shooter equipped with a 400mm lens.

For the past few weeks, I've been shooting with that very same kit, though, in my case, Nikon loaned me a massive 180 mm to 600 mm lens (Nikon told me it's like the newer version of the 400mm lens the astronauts used) and added a 2x teleconverter for good measure. It's inarguably the most camera I've ever used.

The Nikon Z9 is a robust (and at $4,996.95 expensive) camera that NASA chose because it can work in some pretty extreme environments. In addition to its most recent work on the Artemis II mission, it's currently on the International Space Station and is scheduled to join the Artemis III mission, which should land astronauts on the surface of the moon in 2027. The camera should, with a special enclosure, join the astronauts for space and moon walks.

When we reviewed the Nikon Z9 in 2022, we called it Nikon's "most powerful camera" yet, and I have to say it can be both powerful and complex. There are auto shooting modes, including nearly 500 points of auto focus capabilities, but the more I shot with the camera, the more I took control of the aperture, shutter, speed, and ISO, all in an effort to take the best moon shots of my life.

I mostly succeeded while also failing, largely because of my own fault.

The weight of the moment

When the full kit arrived, I realized how much I'd underestimated the size and scale of this equipment. Without a lens, the Nikon Z9 weighs almost three pounds. Part of that is to accommodate one of the largest batteries I have ever encountered in a mirrorless camera (it lasts for days). Even the storage media is larger: the camera uses a pair (one is for instant backup) of pricey Compact Flash Express cards (slightly smaller than classic compact flash).

Even though the Nikkor 180 mm to 600 mm lens ($2,095.96) features an internal zoom mechanism, it's still a foot long. Add the TC-2.0x teleconverter ($280), and it sticks out almost two more inches. The overall weight had to be close to 10 pounds, and I was glad I had a sturdy unipod, which came in handy during my first few days of trying to shoot the moon and various birds.

Let me start with a warning: if you don't have a pro-level, industry-grade tripod, keep walking. The Nikon Nikkor lens arrives with a collar that lets you attach the camera to your tripod with the weight properly balanced between the lens and the camera. That was a huge help for my unipod, but my tripod, which had been more than effective for my Sony Alpha a6000 and 200MM lens, was totally overmatched by this kit. It was barely strong enough to hold the thing still for long exposures.

A lot of my star or astrophotography was ruined by imperceptible vibrations. The other issue, I do blame on Nikon: No matter how hard I tried, I could not connect the camera via Bluetooth or WiFi to the SnapBridge app on my iPhone 17 Pro Max. With it, I could've used a remote shutter. At least I figured out how to gently touch the display to trigger a photo. When I did that, the camera would pause for a second and then snap the shot.

While the camera does not default to this setting, I chose to take every photo in both the highest resolution RAW and JPEG compressed. The former meant I could take all the information the sensor captured for every photo and pour it into Adobe Lightroom, where I would apply maximum adjustments to perfect each image.

Getting hooked

Yes, the Nikon Z9 with that massive lens is expensive and a bit overwhelming, but all it takes is one great photo to hook you, and let me tell you, I got hooked.

My first few days were spent taking some backyard bird photography. I had installed the 2X teleconverter, which essentially doubles your zoom but also amplifies any lens issues and removes a couple of lower f-stops (bigger aperture for more light), meaning you either need to increase the ISO or slow down the shutter.

The first bird photos I took on a cloudy day were grainy, even with some work in Lightroom; I didn't love them. So I opted to remove the teleconverter. After all, the birds weren't that far away (closer than the moon, at least), and with 45MP of sensor to work with, I could crop in quite a bit.

I saw instant improvement, and really appreciate the clarity and speed of the lens. The level of detail I was capturing, say, of local cardinals was stunning.

The big moment

Standing outside with the Nikon Z9 resting on my unipod, I scanned the sky for distant objects. The day was clear, and I noticed a fast-moving jet thousands of feet overhead. Mostly, I could see it's chemtrails, but I pointed the long lens at it and, for this shot, let auto settings (and autofocus do its work).

In the preview on the touch-screen viewfinder, I could see the plane, but I also thought I noticed a faint crescent nearby. Quickly, I opened Lightroom on my MacBook Pro and used the Nikon-supplied card reader and a USB-3-to-USB-C adapter to import the NEF file. That's when I saw it.

Not only had I captured the airplane, but the thin crescent moon; to have them both in frame like that was an incredibly lucky shot. In Lightroom, I did some adjustments to correct the sky color and help bring the moon and jet plane into dual focus.

You can see the photo above.

The moon and stars

I need to make something clear. The Artemis II astronaut's moon photos with the Nikon Z9 and other cameras are even more amazing when you realize that they captured these detailed views with relatively modest lenses. The Nikon Z9 they had was equipped with a 400mm lens. Nothing to sneeze at, but you look at what I captured with the Nikkor 600mm lens and then with the doubling 2x converter (akin to 1,200mm) and realize that my view is still not even remotely as close and clear as NASA's Integrity spaceship astronauts were.

It's humbling, really.

Even so, I did my level best to capture sharp, clear images of the moon and stars. As I noted above, the astrophotography was something of a bust. To do it well, you need not only to change all your settings to manual (including focus), but also to use the Nikon Z9's nifty Starlight setting on the large display viewfinder. It's an ultra-low-light view that helps immensely when you want to view the stars in your frame during what can be a challenging situation: outside, at night, in the dark.

I also had to adjust my ISO far higher to capture more light (I tried everything from 400 to 1250) and slow down the shutter speed, anywhere from 1 second to 30. My problem was the tripod, which, when I tried pointing the camera almost vertically at the stars, could not hold still. Most of my star shots have some starlight squiggles (not the tech term). Ultimately, I captured a couple of semi-usable images of Jupiter and its moons. One is super tight, and you can just make out some moons very nearby.

Fly me to the moon

I had better luck with the moon. It's a lot closer than constellations and planets, and far brighter, which meant I could lower my ISO (for less grainy shots), close down my aperture a bit, and use faster shutter speeds (usually 125th of a second).

