• Codasip sale fuels speculation about GlobalFoundries expanding RISC-V processor capabilities
• Studio licensing strengthens case for customized silicon alongside manufacturing services
• Growing RISC-V portfolio raises possibility of future competition with Arm and AMD

Codasip, a Munich and Bristol-based semiconductor design company that develops RISC-V processor cores, is selling part of its business to an unnamed public U.S. semiconductor company, with speculation that the buyer could be none other than GlobalFoundries, AMD's former foundry arm.

Codasip confirmed the company will divest its low-end RISC-V processor design unit while directing future work toward security-focused chip architectures and system-level products.

That lower-end division includes processor cores used in embedded electronics, automotive systems, and industrial hardware, where efficient and customizable designs are widely used.

A three-part structure

GlobalFoundries is seen as the most likely buyer because it has spent the past year acquiring RISC-V processor technology through multiple deals.

The company already owns MIPS and has agreed to acquire ARC-V processor intellectual property from Synopsys, expanding its reach into higher-performance and specialized processor designs.

Adding Codasip’s entry-level RISC-V cores would extend that portfolio into simpler embedded designs, creating coverage across both low-power and more advanced processor tiers.

The agreement also includes a broad license to Studio, Codasip’s processor development software that allows customization of processor instructions.

That capability allows GlobalFoundries, which began as AMD’s manufacturing arm before becoming an independent semiconductor manufacturer, to support tailored chip development, giving customers the ability to modify processor behavior instead of relying only on fixed designs licensed from companies such as Arm.

Arm has long dominated embedded processor markets through royalty-based licensing, collecting fees from companies that build chips using its architecture.

A manufacturer that offers both fabrication services and tailored processor designs creates an alternative path, especially for companies developing automotive, industrial, and edge computing systems.

Continued expansion into processor technology could eventually place GlobalFoundries into competition with established chip designers.

Codasip described the sale as part of a broader change in direction toward what it calls cyber-resilient semiconductor architectures.

Its future development work centers on CHERI, short for Capability Hardware Enhanced RISC Instructions, a technology focused on limiting software exploits by enforcing stricter memory access rules directly in hardware.

“Cyber-resilience has become a strategic requirement for governments, infrastructure operators, and technology providers worldwide,” said Ron Black, Chief Executive Officer of Codasip.

“Traditional approaches inefficiently bolt security onto systems after the fact. Our focus is on enabling partners to build security into the fundamental architecture of compute systems from the beginning.”

The transaction is expected to close within about a month, and we'll know then whether GlobalFoundries is indeed the buyer of Codasip's low-end RISC-V processor design unit.

• Arm enters silicon production with a CPU designed for large-scale AI workloads
• New AGI CPU doubles rack performance compared with traditional x86 systems
• Meta and OpenAI adopt Arm chip for next-generation infrastructure

Arm has extended its compute platform into production silicon for the first time with the introduction of what it calls the “next evolution of the Arm compute platform,” the AGI CPU.

The companys says the CPU is designed specifically for AI data centers, supporting agentic AI workloads which involve continuously running agents capable of reasoning, planning, and acting.

The processor features up to 136 Neoverse V3 cores per CPU, with 6GB/s memory bandwidth per core and sub-100ns latency, allowing higher workload density and improved system efficiency.

Performance and capacity

The Arm AGI CPU promises deterministic performance under sustained load with a 300-watt TDP and a dedicated core per program thread.

The processor supports air-cooled 1U server chassis with up to 8,160 cores per rack, and liquid-cooled deployments reaching 45,000 cores per rack.

Compared with x86 CPUs, the Arm AGI CPU can provide more than double the performance per rack, supporting larger AI workloads while remaining energy efficient.

These capabilities aim to improve compute density, accelerator utilization, and overall infrastructure efficiency.

Meta serves as the lead partner and co-developer of the Arm AGI CPU, integrating it with its Meta Training and Inference Accelerator (MTIA) to optimize data center performance.

Early commercial adoption also includes the likes of OpenAI, Cerebras, Cloudflare, Positron, Rebellions, SAP, and SK Telecom.

Arm is collaborating with OEMs and ODMs such as Lenovo, Supermicro, Quanta Computer, and ASRock Rack to deliver early systems, with broader availability expected in the second half of 2026.

More than 50 industry leaders across hyperscale, cloud, semiconductor, memory, networking, software, and system design sectors support the CPU’s rollout.

“Over the last decade, we’ve partnered closely with Arm in building Graviton here at AWS, and it’s been a remarkable success — the majority of compute capacity AWS added to our fleet in 2025 was powered by Graviton,” said James Hamilton, SVP and Distinguished Engineer, Amazon.

“This collaboration has been great for both companies, and Graviton continues to deliver better price/performance for our customers.”

Industry partners also pointed to the broader infrastructure implications of the new CPU.

“The new Arm AGI CPU will further unlock the Arm ecosystem for a broad range of customers, creating new opportunities for everyone...” said Charlie Kawwas, President, Semiconductor Solutions Group, Broadcom Inc.

“As Broadcom builds the world’s most capable XPU and networking solutions for hyperscalers…our partnership with Arm has enabled us to move with unmatched intent and speed.”

The Arm AGI CPU is intended to serve as a foundation for agentic AI workloads, enabling organizations to deploy AI tools at scale while maintaining high efficiency.

The processor supports large-scale deployment of AI applications, including accelerator management, control plane processing, and cloud- or enterprise-based API and task hosting.

That said, the Arm AGI CPU's success will depend on data center adoption, integration with existing accelerators and memory, and proven performance gains over alternatives.

• Windows 11 26H1 will be for Snapdragon X2 laptops only
• Those Snapdragon X2 PCs won't get the 26H2 update, which will be kept separate for other Windows 11 devices
• This is seemingly part of Microsoft's plan to avoid the scenario seen with the 24H2 update, where changes for Arm PCs caused chaos on other devices

We already knew Windows 11's 26H1 update wouldn't be coming to most PCs – and was for Arm laptops only – but Microsoft has now confirmed that it's just for Snapdragon X2 devices, while providing some more details on the upgrade.

This includes the fact that these Arm PCs – which aren't on sale just yet, but are imminent – won't receive the 26H2 update later this year.

VideoCardz noticed Microsoft's confirmation in a blog post which stated: "Windows 11, version 26H1 is a targeted release that supports some of the new device innovations coming in 2026.

"That means that this release is not being made available through broad channels but is only intended for those who purchase these new devices. At this time, devices with Qualcomm Snapdragon X2 Series processors will come with Windows 11, version 26H1."

Microsoft clarifies that 26H1 is based on a different underlying core codebase than the current version of Windows 11, which is 25H2, and that as a result, 25H2 users won't get an upgrade to 26H1.

In the same vein, those who buy a Snapdragon X2 laptop this year with 26H1 on board will not get the 26H2 update later on – this will be for 25H2 PCs only.

Analysis: a somewhat confusing twist, but a prudent one

This means Microsoft is keeping existing Windows 11 PCs on a completely different track for new OS builds to the new incoming Arm-based Snapdragon X2 laptops. These separate builds will not be merged in the update for later in the year (26H2).

This is quite different to the tack Microsoft took with the original Snapdragon X PCs, where a version of Windows 11 24H2 was floated early, specifically for those laptops only, but later became the full 24H2 update for all machines.

If you recall, that famously led to 24H2 being seriously buggy, and a real headache for non-Arm PCs, and so this is what has presumably led to Microsoft's decision here. It's keeping the new underlying platform Windows 11 needs for Snapdragon X2 – this is called 'Bromine' – completely separate on 26H1, while the current 'Germanium' platform, as seen with Windows 11 25H2 (and first introduced with 24H2), will be where the 26H2 update lands.

Given that, you might be wondering: when will Windows 11 26H1 Snapdragon X2 PCs get their next update, then? It won't be until next year, and maybe not until the second half of 2027 (with 27H2), if Microsoft is going to unify these separate builds at that point – which is seemingly the case.

Essentially, the idea is to keep the 'Bromine'-powered Windows 11 builds distinct from the existing incarnation of the OS until Microsoft has had a whole lot more time to fix any bugs the new platform might cause with AMD and Intel (non-Arm) CPUs. This should avoid a repeat of the bug-fest that was 24H2, and Microsoft appears to have learned its lesson from that disastrous run of glitches.

The downside is that it's admittedly confusing for consumers to have these separate Windows 11 releases, and it might leave Snapdragon X2 laptop owners a bit miffed as to why they are missing out on the 26H2 update later this year. (Note that they shouldn't be out in the cold with new features, mind, which can still be delivered with monthly cumulative updates).

Also noteworthy here is that Snapdragon X2 chips are singled out in Microsoft's blog post, but only that silicon – there's no mention of Nvidia's N1X chip. This rival Arm CPU was rumored to be arriving in Q1 2026, meaning by next month, but with nothing said about Nvidia's processors in this clarification post about 26H1, that's looking a shakier idea now.

Maybe we shouldn't read too much into that, but still – it certainly casts a shadow of doubt for the time being.

• The Xbox app has been ported to Windows 11 PCs with Arm CPUs
• This is another stride forward to make gaming better on Arm laptops, following recent moves to bolster game compatibility
• Microsoft observes that: "Today, more than 85% of the Game Pass catalog is compatible with these PCs."

Microsoft has announced that the Xbox app is now out on Windows 11 PCs with Arm processors, meaning the likes of Copilot+ laptops built around Snapdragon X silicon.

In a blog post, Microsoft noted that: "Players can now download and enjoy a wide range of titles from the Xbox PC app game catalog on Arm-based Windows 11 PCs. Today, more than 85% of the Game Pass catalog is compatible with these PCs."

The firm also said it's working with developers to get more games supported on Arm, and that for those titles which can't be run on these Windows 11 devices, there's always another option – streaming via Xbox Cloud Gaming.

The Xbox app offers a central hub for all your games, acting as a game launcher and tying in social aspects, as well as facilitating cloud gaming, as noted, alongside other neat tricks (such as cross-platform saves for your Xbox and PC).

Analysis: big things could happen for Windows on Arm in 2026

This is certainly a useful step forward for Windows gamers who have an Arm PC, and it's one of a few strides that Microsoft has taken in terms of making Snapdragon X laptops better on the gaming front.

That includes compatibility with quite a few more PC games, with the notable addition of titles that use AVX or AVX2 extensions, which Prism – Microsoft's emulation layer to run x86 (AMD and Intel) games on Arm processors – gained the ability to cope with in the October 2025 update for Windows 11.

On top of that, we saw Epic tune up Fortnite to run on Arm PCs, making the game's anti-cheat tool (EAC) compatible. That has long remained a stumbling block for gaming with an Arm processor. Hopefully, other game developers who use EAC to keep cheats at bay will follow suit. (This includes titles such as Elden Ring).

The porting of the Xbox app over to Arm PCs comes with some interesting timing, as we're about to see new Snapdragon X2 laptops introduced in the near future. On top of that, rumor has it that laptops packing Nvidia's N1X processor could debut in the next couple of months. It's another Arm chip and one that has some seriously powerful integrated graphics. Given that, 2026 could be a very promising year for Windows on Arm devices.

• Arm’s valuation drops massively as investors react to Qualcomm’s shift toward RISC-V development
• Qualcomm’s Ventana acquisition prompts concerns about future Arm demand
• Market response reflects uncertainty over evolving CPU strategies

Arm’s market value dropped sharply following Qualcomm’s announcement of its acquisition of Ventana Micro Systems, triggering renewed concern about the British chip designer’s long term position.

Shares have slipped over the past several days, dropping to about $130.60 at one point, leaving Arm's market cap near $138.85bn.

Qualcomm’s stock moved in the opposite direction, rising to about $170.39 and lifting its market cap to roughly $190.55bn.

RISC-V gets a boost

The reaction came after Qualcomm announced plans to deepen its work on RISC-V through the Ventana deal.

The move points to a future in which Qualcomm could develop RISC-V cores alongside the Arm based Oryon CPUs used in its current platforms, something traders see as a possible shift in long term architecture planning.

The acquisition strengthens Qualcomm's RISC-V engineering base. Durga Malladi, the chip giant’s executive vice president and general manager, said, “At Qualcomm, we are committed to shaping the future of intelligent computing. We believe the RISC-V instruction set architecture has the potential to advance the frontier on CPU technology, enabling innovation across products. The acquisition of Ventana Micro Systems marks a pivotal step in our journey to deliver industry-leading RISC-V based CPU technology across products.”

Balaji Baktha, Ventana's CEO, said, “We are thrilled to join the Qualcomm team and contribute our RISC-V expertise in the development of Qualcomm’s leading Oryon CPU technology. Qualcomm’s acquisition marks an exciting new chapter for our team and we look forward to joining Qualcomm’s team to push the boundaries of energy-efficient high-performance for next-generation products and experiences.”

Ventana, founded in 2018, develops high performance RISC-V chiplet designs aimed at data center and enterprise uses.

Its Veyron V2 design features up to 32 cores, vector units and matrix accelerators built for AI tasks.

The company has also discussed a next generation line expected to deliver higher clocks and expanded math capabilities.

Qualcomm hasn’t provided timelines for any Ventana based products. It continues to work on its Oryon CPU roadmap in parallel, although the acquisition suggests it’s preparing a broader architecture strategy.

Arm shareholders reacted quickly to that possibility, sending the stock downwards as markets weigh how expanded RISC-V adoption from a major customer could negatively influence demand for Arm IP in the years ahead.

• Arm-based Neoverse CPUs can now communicate directly with Nvidia GPUs efficiently
• NVLink Fusion eliminates PCIe bottlenecks for AI-focused server deployments
• Hyperscalers such as Microsoft and Google can use custom Arm CPUs immediately

Nvidia has announced Arm-based Neoverse CPUs will now be able to integrate with its NVLink Fusion technology.

This integration allows Arm licensees to design processors capable of direct communication with Nvidia GPUs.

Previously, NVLink connections were primarily limited to Nvidia’s own CPUs or servers using Intel and AMD processors, so hyperscalers such as Microsoft, Amazon, and Google can now pair custom Arm CPUs with Nvidia GPUs in their Workstations and AI servers.