I did not have the benefit of a full moon during most of my shooting window, but the 3/4 moon I saw the other night was still impressive. As before, I put the shots in Lightroom and did some adjustments to reduce the exposure and increase the contrast. I didn't adjust sharpness, and therefore the level of detail I can show you here (yes, I did crop the image) is still impressive. Since I didn't travel with the Artemis II crew, my photos still only represent the side of the moon we all see. They used the Nikon Z9 to show us the dark side in ways we had never seen before (and to capture views of our home looming in the distance).

There's no frustration that I could not fully match Team Artemis II's photographic accomplishments. Instead, holding that camera in my hand, I imagined what it must have felt like to look through the viewfinder and see the weathered face of our Moon in such impressive, almost touchable detail.

For the briefest moment, I felt a kinship with those astronauts, and while I'm aware that the onboard iPhone 17 Pro Max also captured some inspiring images, I know there's no substitute for a pro camera and big lens when it comes to capturing unforgettable, lucky, and sometimes historic astro and nature images.

• Large-scale cardboard ENIAC replica recreates historic computer layout using thousands of handmade parts
• Teacher credits dyscalculia-driven spatial reasoning as key factor in engineering massive classroom builds
• Students construct nearly 300 square meters of cardboard structure matching original computer dimensions

A full-scale replica of one of the earliest programmable digital computers now fills a classroom space in Arizona, built almost entirely from cardboard and wood by students working under a teacher who credits his own dyscalculia (the math equivalent of dyslexia) for shaping how he engineers.

The life-size recreation of ENIAC (Electronic Numerical Integrator and Computer), widely regarded as the world’s first general-purpose programmable electronic computer, stretches across hundreds of square feet and mirrors the layout of the original machine that once weighed about 30 short tons.

Students at PS Academy in Arizona spent nearly six months assembling the structure, producing around 22,000 custom parts and joining them using roughly 1,600 hot glue sticks, according to I Programmer.

Experience, scale, and structure of ENIAC

Instead of steel cabinets packed with wiring and electronics, the replica uses layered cardboard panels with LED lighting to simulate the appearance of the original units.

Technology instructor Tom Burick said the project focused on recreating the physical experience of the machine rather than making it run.

"This project was never intended to recreate a working ENIAC. It was designed to recreate the experience, scale, and structure of ENIAC — so students and the public can understand what early computing really looked like and what it demanded of the people who built and operated it," he said.

Every major unit found in the original system, including accumulators, function tables, and the master programmer, appears in the correct physical position, matching historical layouts as closely as possible.

Students worked from original patent drawings, military documentation, and detailed photographs while also communicating with historians and museum staff to verify accuracy.

Burick’s path into teaching followed years spent building robots and running his own robotics company before turning toward education after the business closed during the financial crisis of the late 2000s, according to IEEE Spectrum.

He said his own dyscalculia shaped the way he approaches engineering challenges, forcing him to develop alternative ways of solving technical problems.

“People tell you what it takes, but they never tell you what it gives,” Burick said. “To me, [it] has always been a superpower.”

The completed replica includes 18,000 simulated vacuum tubes and multiple function tables arranged in the same U-shaped configuration used by the original machine.

By the time construction finished, the scale of the project left the classroom filled wall-to-wall with towering panels that recreate how early computing environments once appeared.

The ENIAC system itself was dismantled decades ago, leaving only scattered sections preserved in museums, which meant most people would never see the machine assembled in full form again.

Rebuilding it at full scale offered students a way to engage with computing history physically rather than through textbooks or diagrams alone.

IEEE Spectrum notes that previous projects led by Burick include an 8-foot-long drivable Cybertruck replica, and future plans could involve recreating hardware connected to NASA’s Artemis missions, something I'd love to see.

• Chinese national indicted in a huge phishing campaign
• NASA, the military, universities and private companies were targeted
• The perpetrator stole restricted defense software

NASA’s Office of Inspector General (OIG) has released a report detailing an intricate campaign to steal export-restricted software, orchestrated by a Chinese national.

The campaign targeted engineers at NASA, the Air Force, the Navy, the Army, and the Federal Aviation Administration, as well as major universities and private companies.

Posing as a colleague or friend, the perpetrator managed to convince dozens of professors, researchers, and engineers to share sensitive defense software and code for almost five years.

Federal warrant issued

The Chinese national, identified as Song Wu, was investigated by the Office of Investigations Cyber Crimes Division (CCD) after receiving an alert that emails were being received from a Gmail account claiming to be an aerospace professor who had collaborated with NASA.

Song Wu’s phishing campaign began in January 2017 and ended in December 2021, during which time he contacted employees at NASA, the Air Force, the Navy, the Army, the Federal Aviation Administration, and employees at numerous universities and private companies.

According to the CCD, Song Wu and unidentified co-conspirators would rigorously research their targets, and then pose as a friend or colleague requesting access to copies or source code of modeling software used for aerospace design and weapons development.

As the software was related to defense, it was subject to export controls that were unknowingly violated when victims fell for Song Wu’s deception.

A federal warrant has been issued for Song Wu’s arrest, and he has been indicted on 14 counts of wire fraud and 14 counts of aggravated identity theft, but he remains at large, the NASA report said.

“Cases like this underscore the importance of complying with export controls and staying vigilant, even during everyday email exchanges,” said Ryan Pittman, the Special Agent in Charge of CCD. “At NASA OIG, it’s our job to protect the Agency’s cutting-edge technology and expose cyber criminals who attempt to steal it.”

• Reuters has leaked details of SpaceX's plans ahead of its IPO
• They include "manufacturing our own GPUs"
• Whether that means chips in the current line of Tesla AI processors, or a new category of GPU or AI accelerator is unclear

SpaceX is apparently set to make its own GPUs, based on insights gleaned from its S-1 form ahead of the company's IPO, although these won't be products you'll be installing in your PC.

Tom's Hardware picked up the report by Reuters, with the news site claiming to have seen the S-1 form, which is filed by companies going public. SpaceX is expected to be worth in excess of $1 trillion – and probably closer to $2 trillion – when its shares are available on the stock market (next month, in theory).

Part of the form explains major spending plans to develop AI, which includes a mention of "manufacturing our own GPUs" in the listed 'substantial' capital expenditures.

As noted, these wouldn't be a rival for the likes of Nvidia's GeForce GPUs, but rather graphics boards built for AI tasks, commonly known as AI accelerators.

Nvidia makes those solutions, as well, of course – as does AMD – but SpaceX would seemingly rather have its own homegrown AI accelerators, because as noted in the S-1 filing, it does not have "long-term contracts ​with many of our direct chip ​suppliers". And clearly, that's going to be problematic given the current RAM crisis and its knock-on effects.

What isn't clear is whether SpaceX's mention of 'GPUs' in the filing could more broadly refer to Tesla's AI processors, which are currently on the AI4 generation, with the next-gen AI5 set to provide a big leap in performance (of a factor of 40x, it's claimed). It's possible that the company may be referring to this series of AI chips, rather than an entirely separate line of GPUs for AI tasks.

When it comes to making these chips, presumably the 'Terafab' project – a huge chip manufacturing center in Texas which involves SpaceX, xAI and Tesla – is going to be involved, even though GPUs haven't been mentioned directly in connection with this facility.

Analysis: AI5 or something else entirely, it won't be in your PC

At least to me, this does read more like a play for a new product, as opposed to a mention of existing AI chips and calling them 'GPUs' informally. That said, looking at the bigger picture, as Tom's points out, Elon Musk has referred to Tesla's AI5 chips as GPUs, and even noted that it "basically is a GPU" – and so you can see where the confusion comes in. All this could simply boil down to a liberal use of the term GPU, and mean precisely nothing.

At any rate, for the average consumer, the point to bear in mind is that whatever SpaceX is up to on the GPU front, if anything, it won't be relevant to the everyday computer user. This is all about silicon designed to drive AI performance specifically, and whether it's technically an AI chip in the existing line, or something different in terms of a new GPU accelerator designed for AI tasks, is all semantics – it won't be a graphics card as such (as in one designed to process 3D graphics).

Any confusion aside, the overall aim for Musk is clear enough, and that's to ensure AI chip supply by producing them himself, given the mentioned lack of 'long-term contracts' with chip makers.

How comforting this leak will be to would-be investors in SpaceX remains to be seen, but Musk appears to have some grand plans here, and not for the first time.

• NASA has released an enjoyable free tool for Earth Day
• It lets you spell your name using landscape features in satellite photos
• You can share your creations or download them and print them later

You’ve heard of writing your name in the stars, but what about writing your name on Planet Earth? That’s what's NASA’s latest tool is offering up, and it’s a fun way to celebrate Earth Day and demonstrates the variety of landscapes and features found on our blue planet.

Specifically, you’ll need to head over to NASA’s ‘Your Name in Landsat’ website and enter your name — or whatever word you want — into the box near the top of the page. Click the Enter button and you’ll see it spelled out, with each letter being represented by a letter hidden in a satellite photo.

For instance, try entering ‘TechRadar’ and you’ll get nine images lined up next to each other. The first could be a rocky formation that resembles the letter T, the second a swirling oceanic current that looks like an E, and so on.

The result is an entertaining little digital toy that is an enjoyable way to brighten someone’s day. You can share a link to your creation or download it, ready to be sent over email or maybe printed out as a postcard.

And if you mouse over any of the images used to spell your chosen word, you’ll see the location it was sourced from, as well as its longitude and latitude. That enables you to check it out on a site like Google Maps if you want to see the landscape in more detail.

An enjoyable distraction

NASA’s latest plaything is designed to coincide with Earth Day, which landed on April 22 this year. The event, which has been running since 1970, is dedicated to supporting environmental protection efforts around the world — you can read our series of articles celebrating it during Sustainability Week 2026.

It therefore makes sense for NASA to time its Your Name in Landsat tool to line up with this annual event. Its images show the vast variety of environments found across the globe and might help to inspire a touch of activism in those who use it.

According to NASA, “The satellite images used in this interactive [tool] are part of Landsat's extensive record, spanning more than 50 years.” They were sourced from NASA Earth Observatory, NASA Worldview, USGS EarthExplorer, and ESA Sentinel Hub, and you can download them individually if you like. Landsat is the longest-running satellite imagery program in the world and has been going strong since 1972.

Unfortunately, the downloaded images are not particularly large, so aren’t ideal for using as a poster or similarly-sized object. But they’re still a good way to explore the world and have a little fun while doing it.

• Star City, a new sci-fi series from the creators of Apple TV's For All Mankind, has dropped its first trailer
• Starring Rhys Ifans, Anna Maxwell Martin, Agnes O’Casey and Alice Englert
• Explores the space race from "behind the iron curtain, when the Soviet Union became the first nation to put a man on the moon"

The first trailer for Star City — a new Apple TV sci-fi series from the creators of For All Mankind — has arrived, giving us our first look at the "alt-history" of the space race. According to the streamer, the series will "explore the space race from behind the iron curtain, when the Soviet Union became the first nation to put a man on the moon, from May 29."

The cast is lead by Rhys Ifans, Anna Maxwell Martin, Agnes O’Casey and Alice Englert, and in the first trailer (which you can catch up with below) we see a traitor is in their midst, with an unknown source leaking military secrets out of their Star City base.

Compared to For All Mankind, Apple TV's alternative take on real history is a whole lot darker... and that's a good thing.

Star City on Apple TV turns the space race into a 'paranoid, propulsive thriller'

There's no doubting from the above trailer that Star City is technically fiction, but it's perhaps more linked to real-world events than we realize. For a start, the show's title alludes to the Yuri Gagarin Cosmonaut Training Center in Russia, which had the same nickname.

As Apple TV goes onto explain, “Star City is a propulsive paranoid thriller that takes us back to the key moment in the alt-history retelling of the space race — when the Soviet Union became the first nation to put a man on the moon.