Expansion of NVLink beyond proprietary CPUs

The development also enables Arm-based chips to move data more efficiently compared to standard PCIe connections.

Arm confirmed its custom Neoverse designs will include a protocol that allows seamless data transfer with Nvidia GPUs.

Arm licensees can build CPU SoCs that connect natively to Nvidia accelerators by integrating NVLink IP directly.

Customers adopting these CPUs will be able to deploy systems where multiple GPUs are paired with a single CPU for AI workloads.

The announcement was made at the Supercomputing ’25 conference and reflects participation by both CPU and GPU developers.

Nvidia’s Grace Blackwell platform currently pairs multiple GPUs with an Arm-based CPU, while other server configurations rely on Intel or AMD CPUs.

Microsoft, Amazon, and Google are deploying Arm-based CPUs to gain more control over their infrastructure and reduce operational costs.

Arm itself does not manufacture CPUs but licenses its instruction set architecture and sells designs for faster development of Arm-based processors.

NVLink Fusion support in Arm chips allows these processors to work with Nvidia GPUs without requiring Nvidia CPUs.

The ecosystem also affects sovereign AI projects, where governments or cloud providers may want Arm CPUs for control-plane tasks.

NVLink allows these systems to use Nvidia GPUs while maintaining custom CPU configurations.

Softbank, which previously held shares in Nvidia, is backing OpenAI’s Stargate project, which plans to use both Arm and Nvidia chips.

NVLink Fusion integration, therefore, provides options for pairing Arm CPUs with market-leading GPU accelerators in multiple environments.

From a technical perspective, NVLink expansion increases the number of CPUs that can be used in Nvidia-centric AI systems.

It also allows future Arm-based designs to compete directly with Nvidia’s Grace and Vera processors, as well as Intel Xeon CPUs, in configurations where GPUs are the main computational units.

The development may reduce the appeal of alternative interconnects or competing AI accelerators, but chip development cycles could affect adoption timing.

• Fortnite finally runs on Windows on Arm laptops as the game's anti-cheat tool (EAC) is now compatible with the architecture
• The game runs well on these Snapdragon X chips, too, going by reports
• This move was facilitated by a new Epic Online Services SDK, and other EAC games should work on Arm soon

It's finally happened – you can now play Fortnite on a Windows 11 laptop with an Arm processor.

Windows Latest reports that Epic has made good on its promise to get Fortnite's anti-cheat tool working with Windows on Arm machines, those super-slim Copilot+ laptops with Snapdragon X chips which use a different architecture compared to traditional Windows 11 laptops.

What's the problem with that? Well, the Arm architecture being different to x86 processors (from AMD and Intel) means that emulation (specifically a translation layer called Prism) is required to run apps and games on Snapdragon X laptops, if they aren't natively coded for Arm (and many aren't, of course).

That translation layer exacts a performance overhead, as you might guess, but more problematically certain tech just doesn't work on Arm systems at all, notably anti-cheat tools in the gaming world. Hence online games like Fortnite, Call of Duty, Valorant, Apex Legends and so on won't work on a Snapdragon X device, as they all use some kind of anti-cheat utility.

As of version 38.00 of Fortnite, the shooter now works on Arm PCs as the Easy Anti-Cheat (EAC) system is now natively supported, although the change wasn't documented in the release notes, oddly.

The game has been verified as running fine on Snapdragon X laptops, though, by Windows Latest and also other reports on Reddit. All of which agree that Fortnite runs nice and smoothly, even on the lower-end Snapdragon chips.

Analysis: other online games will hopefully follow soon

It's good to see Fortnite finally up and running on Arm devices, and performing well considering that the game is emulated. (There's native support for EAC now, but not for the game itself, which works via Prism – and the translation layer does a good job here by all accounts).

EAC compatibility was provided in the Epic Online Services SDK released back in August, but it's still up to game developers to implement that support in their respective titles – it took Epic a few months to get this in Fortnite itself, of course. Other EAC games include Apex Legends and Elden Ring, so hopefully we might see those running on Snapdragon X laptops before too long.

Outside of EAC, Arm CPU support will be a thornier issue, and cheating prevention systems such as Riot Vanguard (Valorant) and BattlEye (which is part of PUBG's armory of anti-cheat tech, as well as Rainbow Six Siege and many other shooters) will need their own solutions – which may or may not be forthcoming. Given that Arm isn't exactly a pressing issue for many developers, we could be waiting a long while – maybe forever if you're being pessimistic (and the same is true of getting Linux support for these anti-cheat tools, too, for SteamOS and handheld owners).

It's worth noting that Microsoft also applied improvements to Prism recently to ensure more games run on Snapdragon X chips – resolving a problem tied in with AVX and AVX2 extensions – so gaming with Windows on Arm has come on a long way in the past month.

That's good news because these Copilot+ laptops are highly portable devices with impressive battery chops, so to be able to engage in some meaningful gaming while on the go is a definite boon.

• Windows 11 is rumored to get an update early next year
• It might be called the 26H1 update, but Microsoft may avoid that name
• Whatever the case, it'll be for Snapdragon X2 Elite laptops only – but any features therein will come to all Windows 11 PCs in the 26H2 update

Windows 11 25H2 hasn't long been on the scene, but already talk has turned to next year's big update – and the possibility that there could be two major updates for 2026.

At this point, this is just a rumor that Neowin spotted – and even the source of the speculation isn't sure there will be both a version 26H1 and 26H2 next year – but seemingly there will be a major upgrade for Windows 11 early in 2026.

In a post on YouTube, leaker PhantomOfEarth informs us that this update won't be for most people. It will only be deployed for Copilot+ PCs with Arm chips. Specifically, we're told Microsoft will be deploying this new version of Windows 11 alongside the launch of Snapdragon X2 Elite laptops (early in 2026).

According to PhantomOfEarth, "[It looks] like a Windows 11 version 26H1 is about to happen for Snapdragon X2 Elite launch ONLY (this will NOT be releasing to existing PCs, that'll be 26H2)."

In a further post, PhantomOfEarth clarifies that this new version of Windows 11 is definitely in the works – add salt with this claim, naturally – for those X2 Elite PCs only, but that it might not be called version 26H1.

Although a reference to '26H1' does apparently exist, so that's a sizeable hint this could be happening. Whatever the case, in theory, there's a big update for Windows 11 due early in 2026, but most Windows 11 PCs won't get it.

Analysis: an Arm-full of updates – but everyone will get the goodies eventually

For most people, then, the big release for next year will be Windows 11 26H2 as expected.

One reason Microsoft might not name the purported first upgrade for 2026 as 26H1 is that it could make folks feel like they're missing out on something. Although, on the other hand, such a naming decision might make Snapdragon X2 Elite laptops feel like they're more on the cutting-edge of Microsoft's desktop ecosystem, and perhaps help to drive sales (of those upgrading from Windows 10 – there being quite a few holdouts still, myself included for now).

In reality, this isn't much different from what Microsoft already did in the past with the introduction of the first Snapdragon X laptops. If you remember, these devices received Windows 11 24H2 before everyone else when they were launched in the middle of the year. (24H2 didn't roll out to all Windows 11 users until October 2024, a few months after the initial Copilot+ PCs began shipping with 24H2 on board).

It's still a pretty sure bet that Windows 11 26H2 will incorporate all the work in 26H1 (or whatever it's called) anyway, so all this means is that non-Arm PC owners are just waiting a bit longer for an upgrade (again). Essentially, the early 2026 update for Windows 11 is likely to contain some vital work to ensure the new Snapdragon X2 Elite CPUs work as they should. (Again, in the case of 24H2, that packed some fundamental changes in terms of a new underlying platform for Windows 11, which was necessary for Snapdragon X chips to actually work).

Of course, there'll be some shiny toys for Snapdragon X2 devices, too, no doubt.

• Google is moving thousands of internal workloads from x86 to Arm CPUs
• The company built an AI tool called CogniPort to automate migration fixes
• Google engineers spent months fixing test failures linked to x86 infrastructure

Google has embarked on a hugely ambitious project to migrate all its internal workloads from x86 to Arm-based CPUs, a process that involves one of the largest hardware transitions ever attempted by a global tech company.

The effort aims to allow its systems to run efficiently on both x86 processors and its custom-built Axion silicon.

With roughly 30,000 applications already converted, Google continues to rely heavily on automation to handle the massive codebase involved in the process.

Porting workloads at warehouse scale

In a blog post outlining the project, Google’s engineering fellow Parthasarathy Ranganathan and developer relations engineer Wolff Dobson noted the migration began with some of the company’s most critical systems, including F1, Spanner, and Bigtable.

Initially, teams relied on conventional software development practices with dedicated engineers and weekly coordination meetings.

Although they expected major architectural hurdles, modern compilers and debugging tools helped reduce many of the anticipated issues.

However, a large amount of time was still devoted to adjusting thousands of tests that were closely tied to Google’s existing x86-based infrastructure.

Engineers also faced challenges updating legacy build and release systems, managing production rollouts, and ensuring stability across mission-critical environments.

To accelerate the transition, Google developed a new AI tool known as “CogniPort.”

The system works by analyzing build and test errors and then attempting to automatically fix them, particularly in cases where an Arm-specific library or binary fails to compile.

CogniPort has shown a success rate of around 30%, performing best when handling test corrections, data handling inconsistencies, and conditional platform code.

While not flawless, the tool represents a key step in enabling automation at warehouse scale and reducing the human workload required for such conversions.

The long-term motivation behind Google’s move lies in performance and efficiency - its Axion-powered Arm servers reportedly deliver up to 65% better price-performance and can be as much as 60% more energy-efficient than comparable x86 instances.

This shift could result in fewer x86 processors across Google’s vast data infrastructure, potentially transforming the makeup of its internal compute clusters.

For now, major applications such as YouTube, Gmail, and BigQuery already operate on both x86 and Arm-based systems.

As Google migrates the remaining 70,000 packages, doubts persist about whether AI tools can handle such scale without adding new maintenance challenges across its systems.

Via The Register

• Windows 11's latest update packs an important feature for Arm CPUs
• It introduces support for AVX and AVX2 with Prism emulation
• Some PC games use these extensions, and previously wouldn't work with the likes of Snapdragon X chips – but they now should

Microsoft just gave Windows 11 laptops with Arm processors better support for games (and software too) courtesy of the latest update for the OS.

Windows Latest reports that the October 2025 update (KB5066835) for Windows 11 (both versions 24H2 and 25H2, which are essentially the same thing) comes with a big update for Prism. This is Microsoft's emulation layer to run x86 (AMD and Intel) apps and games on Arm chips like Qualcomm's Snapdragon X.

Not all games work with Prism, and one problem area has been titles that use AVX or AVX2. These are extensions to the x86 instruction set which, on a basic level, pull off some fancy CPU tricks to run faster – and now Prism can handle those extensions. (And others besides, including BMI, FMA, F16C, and more, but AVX is the main advance here).

In short, some games that use AVX or AVX2 which didn't load up at all will now work, so gamers have a wider range of compatibility with Windows on Arm.

As mentioned, it's not just games, but also software, and that includes some Adobe Creative Cloud apps, for example, which those with a Snapdragon X laptop haven't had access to before.

Analysis: a year-long wait

This expanded support for Prism entered testing almost a year ago (November 2024, in fact), so it has taken Microsoft quite some time to work this through and get it out of preview.

Windows Latest tested the beefed-up version of Arm emulation, and found that it allowed most of the PC games in their Steam collection to successfully launch and work. However, the tech site observed that there was "wildly varying performance" from game to game.

Reports on Reddit reflect this, so sadly this isn't a panacea for all AVX-related gaming woes on Windows 11 with Arm. We must also bear in mind that emulating a game or app will always have some overheads, and performance is never going to be as good as running a native piece of software (coded for Arm, rather than x86 AMD or Intel).

Still, Microsoft finally getting this new feature for Prism live is a big step forward, even if it did take a long time. While complaints remain around the frequency of Qualcomm's (integrated Adreno) GPU updates, and hardware driver compatibility issues with Windows on Arm in general, progress is definitely being made towards making Arm laptops a more reliable platform. Anti-cheat tools remain a problem for gamers, too, but a solution is finally in the works and seemingly imminent.

You might also like

• I was about to upgrade to Windows 11, but I've decided to stick with Windows 10 – here's why
• Windows 10's demise could be Linux's gain if the flood of defectors is anything to go by
• Microsoft is 'removing known mechanisms for creating a local account' from Windows 11 setup - get ready to use a Microsoft account

• OpenSFI aims to standardize how CPUs and firmware talk across architectures
• ByteDance’s participation signals rare cooperation between Chinese and U.S. tech giants
• The project could simplify cross-vendor firmware integration for future chip generations

A new collaboration between leading chipmakers and cloud firms will bring together AMD, Arm, Intel, Google, ByteDance, Microsoft, MiTAC, HPE, and others to create a more unified firmware foundation.

The initiative, known as openSFI (Open Silicon Firmware Interface), seeks to define a single, architecture-neutral interface for how host firmware communicates with CPU silicon during initialization and operation.

The project is notable for its broad membership, as ByteDance stands out as the only Chinese participant among American and European technology giants.

How openSFI fits into the firmware stack

The openSFI project builds upon and complements AMD’s ongoing openSIL effort, which aims to replace the proprietary AGESA codebase for silicon initialization with a transparent, open-source alternative.

Within this new structure, openSFI acts as a unified layer sitting atop AMD’s openSIL and Intel’s FSP (Firmware Support Package).

In effect, it serves as a shared abstraction point that lets host firmware call standardized functions without concern for the underlying silicon vendor.

This approach could streamline platform development, reduce redundant engineering, and allow firmware to be reused across different CPUs.

The published openSFI 0.3 specification lays out the project’s design principles and integration model, noting its goal is to, “enable modular, scalable, and vendor-agnostic integration of silicon into host firmware environments.”

The specification emphasizes a stable API contract for host firmware to call silicon initialization functions predictably and consistently.

It also outlines the primary objectives of openSFI: unifying silicon initialization interfaces, simplifying firmware integration, and promoting reusability between vendors to cut validation costs and improve sustainability.

ByteDance’s involvement is notable in a field often dominated by Western and Japanese semiconductor firms.