"But this time, we explore the story from behind the Iron Curtain, showing the lives of the cosmonauts, the engineers, and the intelligence officers embedded among them in the Soviet space program, and the risks they all took to propel humankind forward."

While For All Mankind isn't exactly a bundle of joy when it comes to imagining the inner workings of space science (especially if you've been watching season 5), the cold color grading and eerie visuals in the above Star City trailer indicate something much more sinister is afoot.

Either way, I'm sat — not least because I know from Line of Duty and Steal how incredibly well Anna Maxwell Martin plays an vicious, aloof villain.

The eight-episode series will debut globally on Apple TV with two episodes on May 29, followed by one new episode every week through July 10.

Fusion propulsion has been proposed for decades as a highly efficient way to move spacecraft faster across the solar system without the need for massive amounts of fuel. Making that concept a reality has, however, proved difficult, especially when it comes to controlling plasma and directing it into usable thrust.

Now a UK-based propulsion developer believes it has taken an early step toward achieving that goal.

At Amazon’s MARS Conference in California, Pulsar Fusion demonstrated “first plasma” in its Sunbird exhaust test system, with the experiment conducted in Bletchley, UK, and streamed live to the stage.

The test showed plasma being confined and guided through the exhaust architecture, offering a first look at how a fusion-powered propulsion system might work outside of theory and simulation.

A reusable space tug in orbit

Chemical rockets provide strong bursts of thrust but burn through large amounts of propellant, while electric systems stretch fuel efficiency but accelerate slowly over long periods.

Fusion propulsion is attractive because it offers the possibility of combining high thrust with extremely high exhaust velocity, something current propulsion systems cannot deliver at the same time.

Pulsar’s Sunbird concept takes aim at that trade-off by acting as a reusable space tug operating in orbit.

Instead of launching spacecraft with large amounts of chemical propellant, a lighter vehicle could dock with the tug and receive the push needed to reach destinations such as the Moon.

Internal modelling suggests that approach could reduce in-space propellant requirements by more than 90% for missions similar to Artemis II.

The recent plasma milestone is still an early-stage result, but it provides a clearer view of how fusion propulsion hardware could evolve from laboratory experiments into practical spacecraft systems.

I wanted to find out more, so spoke with Pulsar Fusion CEO Richard Dinan about the technical hurdles ahead, the economics of fusion propulsion, and how systems like Sunbird could reshape future missions.

• What breakthroughs has Pulsar Fusion achieved in order to deliver this first successful demonstration of plasma control?

The breakthrough here was not a single moment, but the successful integration of several difficult elements into one working system.

We were able to generate plasma within the exhaust structure and demonstrate that it could be guided and shaped using electromagnetic fields in a propulsion-relevant architecture.

That matters because controlling plasma is fundamental. Before you can talk seriously about higher-energy systems, you first have to show that you can create the flow, contain it, and direct it in a stable and measurable way.

This demonstration was an early but important step in proving that the architecture behaves as intended.

That matters enormously in space, because every kilogram of propellant launched from Earth is expensive. A high-Specific Impulse propulsion system can dramatically reduce the amount of fuel that must be carried for deep-space maneuvers.

• Specific Impulse will be a new concept for our audience. Can you describe why this metric is so important and how an Artemis mission based on plasma technology would differ in terms of fuel carried, risks, and CAPEX/OPEX?

In an Artemis-type architecture, that could mean a very different mission design. Instead of relying entirely on chemical propulsion launched from Earth in one go, you could imagine lighter spacecraft being delivered to orbit and then transferred onward by highly efficient in-space propulsion systems.

That changes the fuel equation, reduces dependence on brute force launch mass, and opens the door to reusable orbital infrastructure.

This ultimately means lighter launch craft, and therefore less fuel and less weight, resulting in faster, cheaper, more efficient missions.

• You argue that a shift away from single-launch delivery mechanisms to a reusable (and therefore cheaper) model lies in our common future. Can you tell us more about your vision and what sort of price point you predict, given that Orbital predicts $10/kg?

Our view is that the future of space will look less like single-use expeditionary missions and more like a transport network.

Today, most missions still depend on launching everything needed for the mission in one stack, from one gravity well, at one moment in time. That is expensive and inherently inefficient.

We think the long-term future is an orbital economy built around reusable infrastructure: vehicles that move cargo, fuel, and eventually strategic assets between orbits and beyond, without being discarded after one mission.

That does not just reduce cost, it changes what becomes commercially possible. Once mobility in space becomes reusable, you can start treating propulsion as an operational service rather than a disposable hardware event.

• One of the facts you mentioned surprised me. Sunbird delivers megawatt-scale onboard power. How does that happen? What sort of energy ranges (in MWh) are we looking at, and what use cases does it unlock?

It's an important point that a fusion-based propulsion architecture is not only interesting because of thrust. It is also interesting because, in principle, it is tied to very large energy flows. That means a mature system could potentially support both propulsion and substantial onboard power availability.

At this stage, however, we would be careful to distinguish between long-term system potential and what is being demonstrated today.

The present work is about validating elements of plasma control and exhaust architecture, not delivering a flight-ready power platform.

In the long run, if such systems mature as hoped, megawatt-class onboard power could unlock a very different category of spacecraft capability, including deep-space operations, power-hungry payloads, and entirely new classes of orbital and interplanetary missions.

• How did Pulsar Fusion achieve what looks like a major breakthrough in propulsion while titans (SpaceX, NASA) haven't yet cracked it? What is the secret sauce, and how do you plan to keep abreast of the competition?

I would be careful with the framing. We are not claiming to have solved fusion propulsion in its entirety, and there is still a great deal of work ahead.

What we have done is move quickly and show early progress in a specific architecture that we believe has long-term promise.

Large organizations often have different incentives. They may be focused on near-term operational systems, legacy architectures, or broader institutional priorities. Smaller companies can sometimes move faster in frontier areas because they are willing to concentrate their effort on a narrower technical pathway.

• You mentioned nuclear fusion space propulsion in your March 2026 press release. How is that different from the "holy grail" represented by the tokamak?

A tokamak is designed first and foremost as a terrestrial fusion power machine. Its job is to confine plasma for long enough, at sufficient temperature and pressure, to produce useful net energy in a stationary power-generation context.

A space propulsion system has a very different objective. It is not simply trying to replicate a grid-scale fusion plant in orbit. It is trying to create a system in which high-energy plasma can ultimately be used for propulsion in a mass-constrained, mission-driven environment.

So while both involve plasma physics and fusion-related principles, the design logic is very different.

One is optimized for stable power production on Earth; the other is being pursued as a propulsion architecture for space, where efficiency, mass, controllability, and exhaust behavior are central.

• Largest 3D map built using 47 million galaxies and quasars
• Thousands of robotic fiber arms captured light from distant galaxies
• New data could reshape understanding of dark energy behavior

Astronomers have completed the largest high-resolution 3D map of the universe ever recorded, after a five-year observing campaign that tracked tens of millions of galaxies.

The project, known as the Dark Energy Spectroscopic Instrument, or DESI, focuses on understanding dark energy, the mysterious force driving the universe’s accelerating expansion.

The survey exceeded its original targets, recording data from more than 47 million galaxies and quasars instead of the planned 34 million. Researchers also gathered observations from more than 20 million nearby stars to study the structure of the Milky Way.

Tiny robot arms

At the center of the project is a system of 5,000 robotic fiber-optic positioners mounted on a telescope in Arizona. These tiny robot arms move into position roughly every 20 minutes, aligning optical fibers to capture faint light from distant galaxies.

That light is then fed into spectrographs that split it into its component colors, allowing scientists to calculate how far away each galaxy lies from Earth. By combining distance measurements with sky positions, the system builds a layered 3D map showing how matter is spread across the universe.

The telescope used for the project is the 4-meter Nicholas U. Mayall Telescope at Kitt Peak National Observatory. Engineers replaced its original camera with DESI’s fiber-based system, allowing it to measure thousands of galaxies at once.

DESI was built to study how galaxies cluster across different distances and times. Those patterns act as markers of how quickly the universe expanded in the past, revealing how dark energy influenced cosmic growth over billions of years.

Earlier results from the project suggested that dark energy might not behave as a steady force. Instead of remaining constant, early data hinted that its influence could change over time, although researchers caution that additional data could still alter that conclusion.

Completing the planned map doesn't mark the end of the project. Scientists plan to cover additional regions of the sky and to capture more distant galaxies.

Future observations will also revisit existing areas to gather denser data and improve the precision of recorded measurements to help researchers test whether earlier hints about changing dark energy persist in the larger dataset.

Via Science.org

• Blue Origin reused a rocket booster on its latest mission
• This was the first time it has done this, marking a major milestone
• However, the mission was a failure as it didn't place its satellite payload high enough

Over the weekend Jeff Bezos’ Blue Origin space company hit a major milestone by successfully re-using a rocket booster for the first time — deploying the same ‘Never Tell Me The Odds’ rocket in Sunday’s NG-3 mission that it used back in November for NG-2.

This should be a time for celebration as Blue Origin proves it can better compete with SpaceX on sustainability and space travel prowess, but it isn’t. Why? Because NG-3’s primary objective was an utter failure.

Beyond a rocket booster reusability test, Blue Origin's primary goal for this flight was to drop an AST SpaceMobile communications satellite – BlueBird 7 — into orbit. It did technically do this, but NG-3 placed the satellite “lower than planned” as AST phrased it.

Per an AST SpaceMobile statement, “While the satellite separated from the launch vehicle and powered on, the altitude is too low to sustain operations with its on-board thruster technology and will [be] de-orbited. The cost of the satellite is expected to be recovered under the company’s insurance policy.”

This means the satellite will be repositioned and left to burn up in the Earth’s atmosphere.

It IS rocket science

For AST SpaceMobile this failure isn’t world-ending — it says it still has plans to launch 45 more satellites before 2026 ends — but for Blue Origin this failure will leave an embarrassing stain on what should have otherwise been a momentous flight.

The silver lining is, if this mission had to be a failure, at least Blue Origin messed up with a satellite launch and not the first launch of its lunar lander — which was originally meant to be NG-3’s purpose.

The next NASA Artemis mission is expected to drop people onto the moon for the first time in over 50 years, but it still needs a lander. SpaceX and Blue Origin are currently racing to design and test the craft that will take astronauts from their spaceship to the lunar surface, but neither has completed the project yet.

Because the Artemis missions are intended to establish a permanent lunar base, the new lander approach can’t be a repeat of the one-and-done style previous missions to the moon — it needs to be something that can repeatedly shuttle crew and cargo.

There’s also a lot less room for failure when carrying crew compared with cargo. Things can be replaced, people’s lives cannot.

We’ll have to wait and see how this mistake affects Blue Origin’s lunar lander bid. Rocket science is famously difficult, so hopefully this will simply be a learning experience for the company.

• NASA astronaut Reid Wiseman has filmed an 'Earthset' on an iPhone 17 Pro Max
• The footage shows the Earth 'setting' behind the moon
• It's the first time we've seen video of this and it taken using the 8x zoom lens

It’s not every day that you see footage of the Earth setting behind the moon. But today NASA astronaut Reid Wiseman released the first video taken of that event in human history — and it was shot on an iPhone 17 Pro Max.

Posting on X, Wiseman shared a 53-second video shot through Integrity’s docking hatch window during the Artemis II mission, showing the Earth slowly disappear behind the moon.

It’s a stunning clip, and all the more impressive for having been shot on an iPhone — in fact, it would make a great advert for Apple’s phones.

A once in a lifetime scene

Wiseman claims the footage is uncropped and uncut, with “8x zoom which is quite comparable to the view of the human eye.” We’re slightly confused by that claim as 8x zoom on this phone should be around a 200mm equivalent focal length, whereas the human eye is usually considered to be between 40-50mm.