By joining openSFI, ByteDance opens an unusual level of engagement between a Chinese tech firm and major U.S. semiconductor companies.

Although the ambition is promising, there is no information on how openSFI will be adopted across competing hardware ecosystems.

Previous attempts at firmware standardization have struggled due to differing vendor interests and evolving hardware platforms.

Still, the shared backing of major chip and data center firms gives openSFI a stronger starting point than past efforts.

• Arm’s Lumex chips promise huge improvements to on-device AI
• Its CPUs offer up to 5x better AI performance
• CPUs are seen as the powerhouse for AI

Arm has lifted the wraps off its next-generation Lumex chip designs, optimized to run some local AI workloads on mobile devices.

Its architecture allows for four different design types, ranging from energy-efficient cores for wearables to high-performance cores for flagship smartphones.

Slating accelerated product cycles, which result in tighter timescales and reduced margin for error, Arm says its integrated platforms combine CPU, GPU and software stacks to speed up time-to-market.

Arm’s Lumex could be used in your next smartphone

Arm described Lumex as its “new purpose-built compute subsystem (CSS) platform to meet the growing demands of on-device AI experiences.”

The Armv9.3 C1 CPU cluster includes built-in SME2 units for accelerated AI, promising 5x better AI performance and 3x more efficiency compared with the previous generation.

Standard benchmarks see performance rise by 30%, with a 15% speed-up in apps and 12% lower power use in daily workloads compared with the prior generation.

The four CPUs on offer are C1-Ultra for large-model inferencing, C1-Premium for multitasking, C1-Pro for video playback and C1-Nano for wearables.

The Mali G1-Ultra GPU also enables 20% faster AI/ML inferencing than Immortalis-G295, as well as improvements across gaming like 2x better ray tracing performance.

Lumex also offers G1-Premium and G1-Pro options – but no G1-Nano.

Interestingly, Arm positions CPUs as the universal AI engine given the lack of standardization in NPUs, even though NPUs are starting to earn their place in PC chips.

Launching with Lumex is a complete Android 16-ready software stack, SME2-enabled KleidiAI libraries and telemetry to analyze performance and identify bottlenecks, allowing developers to tailor Lumex to each model.

“Mobile computing is entering a new era that is defined by how intelligence is built, scaled, and delivered,” Senior Director Kinjal Dave explained.

Looking ahead, Arm notes that many popular Google apps are already SME2-enabled, meaning that they’re prepared to benefit from improved on-device AI features when next-generation hardware becomes available.

You might also like

• Need an upgrade? These are the best business smartphones
• I spent some time with the iPhone 17 Pro and am already in deep like
• The best cloud computing providers give you access to powerful chips today

• Arm moves closer to silicon manufacturing, risking strained ties with major partners
• Strategic hires reveal Arm’s deepening ambition to design complete chips independently
• Sinno’s Amazon track record signals Arm’s growing focus on artificial intelligence

Arm, long rumored to be considering building its own chips, is reportedly moving ahead with its “Picasso” project, potentially competing with major customers while also planning royalty hikes.

Long known for licensing its processor designs to customers rather than making its own chips, Arm has made some strategic hires, including engineers with backgrounds at major firms, such as Nicolas Dube from HPE and Steve Halter, who worked at Intel and Qualcomm.

The latest addition is Rami Sinno, a former vice president of engineering at AWS’s Annapurna Labs, where he directed the development of Amazon’s “Trainium” and “Inferentia” processors tailored for artificial intelligence workloads.

Investment and ambition

Sinno had earlier played a key role in Amazon’s effort to build chips that could undercut Nvidia’s dominance in AI computing, both on price and performance.

Arm’s latest moves suggest that the company is strengthening its ability to design complete chips and systems.

Its designs underpin nearly every smartphone on the market, and in recent years, it has also gained traction in servers once dominated by Intel and AMD.

As it edges closer to producing its own silicon, the market dynamics are set to change, and Arm is setting funds aside to achieve its goal.

In July 2025, Arm disclosed that it would allocate part of its profit toward building not only traditional chips but also chiplets.

These smaller, specialized components can be stitched together into larger systems, a strategy increasingly adopted across the industry.

CEO Rene Haas has described this as a natural extension of Arm’s design business, moving into areas where the company currently provides intellectual property but not finished hardware.

Until now, Arm’s income has depended heavily on the royalties it collects from chipmakers who integrate its designs into their products.

However, with data centers shifting toward Arm-based servers, the company appears to see an opportunity to sell more complete solutions.

Such a move could put it in direct competition with its largest customers, including Apple, Qualcomm, and Nvidia, all of which rely on Arm’s intellectual property.

According to industry experts, this carries strategic risks.

Arm could alienate companies that form the backbone of its business while attempting to challenge entrenched rivals in the server and AI markets.

The company’s attempt to raise royalty rates may also create another source of tension with partners.

While the hiring of Sinno and other experienced executives underscores Arm’s seriousness, the extent to which the company can transform itself from a design house into a chipmaker remains uncertain.

Via Reuters

You might also like

• We've also listed the best AI tools right now
• Here's a list of the best mini PC for every budget
• Ever wanted to listen to your Google Docs? Well good news - you finally can, thanks to Gemini

• Epic has made its Easy Anti-Cheat tool compatible with Windows on Arm
• This means online games that use the tool work with Arm-based laptops
• That includes Fortnite, which is going to be one of the first games to implement this support for Snapdragon X (Arm) chips

If you've got a Copilot+ laptop with a Qualcomm Snapdragon X (Arm) processor that you also use for gaming duties, here's some great news: one of the biggest stumbling blocks for gamers on Arm has now been banished.

As Thurrott.com reports, Epic has made good on its promise to make its Easy Anti-Cheat (EAC) tool compatible with Windows on Arm (the version of Microsoft's OS that's installed on Copilot+ PCs with Arm chips, as opposed to AMD or Intel silicon).

The move comes with the new version of the Epic Online Services SDK (for game developers) which now boasts EAC compatibility to offer "secure multiplayer experiences on Arm-based Windows 11 devices", as Epic notes.

EAC is used by some big-name online games to weed out problem players who are using various shady tricks or exploits, and the lack of compatibility with the anti-cheat system meant that those games didn't work at all on Arm-based laptops.

Said games include Epic's Fortnite among other online shooters (Apex Legends, PUBG, and more), as well as the likes of Elden Ring.

Analysis: the ball is now in the developers' court

Epic previously promised that it was going to make EAC work with Windows on Arm back in March 2025, and what we see here is the necessary work being implemented in the mentioned SDK (software development kit).

However, that isn't the end of this story – not quite. It's now up to game developers to incorporate the support in their titles, and Epic is going to be on the ball itself in that respect when it comes to bringing support to Fortnite.

A Qualcomm spokesperson told Thurrott.com: "Fortnite will be among the first titles to take advantage of this compatibility, bringing one of the world's most popular games to Snapdragon-powered laptops."

At any rate, support is going to filter through gradually to a number of games – hopefully quite swiftly – and this will remove one of the black marks in the 'against' column for Arm-based laptops.

Those devices running Snapdragon X chips can make a good enough effort at running even more demanding games – recall the early days of being impressed at these notebooks running Baldur's Gate 3 in 4K.

Being unable to play Easy Anti-Cheat games on these machines was a notable weakness, and particularly with Fortnite, given that this is a popular choice for laptop gamers (and not a demanding title), it was a shame to see it out of the grasp of those with Arm-based Copilot+ PCs – but that won't be the case for much longer.

You might also like...

• Microsoft promises to crack one of the biggest problems with Windows 11: slow performance
• No, Windows 11 PCs aren't 'up to 2.3x faster' than Windows 10 devices, as Microsoft suggests – here's why that's an outlandish claim
• Microsoft has all but given up on Windows 11 SE – and it looks like the war against Chromebooks has been lost

Formula 1 has long been at the forefront of technology development - staying on top of the mountains of data the cars produce each race weekend isn’t just a useful tool, it’s the heart of the team.

That means a team’s technical partnerships are paramount to its success, and Aston Martin’s official AI compute platform partner, Arm, is no exception.

Arm claims to be the industry’s, ‘most power-efficient and highest-performing compute platform with unmatched scalability,’ and I spoke to John Kourentis, Arm’s Director of Automotive Go-To-Market in EMEAI, to find out more about their involvement.

Turning data into downforce

Data is everything in Formula One, with the margins of just a few seconds deciding the difference between first and last place.

Changing the front wing by a few millimetres can make or break a team’s race, and preparation is key to understanding how the car will perform on any given weekend.

"The trajectory for Aston Martin looks really strong, and so from that ambition, we can be ambitious in terms of where the technology path leads to as well. So, it really maps well to a partner that really wants to invest in driving performance through electronic investment and electronic innovations," Kourentis tells TechRadar Pro.

Crucial to the car development is the use of a new, state-of-the-art CoreWeave wind tunnel, and whilst drivers and strategists are away on race weekends pushing the car to its limits, engineers back in the team's Silverstone base are collecting as much aerodynamic information as they can in the limited wind tunnel time they are allowed;

“The detail and the technology that they deploy to understand the aerodynamics of the vehicle inside the wind tunnel is mind blowing” Kourentis explains; “they need the best sensor data capture technology that they can possibly get.”

That’s where Arm comes in. Aston Martin uses miniature sensors in the body of the car, and these sensors deliver real time information back to engineers, where Arm-powered compute is used to filter and process the signals in real-time.

In this wind tunnel, there's a scaled down model of the car, around 40% of the car’s actual size.

This model is packed with over a thousand sensors, and the engineers and aerodynamicists then rotate the car’s position and analyze the wind flow and aerodynamics of the car, adding and modifying car parts to see how these changes affect the air flow.

“All of that requires a really, really comprehensive suite of sensor data technology. That's where this aspect of the partnership is really focused, in making sure that we've got the best sensor data technology they can possibly get, getting the right fidelity of data that they can then actually use for aerodynamic optimizations in the future,” Kourentis says.

Optimized for success

Wind tunnels haven’t always been a feature of Formula One. As the sport has evolved, the margins have only gotten smaller, and aerodynamics in particular has taken centre stage as one of the sport’s most important elements, despite what Enzo Ferrari might have thought.

In the early days of the competition, aerodynamic testing was primarily done on-track, with engineers relying on race-weekends to collect data. This meant their data was at the mercy of track, weather, and driving conditions - and produced a desperate need for a more controlled environment.

Now, regulations and budget caps are the factors that influence data collection - it’s only in the last few months Aston Martin has had its own facility;

“Before having their own wind tunnel, they were sharing a wind tunnel with Mercedes,” Kourentis points out.

“And what that meant was that it was never really set up and always optimized for their use. I'm not sure there's much love shared between the various different F1 teams. So the first major benefit is they have their own wind tunnel facility, is that it's always set up and optimised exactly how they want it to be.”

Unfortunately, just because Aston Martin has its own wind tunnel, that doesn’t mean it can use the tech whenever it wants.

Formula 1’s governing body, the FIA, has a strict limit on the amount of time each team can spend using the tunnel, so as Kourentis says, “that means that every minute, every second of analysis that they do in the wind tunnel really, really does count.”

“That was part of the considerations in terms of some of the sensor technology and the Arm-based microcontrollers that are deployed in that model, is they need to get really, really high fidelity data in actually quite a short space of time because they don't have the luxury of being able to run it 24/7.”

Removing friction

On a more technical level, Arm is focused not just on delivering large volumes of high-quality data, but also on making this process as seamless as possible.

In a sport as precise as F1, any delay or barrier to information can be extremely costly, Kourentis notes;

“So a lot of the work we do is around removing friction from the software ecosystem to allow them to officially deploy their software on Arm architecture. And that's a lot of the work that we've been doing with the Aston Martin team.”

Arm is able to do this by keeping things consistent, allowing the team to utilize the best sensor technology and microcrontroller technology, and pair them with the most efficient software components, whether these are real-time operating systems, middleware, or other applications.

Arm works with software partners to make sure all designs are optimized for the underlying technology;

“One of the beauties of Arm and one of the reasons why we've had the success that we have is that we have managed to keep a consistent architecture for all of our CPU designs, our GPU designs and our ISP designs, which means it delivers a very, very consistent way to land software on those CPUs and families of CPUs,” Kourentis argues.

“To consistently do that over decades and decades and keep that entire ecosystem covering along with the latest and greatest architecture that we change is the real skill of Arm, to be honest with you.”

All change for 2026

Most Formula 1 fans will be looking to Aston Martin with anticipation, especially as the upcoming 2026 regulation changes come in, potentially bringing a shake-up of the grid.

The regulations represent a serious opportunity for the team, whose performance recently could perhaps be described as underwhelming, given previous glimpses of success.

“[New regulations will be] Engine heavy for sure. More of a hybrid focus, more electrical energy coming in in 2026. There are some considerable regulation changes around aero as well,” Kourentis confirms.

Aston Martin recently signed aerodynamicist and legend of the sport, Adrian Newey, an engineer who has designed a staggering 12 constructor's championship-winning cars, in a bid to boost its fortunes, and although the new regulations are said to bring a focus on engine power, aerodynamics will undoubtedly be of huge importance, Kourentis notes;

“I'm not an aerodynamic expert by any means, but understanding how the car performs with aerodynamic changes when it's cornering along different types of corners, in different circuits, different conditions, different speeds.”

“You could see that having a really, really strong sense of data platform in the wind tunnel to be able to capture those new aspects of how the regulations evolve is super important. And that's where most of the time [will be spent]. Adrian and the team are really focused on next year's car right now.”

Aston Martin fans will no doubt be eagerly awaiting the new season, where they’ll be benefiting not just from the wealth of Newey’s expertise, but also from the use of its very own wind tunnel, and the hoards of valuable data they can collect.

• “The sport is relentless- if you're standing still, you're going backwards” - McLaren F1 tells us what winning a world championship means for its technology stack

• Nvidia's N1X chip has been spotted in a Geekbench result
• The specs show the integrated GPU has 6,144 CUDA cores
• That equals the RTX 5070 for pure core count, but there's much more to factor in when it comes to performance

Remember Nvidia's rumored CPU that caused quite a buzz on the grapevine last year? We've apparently now seen this consumer chip in a benchmark leak, where the spilled spec details are the key aspect.