Still, this is no time to get bogged down in technicalities — Wiseman described the scene as something that’s “only one chance in this lifetime” and “like watching sunset at the beach from the most foreign seat in the cosmos.” The astrophotographer Andrew McCarthy, meanwhile, said it was "quite possibly the most incredible video ever captured by a phone. Bravo."

While this is among the more impressive scenes to come from the Artemis II mission, we’ve seen plenty of other shots and videos too — including some shot on iPhone.

Sadly, with the Artemis II mission now concluded there might not be too many more iPhone photos and videos to see from it, but it makes us all the more excited for what’s to come on future NASA launches.

• Blue Origin has reused one of its rockets for the first time
• 'Never Tell Me The Odds' was part of the New Glenn 3 mission
• The mission launched a mobile connectivity satellite

We've got another space first to report today: Blue Origin has successfully reused a rocket booster for the first time, managing to get the 'Never Tell Me The Odds' rocket back to its landing pad in the Atlantic Ocean as part of the third New Glenn (NG-3) mission.

The rocket booster was first used with the NG-2 mission last November, and having recovered it that time, Blue Origin has now shown it can be deployed again. It means the company can now better compete with SpaceX when it comes to these reusable rockets.

Liftoff happened at 7.25am ET (4.25am PT / 12.25pm BST / 9.25pm AEST), and the 'Never Tell Me The Odds' booster arrived back on its floating Jacklyn droneship around 10 minutes later, after finishing the job of pushing its payload into orbit.

And that payload is also notable: it's the AST SpaceMobile BlueBird 7 satellite, which is effectively going to act as a giant cell tower in space (it boasts a massive antenna measuring 2,400 square feet or 223 square meters). It gets us a big step closer to global satellite coverage for modern smartphones.

A major satellite expansion

As with reusable rockets, Amazon and Blue Origin are competing with SpaceX and Starlink when it comes to satellite communications. SpaceX is hoping to begin the process of launching thousands of its Starlink V2 satellites sometime next year.

The idea is that our phones will be able to connect to them without any hardware or software upgrades — these satellites will be smart enough and powerful enough to mimic a regular cell tower, with 4G and 5G speeds, from all the way up in orbit.

It'll take years before the space satellite network is ready to go beyond emergency communications — and pricing and carrier partnerships will need to be worked out along the way — but BlueBird 7 is a sign of the direction we're heading in.

Reaction to the launch on social media has been overwhelmingly positive and enthusiastic (perhaps buoyed by the recent Artemis II mission), with livestream viewers describing the experience as "a wonderful show for any rocket enthusiast" and simply "amazing". You can read more about the New Glenn missions on the Blue Origin website.

Only a very lucky few people ever get to travel into space, but more than 5.6 million names just completed a journey around the Moon, stored on a microSD card carried aboard NASA’s Artemis II mission.

That flight sent four astronauts farther from Earth than humans have ever traveled, completing a roughly 10-day mission that orbited the Moon before safely returning to Earth.

The trip tested critical systems for future deep-space exploration and marked a major milestone for the Artemis program which will ultimately return humans to the surface of the Moon in 2028.

Today's top memory card deal

The microSD card, carrying 5,647,889 submitted names, was zipped into Rise, the mission’s mascot, a cartoonish Moon wearing a cap covered in stars. The mascot itself was designed by a year three student from California, called Lucas Ye, whose artwork was selected from more than 2,600 entries from over 50 countries.

While the specific card zipped inside Rise was certified for spaceflight conditions, it traces its lineage to the SanDisk Ultra series used by people here on Earth inside cameras, handheld devices, and portable recording gear.

The consumer version of the SanDisk Ultra microSD series comes in 16 and 32GB capacities and supports both microSD and microSDHC formats, making it compatible with a wide range of devices used for everyday recording and storage tasks.

Rated at Class 10 speeds, the card supports read speeds of up to 80MB/s, which suits Full HD video capture, burst photography, and quick file transfers.

It's durable, with protection against water, temperature extremes, and X-ray exposure, and maybe (but we wouldn't bet on it), trips around the Moon.

After splashdown, the mission mascot didn’t stay tucked away inside the spacecraft as it should have done according to NASA's mission rules, Artemis II commander Reid Wiseman later revealed on X.

“I was supposed to leave Rise in Integrity….but that was not something I was going to do. I stuffed that little guy in a dry bag we had in our survival kit and hooked the bag onto my pressure suit,” he wrote.

• Shop all SD cards at Amazon

• Artemis II generates data volumes that old systems cannot handle efficiently
• Laser communications transmit far more data than traditional radio systems
• Infrared light enables high-speed space communication across vast distances

The sheer volume of data generated during modern lunar missions has rendered old radio systems nearly obsolete.

Artemis II was expected to produce somewhere between 300GB and over 400GB of high-resolution imagery and telemetry by the mission’s end.

By comparison, the Apollo 13 mission operated with a fraction of that capacity, and the difference is not just incremental — it’s a fundamental overhaul in how spacecraft talk to Earth.

The engineering shift that made the leap possible

Traditional radio frequencies could not move that much data quickly enough, so engineers turned to an entirely different method: laser communications.

Laser communications rely on invisible infrared light, which travels at the same speed as radio waves but carries far more information.

Because infrared light has a higher frequency, it can pack more data into each transmission - and the Orion Artemis II Optical Communications System (O2O) demonstrated the ability to downlink over 100GB of data.

This system could move roughly 36GB in a single hour, outpacing traditional radio systems on the S-band, which could manage only about 7GB per day.

NASA noted, “More data means more discoveries,” although the practical benefits for crew safety and real-time decision-making remain to be fully proven.

However, this system came with its earthly limitations, and any weather disruption could interrupt the flow of information.

Ground station telescopes at NASA’s White Sands Complex in New Mexico and the Table Mountain Facility in California had to operate in high, dry environments with minimal cloud cover to maintain a strong laser link.

Still, the O2O terminal — comprising a 4-inch telescope, two gimbals, a modem, and a controller — passed multi-day readiness reviews.

A NASA official described the achievement as “an impressive leap” forward, yet the system was not used on Artemis III, raising questions about the pace of adoption.

While a 100,000-fold improvement over Apollo 13 sounds extraordinary, the comparison deserves scrutiny.

Apollo 13’s radio systems were designed in the 1960s, and modern radio technology has also improved considerably.

The real test will be whether laser communications prove reliable over deep-space distances without frequent ground-station interventions.

The Australian National University attempted to receive O2O’s laser links using affordable commercial components — a demonstration that could validate or undermine claims of scalability.

For now, the numbers are impressive, but space history is littered with promising technologies that struggled outside controlled conditions.

• The Orion spacecraft uses eight processors running identical instructions simultaneously
• A fail-safe design prevents faulty computers from sending incorrect commands
• Triple redundant memory corrects single-bit errors automatically on access

The NASA Artemis II mission relies on a computing system built to remain operational under extreme conditions and hardware faults.

Unlike the Apollo program, where onboard computers handled limited functions, the Orion spacecraft manages life support, navigation, and communication through integrated flight software.

The Orion capsule carries one of the most fault-tolerant computer systems ever built for spaceflight, operating 250,000 miles from Earth, where no repairs are possible.

From Apollo's limits to Orion's full system control

Apollo astronauts relied on a 1MHz computer with just 4 kilobytes of memory, but today's spacecrafts need much more, considering the distance.

The Orion spacecraft uses two vehicle management computers, each containing two flight control modules.

Each module consists of a pair of processors that continuously check each other's outputs, resulting in 8 processors executing the same instructions simultaneously.

If a processor produces an incorrect result, the paired design detects the mismatch immediately and prevents the output from being used.

"We still architect to cover for hardware failures," said Nate Uitenbroek, Software Integration and Verification Lead in NASA's Orion Program.

"Along with physically redundant wires, we have logically redundant network planes. We have redundant flight computers."

Instead of relying on majority voting, the system selects outputs from available modules based on a defined priority order.

The system is designed to tolerate rapid failures during flight. Uitenbroek stated, "We can lose three FCMs in 22 seconds and still ride through safely on the last FCM… A faulty computer will fail silently, rather than transmit the wrong answer."

Failed modules are reset and re-synchronized, allowing them to rejoin the system during the mission.

Orion uses a time-triggered Ethernet network that distributes a shared time reference throughout the system - so if a module fails to meet its execution deadline, it is automatically isolated, reset, and re-synchronized before returning to operation.

The computing system includes triple-redundant memory capable of correcting single-bit errors during every read operation.

Network interfaces use dual communication lanes that are continuously compared to detect inconsistencies, while the overall network is replicated across three independent planes.

Orion carries a separate Backup Flight Software system that operates on different hardware and software, running continuously in the background.

"It is intentionally different to ensure that a common mode software failure in the primary flight software isn't also implemented incorrectly on the backup," Uitenbroek said.

The spacecraft also includes procedures for full power loss scenarios, allowing systems to restart, stabilize, and re-establish communication once power is restored.

The system is overengineered by any commercial standard, but deep space offers no second chances.

Whether all 8 processors will perform as designed under real radiation conditions remains untested, and the backup software has never faced an actual emergency.

Still, for a mission where the nearest hardware store is 250,000 miles away, this architecture makes a brutal kind of sense.

Via Communications of the ACM

Having successfully looped around the Moon and returned safely to Earth, taking its four man crew farther from home than any humans before, Artemis II's mission is now complete.

During that historic journey, iconic Earth images were stored on ProGrade Digital CFexpress Type B Iridium memory cards — including the same 400GB version that's currently available to buy on Amazon for $459.99.

That connection between a real lunar mission and a product anyone can buy makes this one of the cooler, more unusual pieces of camera gear I’ve come across. It isn’t tied to simulations or lab claims, but to hardware trusted during an actual crewed flight beyond the Moon.

Today's top memory card deal

ProGrade Digital entered into a Space Act agreement with NASA to support Artemis missions using its CFexpress Type B Iridium cards.

The hardware was chosen from several commercial options after testing showed it capable of handling harsh conditions similar to those found in space.

“The lunar surface presents very challenging environments for cameras and related storage media,” said Jeremy Myers, HULC Handheld Universal Lunar Camera COTS hardware lead. “The partnership developed between NASA and ProGrade will help ensure images and video are captured reliably when humanity returns to the Moon”.

“Through diligent teamwork between the ProGrade San Jose engineering team, NASA’s Space Systems Department at the Marshall Space Flight Center, and key memory technology supplier – Phison Electronics – we were able to go from sample submission to hardware delivery in just under 18 months," said Wes Brewer, founder, and CEO of ProGrade Digital.

"During qualification, our cards were placed into the extreme environmental conditions uniquely encountered in outer space, over many hours of testing to maximize the reliable capture and playback of still images and video in the specially designed Nikon Z9 camera was possible."

"Perhaps the best part of this accomplishment is for our everyday customers, who can now buy memory cards that will be used in the Handheld Universal Lunar Camera during future Artemis missions.”

The same card family is part of the Space Act agreement supporting the Handheld Universal Lunar Camera, or HULC, system planned to land on the lunar surface as part of NASA’s Artemis III mission.

This 400GB card supports read speeds up to 3550MB/s and sustained write speeds up to 850MB/s, giving it the bandwidth needed for large RAW files and high-resolution video. That speed will perfectly suits photographers (on or off Earth) working with 4K or 8K footage.

Compatibility with CFexpress Type B cameras and backward support for XQD devices makes it useful across a wide range of professional camera bodies.

Owning the same storage hardware used to capture real images during humanity’s latest journey around the Moon is a rare crossover between space exploration and everyday photography equipment. It's a cool thing to own, even if the photographs and videos you store on it are likely to be of a more down to Earth nature.

• Shop all SD cards at Amazon

The race to build ever larger AI models has created an unexpected bottleneck. It is no longer chips that limit progress, but power. Modern data centers already consume vast amounts of electricity, and demand is rising faster than infrastructure can keep up.

Permitting new facilities is becoming harder, communities are pushing back, and the cost of cooling thousands of GPUs continues to climb. The result is an energy ceiling that threatens to slow the pace of AI development.