Tom's Hardware reports that the N1X chip, which is Arm-based (like Qualcomm's Snapdragon X CPUs), has been spotted in a Geekbench result, specifically for the OpenCL (graphics) test, where it scored 46,361.

That score is pretty much meaningless at this point. This is an early engineering sample of the N1X (in theory), and even then, if you want to gauge graphics performance, Geekbench is far from the first choice of synthetic benchmarks.

As noted, though, this gives us a tantalizing glimpse of the spec, which shows that (add salt now) the N1X will have 20 cores, apparently split into a pair of 10-core clusters. That's the processor itself, but we also see the integrated GPU here, which is shown to have 48 Streaming Multiprocessors - that equates to 6,144 CUDA cores.

That sounds like a lot, right? Well, it is, and in fact, those familiar with Nvidia's graphics cards will realize that this is in the ballpark for a mid-range current-gen GPU - to be precise, the RTX 5070, which, in fact, has that exact core count.

Analysis: cautiously optimistic

So, are we getting a compact consumer chip that could go in budget laptops or handhelds to deliver the same frame rates as the mighty RTX 5070? In a word, no, but the N1X still looks to be shaping up as a promising piece of silicon, and one that will have rivals sitting up and taking notice.

As to the reasons why performance can't simply be drawn from the number of cores seen on the GPU here - it's not a patch on the RTX 5070 in this benchmark, of course - there are other important factors at play aside from the basic core count.

That includes the clock speed and the power supplied to the GPU, which is a very different scenario with integrated graphics in a chip like this versus a full-on graphics card in a desktop PC. As well as considering the power envelope, throw in bandwidth limitations too - in terms of piping tasks over to the system memory, with no on-board VRAM of course - and the upshot is a good deal of headwinds.

That won't stop the N1X from being a potentially sterling performer for an all-in-one chip, but there's not much point trying to guess at the exact level of graphics performance that it might provide at this stage. (Certainly not from the leaked benchmark here, as already noted).

Tom's makes an interesting observation, which is that the leaked specs match Nvidia's GB10 'superchip' built for powerful AI performance and ushering in the era of the tiny AI supercomputer (pictured above). There's no reason why Nvidia couldn't put out another spin on this for consumer-targeted devices, including mini PCs and laptops, and indeed, gamers are getting particularly excited about the possible use in handhelds.

For now, though, this is still very much in rumor territory. If previous speculation is to be believed, we might see Nvidia's consumer CPU revealed later this year, ahead of a launch in early 2026.

You might also like

• Nvidia GeForce RTX 5070 Ti review: nearly perfect, but with one major flaw
• The best Nvidia graphics card in 2025: top picks from Team Green for gaming, creating, and more
• AMD could be working on a faster RX 9070 Extreme graphics card

• Acorn’s ARM1 chip kicked off a 40-year computing legacy
• Arm chips now power over 300 billion devices worldwide and counting
• 99% of smartphones run on Arm and there's growing adoption in IoT, cloud, and AI workloads

In April 1985, a small team at Acorn Computers in Cambridge, UK, set out to rethink what a processor could be. Engineers Sophie Wilson and Steve Furber developed the ARM1 (it originally stood for Advanced RISC Machines), an unassuming chip with just 25,000 transistors, to power the BBC Micro, crafting a 32-bit processor that emphasized reduced instruction sets for faster, more efficient computation.

The design's low power consumption was partially driven by practical constraints, namely the need to run in cheaper plastic packaging. ARM2 soon followed, incorporated into the Acorn Archimedes, the first RISC-based home computer. ARM3 introduced a 4KB cache and further improved performance.

After the spin-off from Acorn in 1990, ARM Ltd. was founded as a joint venture between Acorn, Apple, and VLSI. One early commercial success was the Apple Newton, followed by widespread adoption in mobile phones like the Nokia 6110, which featured the ARM7TDMI.

Looking to the future

ARM6, introduced in 1991, brought full 32-bit processing and an MMU, key to powering GSM mobile phones. In 2005, Armv7 architecture debuted with the Cortex-A8 processor, which brought SIMD (NEON) support and powered many early smartphones.

In 2011, Armv8 introduced 64-bit support and became the foundation for cloud, data center, mobile, and automotive computing. Features like SVE and Helium pushed performance and AI capabilities further.

The 2021 launch of Armv9 marked the architecture's shift into AI-centric workloads. It introduced Scalable Vector Extension 2 (SVE2), Scalable Matrix Extension (SME), and Confidential Compute Architecture (CCA).

These features made it suitable for everything from smartphones with advanced image processing to AI servers handling generative workloads. SME accelerates generative AI and MoE models, while SVE2 brings enhanced AI capability to general-purpose compute.

Arm's compute subsystems (CSS), based on Armv9, now serve client, infrastructure, and automotive markets. By integrating CPUs, interconnects, and memory interfaces, these CSS platforms support rapid development of specialized silicon.

From the original ARM1 with just 25,000 transistors to today’s Armv9 CPUs packing 100 million gates, the architecture has consistently driven computing forward for four decades. Arm-based chips now power over 300 billion devices worldwide, from tiny embedded sensors to full-scale data centers.

With 99% of smartphones running on Arm and growing adoption in IoT, cloud, and AI workloads, the architecture continues to scale thanks to its energy-efficient design and flexible licensing model.

Looking ahead, there have been growing rumors that Arm could move beyond licensing and into chip production, something that would put it in competition with its biggest customers. This speculation intensified recently following the acquisition of Ampere Computing, Arm’s only independent server chip vendor, by SoftBank, Arm’s Japanese owner.

You might also like

• Softbank buying Ampere for $6.5 billion and could integrate it with Arm
• Is an AMD Arm superchip in the works?
• Arm wants to build its own CPU server for hyperscalers, reports say

• Arm says Neoverse powers nearly half of 2025 hyperscaler compute shipments
• AI growth makes power-efficiency essential for hyperscale data centers
• AWS, Google, Microsoft are leading shift from x86 to Arm infrastructure

Arm says its Neoverse platform is becoming the architecture of choice for the cloud, as the likes of AWS, Google Cloud, Microsoft Azure and others drive a broader shift away from x86 in the data center.

Mohamed Awad, SVP and GM of the Infrastructure Business at Arm, says power efficiency and scalability are redefining infrastructure at hyperscale. “Just over six years ago, we launched Arm Neoverse for the next-generation of cloud infrastructure, recognizing a world where delivering new levels of scalable performance on top of Arm’s flexible and power-efficient compute platform could enable a systemic shift in the capabilities and costs of the data center ecosystem,” he explained.

“Fast forward to today, the adoption of Neoverse has reached new heights: close to 50 percent of the compute shipped to top hyperscalers in 2025 will be Arm-based,” he added. Much of this momentum can, inevitably, be attributed to the rise of AI.

AI servers set to grow by 300 percent

AI is reshaping cloud infrastructure, driving explosive growth in compute demand and forcing hyperscalers to prioritize power efficiency at massive scale. Data centers are being designed in gigawatts, not megawatts, making efficiency a requirement rather than a differentiator, something that has been central to Arm's architecture for decades.

That's not the only reason for the architecture's success of course.

"The Arm compute platform is additionally giving our partners the flexibility to create a new generation of customized, differentiated silicon solutions for AI. For example, Nvidia’s Grace Blackwell superchip for AI-based infrastructure combines Nvidia’s Blackwell GPU accelerated computing architecture with the Arm Neoverse-based Grace CPU, integrated with an extraordinarily high bandwidth, coherent mesh network – a system tailor-made to achieve unmatched performance for AI workloads," Awad says.

With ten of the world’s largest hyperscalers developing and deploying Arm-based chips, the future looks bright for the company - but that dynamic could shift if, as rumored, Arm decides to produce its own silicon.

Speculation on that front gained signifcant momentum after Arm’s Japanese owner, SoftBank, acquired Arm-based chipmaker Ampere for $6.5 billion, raising concerns that Arm could end up competing directly with its partners in the fast-growing data center market.

You might also like

• Softbank buying Ampere for $6.5 billion and could integrate it with Arm
• Is an AMD Arm superchip in the works?
• Arm wants to build its own CPU server for hyperscalers, reports say

• Google’s Chrome installer for PCs with AMD and Intel CPUs has been broken for the past week
• Detective work suggests that the version for these PCs was accidentally replaced by the Arm installer, which is for very different Snapdragon CPUs
• While the issue has now been fixed, Google took a long time to resolve it

Those trying to install the Chrome browser in Windows 11 or 10 over the course of the past week may well have been flummoxed by an error telling them the app won’t run on their PC – and I can’t quite believe how long it’s taken Google to fix this.

Nonetheless, the good news is that the glitch is fixed, even if it took the company way longer than it should have.

The problem, in case you missed it, was flagged on Reddit and by Windows Latest a week ago.

What happened was that on firing up the Chrome Installer file (ChromeSetup.exe, download from Google’s website), people watched the process come to a screeching halt, with an error message that read: “This app can’t run on your PC: To find a version for your PC, check with the software publisher.”

As to the cause, Windows Latest did some detective work, and theorized that what Google had done here – somehow – was accidentally swap the Arm installer of Chrome with the x86 installer. Meaning that the version of Chrome for Copilot+ PCs with Snapdragon (Arm-based) processors got swapped with the version for AMD or Intel CPUs.

That conclusion was based on digging into the setup file and finding references to ‘Arm’ which surely wouldn’t be there with a non-Arm executable. Furthermore, 9 to 5 Google backs up this theory, as the tech site tried the faulty version of Chrome on a Snapdragon PC, and found it worked fine (as it would do if it was the Arm installer).

9 to 5 Google was also on the ball in terms of noticing that the issue is finally cured, and you can now download the Chrome installer on a Windows PC with an AMD or Intel CPU, and it’ll work just as you’d expect.

Analysis: A bafflingly basic mistake with only one winner (Edge)

Okay, so mistakes can be made. Indeed, they happen all the time in the tech world, or elsewhere for that matter. But for such a basic glitch to be left in place to annoy a bunch of Windows users for the best part of a week is baffling – especially while reports were flying around about the problem.

I can’t imagine this was something that was particularly difficult to resolve, either, as it looks like a simple mix up of files, as noted (unless there’s something I’m missing here). Apparently, the Arm version of Google Chrome wasn’t affected, and the installer still worked for those with a Snapdragon-powered machine.

This could have cost Google some Chrome users potentially, who might have got fed up with the browser failing to install, and maybe even plumped for Microsoft Edge, its main rival, instead. (Incidentally, Edge is the best overall web browser as far as our roundup of the most compelling offerings out there is concerned).

You might also like

• Microsoft looks to be making a big change to how you install and log in to Windows 11 – and I’m not happy about it at all
• Microsoft gets rid of ‘Edge uninstall’ advice page after facing criticism over it having nothing to do with removing the app, and just promoting the browser instead
• How to use Microsoft Edge

• SoftBank buying Ampere may signal shift from licensing to manufacturing
• Stargate project could benefit from Ampere’s experienced chip designers
• OpenAI lacks internal chip development team so that makes sense

Arm’s Japanese owner, SoftBank, is set to acquire Ampere - Arm's only independent server chip vendor - for $6.5 billion (approximately ¥973.0 billion).

It’s a big move, and one that could see Arm shift from simply licensing chip designs to manufacturing its own silicon. The move would put it in direct competition with its existing customers but also expand Arm’s footprint in the growing and highly lucrative data center space.

The deal is set to conclude in the latter half of 2025, subject to the usual regulatory approvals, including U.S. antitrust clearance. The Santa Clara, California-based Ampere will continue operating under its current structure until then. The exact reasons for the acquisition aren’t known outside of Ampere and SoftBank, but there are plenty of theories flying around.

It makes a lot of sense

The Next Platform thinks it may have something to do with the Stargate project, which President Trump announced at the start of 2025 and which will see OpenAI working with SoftBank and Oracle (which, incidentally, is a major investor in Ampere) to secure American leadership in AI and boost the US tech sector.

How would Ampere’s acquisition fit into Stargate? The Next Platform notes, “probably somewhere around 1,500 of the nearly 2,000 people at Ampere Computing are chip designers and these people, plus those working at Graphcore, could be tapped by OpenAI to help design custom CPUs and GPUs for the Stargate effort.”

While he’s not claiming any insider knowledge, TNP’s Timothy Prickett Morgan said, “Why else would SoftBank pay $6.5 billion for a company that is hoping to be a second-source processor for the hyperscalers and cloud builders who are all making their own Arm server CPUs and who also buy scads of x86 server processors from Intel and AMD?”

It's a good question. “As far as we know, Sam Altman & Co. has not put together a chip development team of any appreciable size, and even if it had, OpenAI has not created a compute engine and shepherded it through development,” Prickett Morgan concludes.

You might also like

• Softbank buys Ampere for $6.5 billion, could integrate it with Arm and Graphcore
• How Ampere Computing plans to ride the AI wave
• Arm may be planning to hike prices, has considered developing its own chips

• Microsoft has kicked off a new set of Windows 11 preview builds in testing
• These will have “behind-the-scenes platform changes” for the OS
• Rumor has it that Microsoft will be tinkering with the current Germanium platform – rather than switching to an all-new base for Windows 11 – and that should mean fewer bugs

Microsoft is likely switching to work on the next big update for Windows 11, which would be 25H2, based on rumors and what’s going on with the latest preview build.

Microsoft itself has told us that the new builds in the 26200 range, which are now in the Dev channel for testing, are still based on 24H2 – the current version of Windows 11 – but that it’ll be “making behind-the-scenes platform changes in these builds” which might mean they have different issues to the 24H2 builds in the Beta channel (a later branch of testing).

According to Windows Central’s Zac Bowden, who regularly shares rumors relating to what’s going on at Microsoft, those Dev builds are “likely” to be about laying the “early groundwork for version 25H2” which is, of course, due to land later in 2025.

So, in other words, the Beta channel will continue to get builds based purely on 24H2, whereas the Dev channel is now going to get changes under the hood to set the stage for 25H2 (most probably).

Bowden further notes that the changes to the underlying platform Windows 11 is built on – the sprawling mass of code you never see, but is the glue that holds together all the bits that you do interact with – will incorporate the changes needed for Qualcomm’s incoming Snapdragon X2 chip.