Orbital believes the solution lies far above the grid.

The company is developing AI data centers designed to operate in low Earth orbit, powered entirely by solar energy and cooled by radiating heat directly into space. Without weather, night cycles, or grid dependency, solar arrays in orbit can generate continuous power, while the vacuum of space provides a natural way to dissipate heat — two constraints that dominate the economics of terrestrial data centers.

Orbital-1

Backed by funding from a16z Speedrun, Orbital is preparing its first test mission, Orbital-1, scheduled to launch on a SpaceX Falcon 9 in April 2027.

The satellite will host Nvidia-powered compute hardware and is intended to validate sustained GPU operation in orbit, test radiation resilience, and begin running AI inference workloads once initial validation is complete.

The company chose inference over training because inference tasks can run independently across a satellite network. Unlike training clusters, which require tightly coupled GPUs operating in near-perfect synchronization, inference tasks can be distributed across many independent nodes — making them better suited to satellite constellations.

The idea of running data centers in space may sound radical, but the pressures driving it are becoming increasingly real.

I wanted to know more so I spoke to Orbital founder Euwyn Poon about the economics, engineering challenges, and practical realities of running AI infrastructure in orbit.

• Why is everyone suddenly looking at space as the new frontier for AI infrastructure? What has changed? (cratering cost of putting goods in orbit, the OPEX appeal, exploding AI demand, the lack of red tape in space?)

A few things converged at once. Launch costs are going to collapse with Starship, going from $7,000/kg on the Falcon 9 today towards the target of $10/kg.

Meanwhile, AI demand is skyrocketing beyond the capacity of the power grid, with US data centers using about 25 gigawatts today and growing 3 to 4 times by 2030. Building gigawatt data centers in the US is also becoming increasingly difficult.

Communities across the US are pushing back against data center construction. Getting a gigawatt facility permitted and connected to the grid now involves a ton of uncertainty and risk.

• More specifically on the energy infrastructure, can you give us a sense of how much a MwH of power would cost in space as opposed to Earth-based (or perhaps how easily would it be to get solar power up there).

On Earth, data center electricity runs $60 to $100 per megawatt-hour before cooling. Add the 40% thermal overhead that every facility carries and you land at $85 to $140 per MWh in effective energy cost.

In orbit, the marginal cost of energy is zero. The sun delivers 1,361 watts per square meter in LEO, constantly, with no fuel cost, no grid fees, no utility contract.

What you actually pay is the amortized capital cost of the solar array and launch costs. Our ultimate goal is to push amortized space power below $10 per MWh.

• What could prevent rogue nations from disabling/harming/taking down these orbital data center networks in the sky à la Moonraker? What sort of resilience features are being implemented to prevent that from happening?

The short answer is that distributed constellations are inherently harder to attack than buildings on the ground. We are not building one big space station. We are deploying thousands of small, independent satellites spread across multiple orbital planes. Taking out a single satellite removes a fraction of a percent of total capacity.

To meaningfully degrade the network, you would need to disable hundreds or thousands of targets simultaneously. That is an enormously expensive and conspicuous military operation.

On the legal side, the Outer Space Treaty of 1967 establishes that no nation can claim sovereignty over space. Satellites remain under the jurisdiction of the launching country, but no nation can seize or shut down another country's satellite without committing an act of war.

Every terrestrial data center can be raided with a warrant. Orbital infrastructure cannot.

• How does one deal with the extreme heat differences in space where I assume one side of the satellite will remain immensely hot and the part in the shadows, a few Kelvins above absolute zero (dissipation of compute heat but also thermal management).

The thermal environment in space is extreme. On the sun-facing side, solar irradiance hits at 1,361 watts per square meter. On the shadow side, space is roughly 3 Kelvin. Inside the satellite, you have GPU waste heat with no air for convection.

On Earth, you blow air over heat sinks or run water through cooling towers. In vacuum, none of that works. Managing this thermal environment is one of our core engineering efforts that we are not yet discussing publicly in detail.

• Given your focus on inference, how do you plan to deal with signal latency for real-time AI applications? Would latency be inversely proportional to the number of satellites deployed?

At 550 kilometers altitude, the speed-of-light round trip is under 4 milliseconds. A typical API call to a terrestrial cloud provider already takes 20 to 100 milliseconds. For inference workloads like chatbots, code generation, and agentic AI, users already wait hundreds of milliseconds for a response. The orbital penalty is imperceptible.

More satellites helps, but it is not inversely proportional. For inference, these are independent, parallel requests, not a single job being split across multiple nodes.

• Your satellites are going to be launched using Space X's platform. Space X has already disclosed its ambitious to be a major player in YOUR field. How do you plan to tackle this David vs Goliath conundrum?

This is a big space, and we are in different lanes. Elon merged SpaceX with xAI, and that validates the entire thesis. The biggest startup risk is that the category does not exist. SpaceX just eliminated that risk for us.

Elon has announced he will focus on its own custom silicon and its own models. That creates an opening for other operators like us. Just as there are many hyperscalers on earth, there is room for many winners here.

• What happens if the satellites reach EOL? Will they be decommissioned or sent to burn in the Earth atmosphere or actually never decommissioned? What's the rough lifespan you expect to eek out from them?

Every satellite deorbits and burns up in the atmosphere. Each satellite carries ion thrusters with propellant reserved specifically for the deorbit maneuver. The structure is made from materials that burn up cleanly on re-entry.

Design life is roughly matched to the useful life of the GPU architecture on board, around 5 to 7 years. By that point the chips are multiple generations behind and the economics favor launching fresh satellites with current hardware rather than maintaining older ones.

• What else is on Orbital's roadmap? Autonomous robots? Strategic partnership with hyperscalers? Different class/size of orbital data centers — e.g. SSO for training/LEO for inference? Self repairing ones?

Right now we are focused on two things: our test mission and our factory. Orbital-1 launches in early 2027 In parallel, we are standing up Factory-1, a robotic satellite assembly facility in Los Angeles.