Those changes have already been put in place in the Canary channel, apparently – which is the earliest test channel, before Dev – and now they’re coming to the Dev channel, this indicates that Microsoft is progressing towards making the finished version of Windows 11 compatible with devices powered by the Snapdragon X2 CPU.

Those devices, which are expected to usher in a new Snapdragon X Elite Gen 2 processor, should arrive later in 2025, alongside this new version of Windows 11, in theory.

Analysis: Sticking with Germanium?

You may recall that Windows 11 24H2 was built on an all-new underlying platform, dubbed Germanium by Microsoft. That major switch with the very foundations of Windows 11 was made to ensure compatibility with Arm chips (and to improve the performance and overall security of the OS), and the first generation of Snapdragon X processors.

However, that was a huge move, whereas the work now rumored to be underway is seemingly about fine-tuning Germanium for Snapdragon X2 – that isn’t a total change of platform, but a refinement of what was put in place last year. Or at least this is what Bowden feels is the most probable scenario, although it’s still possible Microsoft could switch to a different platform with 25H2, the leaker acknowledges.

If it’s true Microsoft is sticking with Germanium, this is important, because one of the reasons why Windows 11 24H2 has been so buggy is due to that migration to the all-new (at the time) Germanium, which I believe caused quite a commotion in the inner workings of Microsoft’s OS. And that’s the reason why some of the many glitches we’ve seen with the 24H2 update have been so odd (again, in my opinion – take it with caution, of course).

Because Windows 11 25H2 isn’t going to be such a big move, it should be much more smoothly implemented and less buggy overall – at least in theory, and that’s a hope I reckon a lot of folks will be holding onto for now.

At the same time, we have to face the fear that the bad run of bugs might continue, either because there are just that many to stamp out, this firefighting won’t have run its course – or that Microsoft will be switching away from Germanium, which again could mean more than the usual share of bugs winging our way.

Given how 24H2 has panned out – pretty terribly for bugs – hopefully Microsoft will be in a risk-averse frame of mind here.

You might also like...

• Windows 11 users get ready for more ‘recommendations’ from Microsoft – but I’m relieved to say these suggestions might actually be useful
• Are you unable to get security updates for Windows 11 24H2? Here’s the likely reason why, and the fix to get your PC safe and secure again
• Here's a list of all the apps that can run on the Qualcomm Snapdragon X Elite

• The Microsoft Surface 7 has been labeled "frequently returned" on Amazon
• No reason was officially given, but user reviews cite Arm-based incompatibilities
• It's believed this could be happening due to more competition being available

The Qualcomm Snapdragon X Elite-powered Microsoft Surface Laptop 7 has been deemed "frequently returned" on Amazon.

According to our sister site, Windows Central, the warning applies to an Amazon listing for the Microsoft Surface Laptop (2024) featuring a 15-inch touchscreen, Snapdragon X Elite SoC, 32GB RAM, and 1TB of storage.

Despite having nearly 400 user reviews, with an overall rating of 4.2 / 5 stars, the laptop appears to be returned often according to the retailer, with the disclaimer telling prospective buyers to "check the product detais and customer reviews to learn more about this item".

Detailed top reviews on the laptop from verified buyers have rated the Microsoft Surface 7 with five stars, with particular praise for the battery life. However, a common complaint is that "a lot of programs didn't work with Arm" which is something we've said in our own coverage.

In our five-star review, we praised the incredible performance, excellent battery life, great display, and solid pricing, with the only minor negatives being a lacking port selection, lack of an OLED variant and slight issues with the touchpad's sensitivity. With that said, why are users returning it so frequently? Amazon has not issued any reason why.

Another reason the Microsoft Surface 7 could see so many returns is that people could trade the device in for a computer with a competitor's chip. That's because AMD has made waves with its Ryzen AI processors, such as the coveted Ryzen AI Max+ 395, which has just arrived in some of the best laptops and best ultrabooks on the market. Those wanting superior performance for general computing, AI workloads, and greater compatibility may return to x86 architecture.

We can also look at the pricing of the laptop compared to its competition. This listing is available on Amazon with an MSRP of $2,099.99. However, the machine has been discounted by 26% to $1,544.83 for some time. It's possible that those who purchased it for the sticker price could have buyer's remorse, especially when weighing up against similarly priced options like the Arm-based MacBook Pro M4 (starting at $1,599 for the 14-inch model) which came out after the Surface.

Consumers still don't know what they're getting themselves into

If you've kept a keen eye on the laptops over the past 12 months, you'll no doubt have heard the buzzword "AI PCs" when it comes to Snapdragon X Elite, Ryzen AI, and Lunar Lake laptops in 2024 and 2025. There has been a wealth of promise in advertising around all the new tools you can use through having a dedicated NPU, and how many TOPs of performance you can expect from Qualcomm and the competition in general.

However, to the average everyday computing consumer, there's been no solid throughline in what this tech does, and what it can offer you, especially in the jump from traditional x86 architecture and over to Arm-based silicon. Yes, the latter can have superior battery life and snappier performance in isolation, but it comes at the cost of compatibility.

Windows on Arm has come a long way in the last year as more devices (like the Microsoft Surface 7) adopt it over x86-based alternatives, but you're still not going to get to do everything that you could on that architecture. A lot of popular software is yet to offer an Arm version, and it's not been communicated to hardware manufacturers what's available and what isn't.

As such, users could buy the Microsoft Surface 7 expecting it to do everything their old laptop could, and then be disappointed when they run into problems. If you're asking people to spend $2,000, they will want everything to be perfect out of the box, and a common understanding of differing architectures (and compatibility overall) just isn't going to happen overnight.

You may also like...

• My days of forgetting to drink water are over thanks to this new app
• The dawn of PCIe 7.0 could mean faster SSDs for everyone... eventually
• Samsung's rumored smart specs could be coming out before the end of the year

• SoftBank confirms $6.5 billion deal to acquire Arm-based chipmaker Ampere
• The deal could align Ampere with Arm and previously acquired Graphcore
• The acquisition could put Arm in direct competition with its chip partners

We previously reported that Arm is considering making its own silicon, and that the British firm’s Japanese owner, SoftBank, was considering acquiring Ampere Computing, Arm's only independent server chip vendor, to make this a reality.

Ampere, based in Santa Clara, California and backed by Oracle, designs high-performance, energy-efficient processors optimized for AI and cloud workloads using the Arm compute platform.

An acquisition by SoftBank could shift Arm from simply licensing chip designs to manufacturing its own chips - putting it in direct competition with existing customers and expanding Arm’s footprint in the growing data center space.

Complementing Arm's design strengths

Up until now, the acquisition had only been a strong rumor, but it’s now finally confirmed that SoftBank will be buying Ampere Computing for $6.5 billion (approximately ¥973.0 billion).

The deal, announced on March 19, 2025, will make Ampere a wholly owned subsidiary through SoftBank’s investment arm, Silver Bands 6.

SoftBank Group already owns a majority stake in Arm, and an affiliated company, Arm Technology Investment Ltd, also holds an 8.08% stake in Ampere.

The transaction has already been approved by SBG’s Board of Directors but as always with these things it remains subject to regulatory approvals, including U.S. antitrust clearance and review by the Committee on Foreign Investment in the United States (CFIUS).

In a statement, Softbank said, “Through this strategic alignment following the transaction, Ampere’s expertise in developing and taping out Arm-based chips can be integrated, complementing design strengths of Arm Holdings.”

It added that Ampere is expected to collaborate with “group companies, investees, and business partners.”

SoftBank said the purchase will be financed through borrowings from Mizuho Bank and others. The Raine Group is serving as financial adviser, with Morrison & Foerster providing legal counsel.

Ampere will continue operating under its current structure until the deal closes, which is expected to be in the latter half of 2025.

SoftBank previously acquired UK-based chip designer Graphcore for between $400 million and $500 million.

That company was once considered a potential rival to Nvidia and AMD, but fell on hard times after failing to capitalize on the AI boom.

It's not too much of a stretch to think that Softbank could integrate Ampere with Arm and Graphcore to build a more unified AI compute strategy across its portfolio.

You might also like

• Arm's Japanese owner could buy Arm's only independent server chip vendor
• How Ampere Computing plans to ride the AI wave
• Arm may be planning to hike prices, has considered developing its own chips

• AMD's rumored "Sound Wave" chip is claimed to be an Arm-based APU
• It's believed to feature "RDNA 3.5+" CUs that are "enhanced" for machine learning
• MLID's sources allege the chip will feature 16MB Infinity Cache

Alleged new information has been uncovered about the rumored AMD "Sound Wave" APU, claiming it will be more towards machine learning, with "enhanced RDNA 3.5" CUs.

The news comes from reliable hardware leaker Moore's Law is Dead in a new YouTube video, with a long list of updated specs cited from "Late March 2025". It's a substantial update on the previous rumor that we covered back in March 2024.

Previously, it was believed that AMD's Sound Wave processor would be a Zen 6 mobile chipset. However, it's now claimed by MILD that the processor will be an Arm-based APU, similar to the likes of Qualcomm's Snapdragon X lineup instead.

Specifically, MLID's sources claim that AMD's Sound Wave APU will be built on a TSMC 3nm node, targeting 5-10W form factors, with two performance cores, four efficiency cores (six total), and 4MB of L3 cache. Crucially, it's alleged that there will be four RDNA 3.5 CUs with "improved ML Performance" which has been given the name of "RDNA 3.5+" unofficially, to denote a difference from the current RDNA 3.5 CUs available on chips.

Its capabilities for AI computing appear to be a main focus, as can be evidenced by the supposed "4th Generation AI Engine" backed by 16GB LPDDR5X-9600 RAM which is said to be for the "Standard" systems. MLID does not know if the newer RDNA 3.5 CUs will allow for FSR 4. However, gaming appears to be a secondary concern for this silicon.

An interesting development for the Sound Wave APU is that it will allegedly feature 16MB of Memory Access at Last Level (MALL) cache (known to AMD users as Infinity Cache) which is unusual for a 5-10W APU. It's curious because "Strix Point" does not have this, which is reportedly "documented" by MLID.

MLID later points out how the AI engine, CPU, and GPU will all work on the same controller, so having the extra bandwidth of 16MB Infinity Cache shared between them could be beneficial, even on such a modest chip.

AI computing without the power draw

While it's still early days for Sound Wave, all the rumors we've heard from the chipset sound as though it's primarily being geared around AI workloads without the (usually) high power draws of rival hardware. 16MB of Infinity Cache doesn't sound too dramatic an inclusion at first, however, that's before taking into account that the chip is only rated between 5-10W TDP.

Additionally, the RDNA 3.5+ CUs could be a big driving factor in AI workloads, depending on how optimized they really are for machine learning. With a 2026 release date claimed, it's unlikely we'll receive official word from AMD any time soon, but we now know it could be an optimized Arm-based APU rather than the chip that ushers in Zen 6 as we previously thought.

You may also like...

• Windows 11 should soon be faster at extracting files from compressed ZIPs
• Apple's facing a false advertising lawsuit over Bella Ramsey
• Google's NotebookLM adds Mind Maps to its research toolkit

• New dual-core MEGA.mini architecture boosts performance while saving energy
• Dynamic core allocation optimizes workloads
• Mega cores for complex tasks and mini cores for routine processing

At February 2025's International Solid-State Circuits Conference (ISSCC), researchers unveiled a new MEGA.mini architecture.

Inspired by Arm’s celebrated "big.LITTLE" paradigm, this universal generative AI processor, discussed at length in 'MEGA.mini: A Universal Generative AI Processor with a New Big/Little Core Architecture for NPU', an academic paper presented at the conference, promised a revolutionary approach to neural processing unit (NPU) design.

Arm's big.LITTLE architecture has long been a staple of efficient mobile and embedded systems, balancing high-performance cores with energy-efficient ones to optimize power usage. The MEGA.mini project seeks to bring a similar dual-core philosophy to NPUs, which are essential for running AI models efficiently.

MEGA.mini: A game-changing NPU design

This approach will likely involve pairing high-capacity "Mega" cores for demanding tasks with lightweight "Mini" cores for routine processing. The primary goal of this design is to optimize power consumption while maximizing processing capabilities for various generative artificial intelligence (AI) tasks, ranging from natural language generation to complex reasoning.

Generative AI tool workloads, like those powering large language models or image synthesis systems, are notoriously resource-intensive. MEGA.mini's architecture aims to delegate complex tasks to Mega cores while offloading simpler operations to Mini cores, balancing speed, and power efficiency.

MEGA.mini also functions as a universal processor for generative AI. Unlike traditional fastest CPUs that require customization for specific AI tasks, MEGA.mini is being developed such that developers can leverage the architecture for different use cases, including natural language processing (NLP) and multimodal AI systems that integrate text, image, and audio processing.

It also optimizes workloads, whether running massive cloud-based AI models or compact edge AI applications, assisted by its support for multiple data types and formats, from traditional floating-point operations to emerging sparsity-aware computations.

This universal approach could simplify AI development pipelines and improve deployment efficiency across platforms, from mobile devices to high-performance data centers.

The introduction of a dual-core architecture to NPUs is a significant departure from conventional designs — traditional NPUs often rely on a monolithic structure, which can lead to inefficiencies when processing varied AI tasks.

MEGA.mini's design addresses this limitation by creating cores specialized for specific types of operations. Mega cores are engineered for high-performance tasks like matrix multiplications and large-scale computations, essential for training and running sophisticated large language models (LLMs) while mini cores are optimized for low-power operations such as data pre-processing and inference tasks.

You might also like

• We've listed the best business computers available
• We've also listed the best AI video editors right now
• AI-powered no-code vs generative AI: who will win the race to build tomorrow's tech unicorns?

• Epic Games has announced a partnership with Qualcomm
• It will see Easy Anti-Cheat become available for Windows on ARM
• As well as Fortnite, it means the potential for hundreds of games to be playable

Epic Games has announced a new partnership with Qualcomm to support Easy Anti-Cheat for its Arm-based Snapdragon processors. This means Fortnite will finally be available on Copilot+ PC laptops utilizing the likes of the Snapdragon X family for the first time (which are some of the best laptops we've recently tested).

In a press release, Epic Games explained that it is working with Qualcomm Technologies to bring Epic Online Service's Anti-Cheat over to Windows on Arm, which is used by laptops like the excellent Surface Laptop. This decision was due to how Windows 11 on Arm has "transformed the PC landscape" with the company's processors representing a "rapidly growing segment of the PC gaming market".

There are some distinct architectural differences between x64 (64-bit) and Arm, with the latter being far less compatible with the majority of software, especially gaming. Windows on Arm has come a long way in the past year, but there's still plenty to improve. "Due to the way anti-cheat software works with x64 features, additional tooling is required for compatibility with these devices".

Crucially, the testing and implementation of Easy Anti-Cheat on Arm could open the door for the many other games that utilize the software for online play, with compatibility spreading across the Epic Games Store (and beyond). "In addition to releasing Windows on Snapdragon anti-cheat support for Fortnite, we will bring this support to developers through an Epic Online Services SDK release".

You're unable to install Fortnite without Easy Anti-Cheat enabled, as the game will not launch until EAC is located in the install directory. Easy Anti-Cheat is also available in nearly 500 games, including Apex Legends, Rust, iRacing, and Fall Guys, as well as more recent releases like Elden Ring. No timeframe has been offered for when the Epic Online Services SDK will roll out, however, you may not be waiting long to finally experience the popular battle royale game on ultrabook AI-powered hardware.

A step forward for accessible gaming on Arm

It's been an exciting time for laptops with the widespread implementation of Qualcomm's Snapdragon X processors inside some of the best laptops over the last year or so.

We've praised many of the productivity machines powered by the likes of the Snapdragon X Elite SoC, including the excellent Dell XPS 13 9350 (2025) and the Asus Vivobook S 15 Copilot+ just to name a few. While compatibility is still not on par with the traditional Windows x86 experience, the snappiness of performance (and leading battery lives) have made them a viable option for work.

Now, with the news that Epic Games is working on bringing Easy Anti-Cheat software to bring (arguably) the biggest game on the planet to Arm, we could see this as a breakthrough for gaming on ultrabooks.

Last year, we were blown away by the Snapdragon X Elite's gaming capabilities, which included running Baldur's Gate 3 in 4K. While the Qualcomm Adreno GPU baked onto the chip will not rival the discrete graphics found in some of the best gaming laptops, there's certainly potential for a playable future on the hardware, and it's especially exicting as modern MacBooks with M-series chips also use Arm-based architecture.

Some of the most popular PC games use Easy Anti-Cheat as standard, so Epic Games could be taking a massive step forward for accessibility, especially if the titles can be optimized for Qualcomm's hardware.

As we've seen with the rise of the Steam Deck, Asus ROG Ally X, and other handheld gaming PCs, developers have started catering towards integrated graphics solutions through the likes of the AMD Z1 and Z2 chips. While these are x86 devices, we could see future handhelds using Arm with a wider availability of games if more software can be adopted over time.

You may also like...

• AMD promises that RDNA 4 restocks are "priority number one"
• Four ways that ChatGPT can help you take control of your life
• I tried out Google Gemini's new AI image generation tool

• Arm could be gearing up to make its own CPUs, competing with its own customers
• Report says Arm is quietly poaching talent from its own partners
• SoftBank's $6.5 billion Ampere buyout could be the push Arm needs

Arm has built a solid business designing microprocessor architectures, such as ARMv9, and licensing the IP to big players like Apple, Qualcomm, Samsung, Nvidia, AMD, and AWS - but what it doesn’t do is make and sell its own chips, although this could be about to change.

It has long been rumored Arm is considering making its own silicon, and it now appears the British firm is finally taking steps in that direction.

Arm’s Japanese owner, SoftBank, is set to acquire Ampere Computing, Arm's only independent server chip vendor, for $6.5 billion, a move which could help Arm expand its footprint in the competitive data center and AI markets and pave the way for the company to build its own chips, but it’s a risky strategy. This quest for higher profits could put it in competition with some of its larger customers, including Qualcomm and AMD.

Looking to poach staff

Much of Arm’s success stems from the fact that its partners view it as a neutral force in the industry, rather than a competitor. If Arm begins to build its own silicon, that perception will change - and, as a Reuters report citing anonymous sources and a reviewed document suggests, it could upend the industry. Companies that license Arm's IP will have to consider whether they want to continue doing business with a direct competitor. RISC-V could really benefit here.

According to the report, Arm is actively seeking to recruit staff from its own customers, as well as competing with some of them for deals as it moves towards the goal of selling its own CPUs.

The document that Reuters reviewed reportedly states in November 2024, a recruiter working for Arm sent a message to an unnamed executive at an unnamed Arm client, seeking to hire employees to assist with its "transformation from solely designing processor architecture (IP) to also selling its own silicon, with a focus on driving AI enablement in the data center" and on other devices.

Two industry sources also told Reuters Arm recruiters have been reaching out to other chip designers in Silicon Valley for the same purpose.

Arm’s ambitions may extend further still. According to a research note published by J.P. Morgan analyst Harlan Sur, Broadcom has won a contract from Arm and SoftBank to develop a custom-built AI chip for data centers, potentially bringing Arm into competition with Nvidia.

A spokesperson for Arm declined to comment on the Reuters report.

You might also like

• Arm's Japanese owner could buy Arm's only independent server chip vendor
• Is an AMD Arm superchip in the works?
• Arm may be planning to hike prices, has considered developing its own chips

Mullvad VPN has become the latest VPN service to expand its range of device support with its arrival on the next-generation of Snapdragon PCs.

Users running Windows on ARM-based devices can now download and use a native Mullvad VPN application. The Swedish-based VPN provider has ensured that all the service's functionalities, privacy, and security features are included in the new app, so you can expect the same level of performance you'll have on other desktop platforms.

Mullvad's Snapdragon ARM VPN app supports both Windows 10 and 11 for users of ARM64 computers. "The installers are the same for both ARM64 and x86_64," explained Mullvad in its official announcement.

One of the best VPN providers currently on the market, Mullvad has invited anyone to share feedback about their new VPN app with their support team.

Snapdragon ARM devices and VPN support

Last summer's release of the powerful Snapdragon X Elite chip on Windows laptops came as exciting news for users. ARM PCs are expected to be faster than their predecessors, without compromising on battery life.

The problem is that most of the top VPN apps on the market still weren't ready for the next-generation PCs and didn't support Windows ARM devices at that time.

VPN providers couldn't simply emulate their VPN applications from other platforms. Developers had to build their VPN apps from scratch for it to be compatible with Windows ARM devices. Needless to say, it's taken some time.

Six months on and the VPN industry is slowly catching up.

The news makes Mullvad VPN currently one of just a few Windows VPN providers to offer a native application for Windows ARM devices.

Proton VPN unveiled its native app for Snapdragon PCs in November, a few weeks after TechRadar's top pick, NordVPN did the same. Private Internet Access (PIA), Windscribe, and Surfshark launched their ARM-native apps just before.

So far, ExpressVPN is the only VPN provider to come up with a workaround without creating a native app for Snapdragon PCs from scratch. The team at Express decided to take advantage of the latest release of Microsoft’s new Prism emulator and shape its proprietary VPN protocol, Lightway, accordingly.

Despite the technicalities, Mullvad VPN's new ARM-compatible VPN application further strengthens VPN support on the next-generation Windows computers. Users now have more freedom to choose the VPN provider that most suits their needs.

• SoftBank is reportedly close to buying Ampere
• The $6.5 billion deal could shake up AI and data centers
• There are questions about how it affects Arm’s neutrality and SoftBank’s long-term strategy

SoftBank, the Japanese owner of Arm, is close to acquiring Ampere Computing, a company that develops data center chips based on Arm's architecture.

In January 2025, BNN Bloomberg reported SoftBank was exploring a purchase of Ampere, which is backed by Oracle and private equity firm Carlyle Group, in a move widely seen as part of Softbank’s bid to challenge AMD and Intel’s dominance in the AI and data center chip market. However, as we previously reported, there were some hurdles to be overcome - Ampere was gearing up for an IPO, suggesting a desire to remain independent, while SoftBank would require approval from both Oracle and Carlyle Group before a deal could move forward.

However, a new Bloomberg report suggests that the deal is nearing completion and could be announced in the coming weeks. Sources familiar with the matter estimate its value to be around $6.5 billion, including debt. In 2021, SoftBank valued Ampere at over $8 billion during a proposed minority investment.

Strategic questions

Ampere develops high-performance processors for data centers, including the 192-core "Polaris" AmpereOne and upcoming 256-core "Magnetrix", using Arm’s chip designs. Acquiring Ampere could help SoftBank, which previously bought UK-based AI chip designer Graphcore, expand its footprint in the competitive data center and AI markets.

That said, the move does raise some interesting strategic questions. Arm, long dominant in mobile chips, has been working to expand into server and AI processing.

Ampere is one of the few firms independently developing Arm-based server chips. If the deal moves forward, it could impact Arm’s role as a neutral supplier of intellectual property. Integrating Ampere into SoftBank’s ecosystem may also shift competitive dynamics in the industry.

Market conditions add a further level of complexity to the benefits buying Ampere will bring to SoftBank. The semiconductor sector faces rising competition and concerns over slowing AI-related investments.

As Bloomberg notes, “Chinese startup DeepSeek, which unveiled an inexpensively produced AI model last month, has raised the specter that technology providers won’t be generating as much revenue as predicted.”

You might also like

• Why would Arm's owner want to buy a rival to AMD and Intel?
• How Ampere Computing plans to ride the AI wave
• Jaw-dropping 256-core CPU will debut in 2025 from Arm partner

• Softbank is apparently looking to buy Ampere Computing
• Japanese tech giant already owns Graphcore and Arm
• Experts say Softbank's ambition would be to rival global hyperscalers

SoftBank, the majority stakeholder in Arm, is reportedly in talks to acquire Ampere Computing (via Bnnbloomberg).

Ampere, known for its data center chips built on Arm architecture, has garnered attention for its advanced chips, including the 192-core Polaris and the upcoming 256-core Magnetrix.

The deal is currently uncertain but highlights SoftBank’s ambitions to challenge processor industry giants like AMD and Intel amid the ongoing boom in AI investment.

Ampere's strategic significance

Ampere Computing, backed by Oracle, is a pivotal player in the AI and data center chip market, and acquiring it would undoubtedly bolster Arm’s push beyond licensing chip designs to becoming a full-fledged chipmaker.

Ampere’s expertise could also enhance Arm’s push into the lucrative data center market, aligning with CEO Rene Haas’s vision of expanding the company’s presence in this sector.

Ampere processors are designed to handle demanding data center workloads, a critical capability as industries increasingly rely on AI-driven solutions. Combining Softbank Graphcore’s AI accelerators with Ampere’s processors could enable SoftBank to build competitive systems for that sector.

The potential acquisition comes with some caveats for either company. Ampere has been preparing for an IPO, indicating its interest in maintaining independence, while Softbank will have to convince Oracle and the Carlyle Group, both significant stakeholders in Ampere, to ratify the deal.

You might also like

• The AI boom is spurring an energy crisis: is there a sustainable way out?
• Data center demand is booming - but we don't have the power to deal with it
• We've also listed the best AI tools right now

• Qualcomm currently sells AI accelerator processors but has the CPU market in its sights
• Sailesh Kottapalli's expertise is in x86 architecture, not Arm
• But that didn't stop Qualcomm from bringing him onboard to lead its data center team

Qualcomm, renowned for its Snapdragon processors powering business smartphones and laptops around the world, has made a potentially key hire as it looks to challenge the likes of AMD and Intel in the processor market.

The company’s latest coup is hiring Sailesh Kottapalli, a former chief architect for Xeon processors and a 28-year Intel veteran.

Kottapalli joined Qualcomm as senior vice president in early January 2025, bringing extensive expertise in designing high-performance x86 server chips.

Kottapalli's move to Arm

Kottapalli wrote on LinkedIn that “the opportunity to innovate and grow while helping to scale new frontiers was immensely compelling to me—a once-in-a-career opportunity that I could not pass on.”

What makes the move significant, given Qualcomm’s reliance on Arm-based designs, is Kottapalli’s expertise in x86 architecture. His leadership could help bridge the gap between Qualcomm’s existing technology and the demanding requirements of data center CPUs.

A renewed push for the data center

Qualcomm had retreated from server CPU development back in 2018, but the company has now revealed plans to develop high-performance, energy-efficient server solutions tailored for data center applications.

This journey began with its Snapdragon X series for PCs, featuring custom Arm-based cores derived from its $1.4 billion acquisition of the startup Nuvia back in 2021, that built to a legal crescendo in a Delaware court in December 2024 when Arm alleged that Qualcomm's acquisition breached its licencing terms. Though a federal jury sided with Qualcomm, Arm is seeking a retrial.

For now, though, Qualcomm has been expanding its presence in the data center sector, with AI accelerator chips under the Qualcomm Cloud AI brand supported by industry leaders like AWS, HPE, and Lenovo.

You might also like

• Here's a list of the best mini PC for every budget
• We've also listed the best AI tools right now
• Spending on data centers hit a high in 2024, but it won't last

• Arm could be considering move to selling its own chips
• Details of proposed change revealed as part of Qualcomm’s court win over Arm
• Switch in strategy could prove incredibly lucrative for Arm

Update 01/16/25: This article has been updated with a correction to reflect that Qualcomm bought Nuvia, not Arm.

Original article follows.

Semiconductor technology supplier Arm, which almost certainly has its hardware somewhere in your business smartphone, is known for assisting companies to make their own mobile-suited processors, but that could soon change with a mooted shift into manufacturing its own chips.

A report from Reuters discusses Arm’s so called “Picasso” project; a bid to increase revenue by selling its own chips and competing with its own juggernaut customer base - including Qualcomm and Apple - that it normally sells ready-made Arm intellectual property to in order to assist with chip design.

Arm may also be planning to hike the royalty rates for those customers.

Arm v Qualcomm in brief

Details of the proposed strategy were revealed as part of Qualcomm’s court win over Arm in a royalty dispute brought and lost by the latter in December 2024.

This would nominally have included Qualcomm, but its purchase of the start-up Nuvia, in order to use the company's tech to produce its own chips and gradually break away from the prior agreement, led to Arm filing a complaint in US Federal Court in Delaware over a breach in licensing terms.

In the end, however, a jury ruled that Qualcomm’s Nuvia-tech chips were properly licensed, and ruled the company could continue to sell them as part of its route into the personal computing and AI sectors.

Arm’s future plans to “hose” its customers

Filings in those proceedings, that Reuters claims are still under court seal, reveal that the ‘Picasso’ plan for Arm to sell its own chips (or indeed “chiplets”) came at the behest of Arm CEO Rene Haas, who, before he even took on that role, had generally described the company’s biggest customers as “hosed” in an internal Teams message sent in December 2021.

In fact, court evidence suggests Arm executives had been discussing 300% royalty rate increases for its customers using Armv9 - its latest computing architecture, as far back as 2019, in a bid to boost the company’s smartphone revenue by $1 billion USD over the course of a decade.

It’s ultimately unclear as to whether this rate increase will happen or stick at all; use of Arm’s computing architecture doesn’t necessarily require its ready-made component blueprints.

And, as Reuters points out, many of Arm’s biggest customers-turned competitors could survive without those blueprints and still design their own parts.

You might also like

• Snapdragon CPUs could be over 50% of Windows PCs by 2029, Qualcomm’s CEO predicts – so should Intel be worried?
• Microsoft claims its servers were illegally accessed to make unsafe AI content
• We’ve also listed the best business laptops right now

• Qualcomm has officially taken the wraps off its new Snapdragon X CPU
• This processor is very power-efficient, as you might expect
• However, the low price is key here, promising to deliver Copilot+ PCs at around $600

Qualcomm has revealed a new Snapdragon chip to power more affordable Copilot+ PCs at CES 2025.

The new CPU is simply called the Snapdragon X – sitting underneath the Snapdragon X Plus and Snapdragon X Elite SoCs – and it’s designed to enable laptop makers to sell AI notebooks in the $600 range.

Qualcomm had already primed us for the arrival of this vanilla chip, but this is its official revelation at CES.

The 8-core CPU has the same NPU (with 45 TOPS) required to qualify as the engine of a Copilot+ PC, and Qualcomm claims that the processor “delivers up to 163% faster performance at ISO-power than our competitors.”

More broadly, Qualcomm asserts that: “Snapdragon X is an ideal solution for students, freelance workers, and budget-conscious consumers who need a reliable and powerful laptop that can keep up with their busy lives.”

We’re told that laptops with the vanilla Snapdragon X chip will be on shelves in early 2025, so presumably in the next month or two. Leading laptop makers will support the chip, as you might expect, which includes Acer, Asus, Dell, HP and Lenovo portables, in the $600 price bracket as mentioned.

Analysis: All part of the Copilot+ PC masterplan

Coming back to the performance claim Qualcomm aired above, the comparison is made against an Intel Core 5 120U – a Raptor Lake-U CPU from a year ago. It’s something of an odd choice, given that the Snapdragon X is a Copilot+ PC targeted SoC, and has that beefy NPU – and the Intel processor doesn’t – but it’s more about the laptop price bracket than anything else (and Lunar Lake doesn’t have an affordable equivalent, not yet anyway).

The Snapdragon CPU is clearly nicely power-efficient compared to that Intel chip, but this is just one benchmark, and as ever, first-party claims in press releases like this need to be taken with a good handful of seasoning. That said, it’s no surprise that a piece of Snapdragon X silicon would do well in terms of efficiency, as the rest of the family does a very good job on this front.

The key element here is the price, though. When Copilot+ PCs first emerged, they were exclusively Arm (Snapdragon) machines, and they were also very pricey. Over time, with the Snapdragon X Plus, those price tags came down (to more like $800), and with this new vanilla Snapdragon X, these AI laptops are now going to become truly affordable (and perhaps feature more heavily on our list of the best laptops).

Remember, if these Copilot+ PCs emerge at around $600, when they go on sale down the line – during this Black Friday (it’ll be here before you know it) for example – we could see some really tempting bargains.

Whichever way you dice it, this is a win for the consumer, and these much more affordable Arm-based laptops are doubtless part of Microsoft’s bold predictions for the success of Copilot+ PCs.

TechRadar will be extensively covering this year's CES, and will bring you all of the big announcements as they happen. Head over to our CES 2025 news page for the latest stories and our hands-on verdicts on everything from 8K TVs and foldable displays to new phones, laptops, smart home gadgets, and the latest in AI.

And don’t forget to follow us on TikTok and WhatsApp for the latest from the CES show floor!

You might also like...

• Razer Blade 16 announced at CES 2025 with an Nvidia RTX 5090 GPU – but Intel has been dumped
• LG and Samsung have put AI everywhere in their 2025 TVs – and Copilot's coming too
• Here's a list of all the apps that can run on the Qualcomm Snapdragon X Elite

• Roborock Saros Z70 unveiled at CES 2025
• First commercial robovac to have a retractible mechanical arm
• Due to go on sale in the first half of 2025

Robot vacuums are getting smarter and smarter, and today's best robot vacuums can do all kinds of magical things. Unfortunately, those innovations are often in software, meaning they can be a little lost on the average consumer. That is not the case with the Roborock Saros Z70 – which has a whole extra limb.

Yes, the Saros Z70, which has been officially unveiled at CES 2025, has a retractable robotic arm that takes care of tidying for you. It's called the OmniGrip 1.0, and it's a world-first for a mass-produced robovac.

I got a sneak preview of the Z70 a month ago, and it was absolutely mesmerizing. The bot looks much like any other robovac, except for a mysterious shiny hatch that sits on the upper surface. When requested, this hatch majestically opens and from it emerges a mechanical arm with a pincer on the end. The bot then scouts about for objects it has been told to clear – for example socks, or shoes – picks them up, and moves them to a new location. This pincer can extend and twist both horizontally and vertically to reach its prey.

You'll be able to set default behaviors in the app, so you might specify that everything the bot recognizes as a sock needs to go next to your laundry basket, for example. Similarly, discarded shoes could also be relocated to a specified 'shoes area', and ditto things like clothes or trash.

Not only is that a win in terms of being one less task for you to take care of yourself, it also means a more complete clean, as the bot has a clearer floor to vacuum.

If the idea of autonomous robots with limbs is making you feel a little uncomfortable, be reassured that Roborock has built in a few safety features. All of the OmniGrip 1.0's functionality is entirely disabled by default; the owner needs to specify exactly how it wants the arm to operate, via the app. You choose what objects the OmniGrip should interact with, and where they need to go. If you don't do that, it'll just stay behind its door permanently.

There's a child-lock and a safety stop button, plus the pincer has a relatively low-tension grip – the maximum weight is apparently roughly equivalent to a pair of shoes. That means that even if it did try to haul away your cat (and to be clear, it shouldn't), it wouldn't be able to.

Speaking of which, Roborock is apparently working on a function where the Saros Z70 will actually be able to play with your cat while you're out. If the cat deigns to get involved, of course.

There's always plenty of wacky tech at CES, so you'd be forgiven for assuming (as I did) that this was little more than an attention-grabbing prototype, rather than something consumer-ready. Apparently not. Pre-production samples are due to go on sale this month, retail samples are scheduled for March 2025, and open sales scheduled for H1 (i.e. in the first half of the year, before the end of June).

Is this the future of robot vacuums? Maybe…

I saw a slightly older generation version of the Z70 – in fact, not even the latest version at that point, but rather a touring press-event model that had picked up a war wound to its hatch as a result of some over-zealous handling in New York. Even without the latest software updates, and a few glitches during the demo, it was undoubtedly impressive, and certainly like nothing we've ever seen before in the world of robotic vacuums.

Is this the way robot vacuums are heading? Maybe. I'm encouraged by the fact that this bot is coming from Roborock, a brand that consistently impresses us with its new launches, and can be relied upon to deliver useful innovations that actually work. We loved the Roborock Qrevo Curv, which came with quad-bike-like suspension to bounce it over tall thresholds, for example.

While it's exciting to see something totally new, realistically, I suspect that for the moment, sister models the Saros 10 and 10R will be the ones that people actually buy (Update: to find out about those, you can read our Roborock Saros 10 review, our Roborock Saros 10R review, or our Saros 10 vs 10R comparison). The three options are being positioned as joint flagship models, and all three have the latest and greatest tech from the rest of the Roborock fleet.

The 10 and 10R are similar, but the former uses a traditional LDS module for navigation, while the latter uses the newer StarSight system (introduced with the Qrevo Slim, and notable because it doesn't rely on a raised central puck, meaning the overall bot profile can be slimmer). The Z70 is again similar, but with the all-important addition of the robo-arm.

TechRadar will be extensively covering this year's CES, and will bring you all of the big announcements as they happen. Head over to our CES 2025 news page for the latest stories and our hands-on verdicts on everything from 8K TVs and foldable displays to new phones, laptops, smart home gadgets, and the latest in AI.

And don’t forget to follow us on TikTok and WhatsApp for the latest from the CES show floor!

You might also like...

• 5 overdue robot vacuum innovations for 2025
• Is it worth getting a robot vacuum?
• The best cheap robot vacuum: budget-friendly home helpers

Nvidia is set to spring a whole bunch of new graphics cards on us next year, and it’s likely that we’ll see quite a few Blackwell GPUs early in 2025. On top of that, Nvidia apparently has ambitious plans to launch its own processor, an Arm-based chip for PCs that could challenge AMD and Intel (and Qualcomm) if it does indeed come to fruition later in 2025 as rumored.

Ready to explore the ins-and-outs of what Team Green is likely preparing for next year? Let’s cut to the chase and explore the likely possibilities in more depth.

Nvidia’s blitz of GPU launches at CES 2025

The rumor mill is pretty certain that we’ll see next-gen Blackwell GeForce graphics cards at CES 2025, and it’d be surprising if Team Green didn’t reveal those desktop GPUs during CEO Jensen Huang’s keynote (on January 6, 2025). Nvidia has all but confirmed this, at this point.

The question is: which Blackwell GPUs will be revealed? The grapevine seems most sure that two models will turn up initially, the RTX 5090 flagship and RTX 5080, both of which promise a major leap in performance.

Take the leaked specs with plenty of seasoning, but the RTX 5090 is said to be anything from 1.5x to 1.7x faster than the RTX 4090. That huge generational leap may come at a cost in terms of the next-gen flagship guzzling power to the tune of 600W, and there are concerns about the cost, too (with some kind of price hike definitely expected over and above the RTX 4090’s MSRP).

Those pricing worries apply to the RTX 5080, as well, with leaks suggesting Nvidia might pitch it as a ‘professional GPU’ to its partners, indicating the 5080 will be expensive. Spec rumors about the RTX 5080 have been a bit of a wild ride, with some controversy surrounding the apparent choice of 16GB for the VRAM loadout – which may not be the full picture, mind.

The RTX 5070 could also be revealed at CES 2025 alongside this pair of GPUs, at least according to some of the chatter filtering down from the rumor mill. However, other rumor peddlers have suggested that this graphics card might be launched slightly after the RTX 5090 and 5080, possibly in February 2025. Other theories still have indicated we could see the RTX 5070 Ti launch before the RTX 5070 (and that Ti spin could be seriously peppy).

At any rate, there’s something of a consensus that RTX 5070 models, probably two of them, will turn up early in 2025, and quite a few spec rumors have been aired – including, once again, that Nvidia could cheap out with the VRAM. This time Team Green is purportedly planning just 12GB for the RTX 5070, which would be a very disappointing choice in terms of future-proofing for the price tag the GPU will doubtless command.

So to recap, for next-gen desktop GPUs, we should see the RTX 5090 and 5080, possibly alongside the RTX 5070 and/or 5070 Ti, potentially all of which could be unveiled, or teased at the very least, at CES 2025.

The Consumer Electronics Show might also witness the revelation of Blackwell GPUs for gaming laptops. Nvidia could show off mobile graphics cards ranging from the RTX 5050 all the way up to the RTX 5090, with the RTX 5080 rumored to be a powerful beast of a notebook GPU. A further theory is that Nvidia will maintain the RTX 4050 mobile underneath the RTX 5050, to hopefully deliver a punchy enough offering for budget gaming laptops.

DLSS 4 could arrive to complement RTX 5000 GPUs

We might see Nvidia take the wraps off DLSS 4, or at least drop some new info on the frame rate boosting feature, also at CES 2025. It makes sense that the next-gen of DLSS would be tied to RTX 5000 GPUs exclusively, as was the case with DLSS 3 and current-gen Lovelace graphics cards, given it’s an obvious tactic to bolster sales of next-gen GPUs.

That said, we haven’t yet had any firm leaks on what DLSS 4 might do, which at this late stage, casts some doubt on a CES 2025 reveal. You can bet that Nvidia will push out DLSS 4 at some point next year, though.

Other Blackwell desktop graphics cards

After the dust settles from the early Blackwell GPU launches – the RTX 5090, 5080 and likely those 5070 models too – what else might we expect from Nvidia in terms of desktop graphics cards, later in 2025?

The missing piece of the puzzle is the RTX 5060, and rumors have been thinner on the ground with this GPU, making us believe that it might not turn up until later in 2025. Recalling Nvidia’s past form with launch schedules, it took seven to eight months for RTX 4060 models to arrive after the Lovelace flagship GPU first emerged.

However, of late, the RTX 5060 has popped up on the grapevine with suggestions that the graphics card might arrive in Q1 of 2025, maybe in March. The worry with the RTX 5060 is that Nvidia could stick with an 8GB helping of video RAM, the same as the RTX 4060. (Although the RTX 5060 Ti is also rumored to be inbound, and this could have 16GB, although obviously add plenty of seasoning with all this).

Nvidia’s CPU set to challenge AMD and Intel on their home turf

2025 might not just be a big year for graphics cards, with a host of next-gen GPUs, but Nvidia could also make a startling move on a very different front. That’s if reports of Team Green launching a new consumer-targeted Arm-based CPU next year turn out to be true.

This would, in theory, be a desktop chip to take on the might of AMD and Intel on their home turf processor territory, except Nvidia’s effort would be Arm-based (rather than x86), so a challenger akin to Qualcomm’s Snapdragon (Arm) CPUs that are currently making a big splash in Copilot+ PCs.

The Arm-based SoC would, of course, leverage Nvidia’s GPU expertise to provide peppy integrated graphics, and earlier rumors suggested Team Green might be working with MediaTek to produce these chips. With Qualcomm’s exclusivity on SoCs for Windows on Arm reportedly coming to an end, there’s theoretically room for Nvidia – or indeed other players – to move in.

Just like Qualcomm’s Snapdragon X range, Nvidia’s CPU would of course face the same issues due to Windows on Arm. Team Green’s chip would be hit with the inherent compatibility drawbacks pertaining to that platform in terms of performance overheads and software emulation (running x86 apps on Arm).

Arm has strengths, of course, as well – battery life, and affordability (in theory, as much cheaper Snapdragon X laptops are inbound for the start of 2025, certainly). Furthermore, app compatibility could become less of an issue as time rolls on. Why’s that, exactly?

Well, what we’ve got to remember is that Qualcomm has made a big play with its Snapdragon X chips already, and if Nvidia joins that Arm bandwagon as is rumored – and MediaTek also, in theory – that’s going to add a lot of weight to the push with Arm-powered PCs. With big things forecast by analysts for Arm devices, as the hardware starts to carve out more and more sales, with several tech juggernauts behind these PCs, developers writing more software for Arm natively will follow naturally as the market grows.

At least that’s the theory anyway, and doubtless the vision Nvidia has in its mind – if these rumors are correct. Also, we shouldn’t forget that Apple has marched down the Arm road with its own M-series chips, too.

The word is that we could see Nvidia’s Arm-based CPU arrive later in 2025, though it’s not clear if that would also mean a volume launch of laptops with the chips, which may not happen until 2026. Still, this is an apparent planned move for next year by Nvidia which will be potentially earthshaking in the laptop space, a prospect that might give Intel some pause for thought.

Concluding thoughts

It’s a good bet we’ll see Nvidia’s entire range of next-gen Blackwell desktop gaming GPUs emerge in 2025 – probably pretty early on. Although the rumors suggest we are in for some unfortunately underpowered video RAM configurations with some models, and perhaps some eye-watering price tags, too.

Those new GeForce boards will doubtless cause quite a stir in the GPU world, and there’ll be an interesting battle to watch against AMD’s RDNA 4 graphics cards – expected to launch in Q1 2025 – at the mid-range, and lower-end, of the spectrum.

The CPU pot could be even more vigorously stirred, mind, if Nvidia does unleash that theorized Arm processor – albeit in terms of what the further future might hold, rather than the PC landscape in 2025.

Arm laptops are going to be a longer-term game, but if the likes of Nvidia and MediaTek are going to shove their weight behind them alongside Qualcomm, those analyst predictions of Windows on Arm PC sales spiking strongly upwards start to look much more realistic. As the end of the decade rolls around, these devices could be a major force in the computing world, and that PC revolution of sorts could begin next year.

• Intel co-CEO has claimed a ‘large percentage’ of Arm-based PCs are returned
• Qualcomm disagreed claiming return rates are within industry norms
• Neither company backed up these assertions with any hard figures

It’s all kicking off between Intel and Qualcomm of late, as one of the new interim co-CEOs of Team Blue has taken a pointed swipe at Snapdragon-powered Copilot+ PCs. As you might imagine, Qualcomm was quick to return fire – and in no uncertain terms.

Windows Central reports that this started with Intel’s Michelle Johnston Holthaus, who currently helms Intel along with co-CEO David Zinsner, following Pat Gelsinger’s sudden departure.

Holthaus observed that Arm-based PCs, which are a major part of Microsoft’s drive with Copilot+ laptops, powered by Qualcomm Snapdragon X chips, might not be all that popular: “If you look at the return rate for Arm PCs, you go talk to any retailer, their number one concern is, ‘Wow, I get a large percentage of these back,’ because you go to set them up, and the things that we just expect don’t work.”

So, the claim here is not that Arm PCs are less reliable on the hardware front, but that software compatibility is the main sticking point, and reason why some buyers are returning their devices.

Qualcomm then issued a statement in reply to this, published by CRN, which read: “Our devices continue to have greater than 4+ stars across consumer reviews and our products have received numerous accolades across the industry including awards from Fast Company, TechRadar, and many consumer publications. Our device return rates are within industry norm.”

In short – unsurprisingly – Qualcomm was having none of this, with the spokesperson making it clear that return rates of Snapdragon X laptops are within the expected ‘industry norm’ range for PCs.

Analysis: Laptop realities and airy assertions

The gloves are off, then, well and truly. Intel’s been having a turbulent time this year, of course, so is this just a case of lashing out, deflecting, and drawing attention to the weaknesses of rivals?

Not as such, because I can see Intel’s point here. It’s easy to imagine a scenario where an average consumer buys an Arm laptop, gets it home, then finds out a game doesn’t run on it – or an app runs, but rather sluggishly – then they get frustrated and take it back to the shop, complaining that it’s ‘faulty’ and doesn’t work properly. Not the hardware, but the way in which the device copes with the software out there in the Windows ecosystem.

The root issue here is that Arm is a different architecture to AMD and Intel’s x86 chips, and most Windows apps are written for the latter – these are by far the dominant CPUs out there, of course, in Windows land.

Software (and game) developers have to code their products for Arm to run natively and fully performant, and if an Arm incarnation doesn’t exist, Arm-based PCs run the x86 version but have to emulate it (if it can work at all, and some software – and games in particular – can be a non-starter). That emulation involves overheads which can drag down performance somewhat.

Apple made the move to Arm, and its own M-series silicon, with its Macs, using the Rosetta translation layer for emulation (now considerably refined with version 2). Also, a major motivating factor for developers was that Apple was transitioning fully to Arm – so software and game makers coding for macOS had to get on board, or be left behind.

In Microsoft’s case, its Prism emulation – the equivalent of Rosetta – is still in its early stages, so not as refined. But more importantly, Microsoft is offering Arm PCs as an alternative, while still mostly being behind x86 chips.

So, there’s less incentive for developers to code Arm-specific apps or games for what’s a relatively small niche of laptops right now – and if that native software doesn’t come, the mentioned compatibility issues are in play, and put people off buying Arm notebooks.

Essentially, it’s a tricky situation: without the software support in place, growing the hardware base is more difficult, and without the hardware out there, motivating devs to write that native software is an uphill struggle.

So, it’s easy to imagine the rocky road Arm on Windows is currently travelling (and has always been on, in fact). And it’s true that what Intel is claiming is plausible to an extent – and somewhat backed up anecdotally by what I’ve read online. But we’ve got to be very careful about drawing conclusions based on what are theories, ultimately, and what Intel is saying here is all too airy.

We aren’t given numbers, percentages, or facts by Intel – just an assertion that retailers are getting large quantities of Arm PCs returned. Qualcomm’s reply is vague too, merely mentioning industry norms, without setting us (or Intel) straight in terms of an actual figure here.

Also consider that there’s no doubting that Snapdragon X-based laptops are excellent in some cases – TechRadar’s current best laptop is one of these machines (Microsoft’s Surface Laptop) – despite the weaknesses of Windows on Arm, and they are definitely present.

Analyst firms are also predicting some major growth for Copilot+ PCs with Arm chips, something Intel is doubtless not going to mention. But those are just forecasts – and x86 growth is supposedly going to be strong, too, although the claim is that Arm-based PCs could own 30% of the laptop market by 2028.

I can’t tell the future, but what I can tell you is that it feels rather unseemly for Intel to be throwing stones at this point, at the close of a pretty terrible 2024 for the chip giant in many respects. Qualcomm might well be tempted to ask what the return rate is on Intel’s 13th and 14th-gen desktop CPUs, following an actual hardware fault across two entire generations of silicon, perhaps.

You might also like...

• A Chrome-killing browser will be OpenAI's next big shot at Google, according to a new report
• Windows 11’s feature to allow typing with your Xbox controller has vanished in a blow for owners of handheld gaming PCs
• Windows 11 remains an unloved OS – but why won't people upgrade?

• Google Drive beta software now available for Snapdragon processors
• Only available for Windows 11 (and presumably onward)
• Marks a trend of increased support for ARM

Google has confirmed machines running Qualcomm's Snapdragon processors, previously unsupported by its cloud storage software, will now be able to download and run Google Drive natively, albeit currently in beta form.

The welcome news comes amid a rise in interest in Snapdragon processors, which are gaining more support from software makers.

Built specifically for ARM64, it offers users most of the same functions as the x86 version and offers a more efficient and seamless experience for Snapdragon users.

Google Drive app now available for Snapdragon processors

“We're excited to announce beta support for Drive for desktop on Windows 11 devices powered by Snapdragon processors. Compiled natively for ARM64, this release enables users to easily sync and store files online from Windows PCs powered by Snapdragon," the company posted in a blog update.

Although previous Google Drive apps have been available with versions prior to Windows 11, the ARM-based beta app is compatible exclusively with Microsoft’s flagship OS. It also requires Microsoft WebView2, which is usually preinstalled within the OS but could require users to download is separately during setup if that’s not the case.

Google emphasized, because this is a beta product, some bugs may be present, therefore Snapdragon users should not rely on it until general availability is confirmed. Any issues can be reported via Help > Send feedback within drive.google.com.

More broadly, Google isn't the only company optimizing its software for ARM hardware. Other apps, like Arc Browser and NordVPN, have also received recent support for the hardware.

The beta version of Google Drive for ARM devices is available to download now for all Google Workspace customers, Workspace Individual Subscribers, and users with personal Google accounts.

You might also like

• We’ve listed the best free office software
• PC sales are stalling as worries over costs and upgrade hassle hit hard
• Get an upgrade this Christmas from our selection of the best business laptops

• Nvidia’s Project Denver began as x86 but transitioned to Arm
• Insider reports legal constraints drove Nvidia's pivot
• The Arm-based Project Denver CPU debuted in 2011

During a technical session at the recent SC24 event, Dave Ditzel, founder of Esperanto Technologies, offered some fascinating insights into Nvidia’s early server processor efforts.

According to HPCwire, Ditzel, who was previously CEO of Transmeta, revealed that Nvidia’s first server CPU, Project Denver, initially started as an x86 CPU but transitioned to Arm due to legal constraints.

Ditzel says Nvidia’s shift to Arm was influenced by its licensing of Transmeta’s Tokamak technology, which could translate x86 code into a RISC instruction set.

Failed attempt to acquire Arm

IAs he explained, “Nvidia brought out a product called Denver. It was actually that same design. It originally started as an x86 [CPU], but through certain legal issues, had to turn itself into an Arm CPU.”

This decision, he said, laid the foundation for Nvidia’s alignment with Arm architecture. Tokamak, developed by Transmeta, was intended to be its third-generation x86 chip following the Crusoe and Efficeon processors. However, the project was never officially launched and was instead licensed to companies like Intel and Nvidia.

ntel, despite acquiring the design, did not announce a product based on it either. “You can guess as to all the reasons why or buy me a beer sometime,” Ditzel said.

Nvidia officially introduced Project Denver as an Arm-based CPU in 2011, later integrating it into its Tegra lineup. HPCwire reports that while there was initial enthusiasm around Arm servers, adoption was limited by challenges in the software ecosystem. Nvidia has since developed its Grace CPU and abandoned its attempt to acquire Arm after regulatory opposition.

Ditzel founded chip design firm Esperanto about seven years ago and because of his previous bad experiences with licensing x86, he opted for RISC-V because it was cheap and there were no legal concerns to get bogged down by.

“At least we have a playground where we can test some new things out, and some lawyer is not going to be ringing your bell,” Ditzel said.

More from TechRadar Pro

• Nvidia's most powerful chip produced gets tested
• This 192-core RISC-V processor could be AMD and Intel's biggest rival yet
• Google's super powerful Arm-based CPU is now available

• Nvidia APU rumored to equal RTX 4070 laptop GPU in performance
• Laptops are already being made with the APU, including from Alienware
• The Arm-based APU could outgun AMD’s Strix Halo by miles for efficiency

Nvidia’s apparent plan to make an Arm-based CPU for PCs – in conjunction with MediaTek – raised some eyebrows, certainly, and now we’re hearing a bit more about how powerful said APU from Team Green might be in terms of its integrated graphics.

For the uninitiated, an APU is an all-in-one chip (packing a processor, alongside an integrated GPU and NPU) and with a powerful enough graphics solution, these can make for a great engine for thin-and-light gaming laptops, dispensing with the need to have a separate graphics card in the chassis. (They’re also ideal for PC gaming handhelds, too).

Moore’s Law is Dead (MLID) has some fresh info on Nvidia’s rumored Arm-based SoC, namely that said integrated graphics will match the RTX 4070 laptop GPU (with the chip running at around 65W).

MLID further notes that laptop makers are already working with Nvidia to realize this APU, and there’s a rumored deal with Alienware on the boil.

We could see new entrants bursting onto our best gaming laptop list powered by the Nvidia APU as soon as the end of 2025, or the leaker still believes that’s a possibility – though we may realistically be waiting until 2026.

It is, after all, a big move for Nvidia, and one that Team Green and its partners will want to get right.

Analysis: The battle against a Halo product

If you recall, AMD also has powerful APUs on the cards, and is closer to launch than Nvidia – in fact, Team Red is expected to reveal new Strix Halo chips at CES 2025 (alongside a whole heap of other goodies).

Interestingly, the Strix Halo flagship APU has also been strongly rumored to have seriously peppy integrated graphics, and comparisons have been drawn to the discrete RTX 4070 laptop GPU here, as well – or that Team Red’s top chip might even be slightly more powerful.

Still, in Nvidia’s favor, it’s not clear if that flagship APU will serve gaming laptops, or just workstations, as we’ve discussed in the past. And even if the Nvidia APU is a bit behind in terms of the performance of its integrated GPU, it’s targeting much better efficiency. (Strix Halo rumors point to 120W for the flagship, and as we see here, 65W is mentioned for Nvidia’s effort – that kind of power difference is huge in the world of portables).

We can also expect a seriously powerful NPU from Nvidia, and AMD Strix Halo for that matter.

What’s going to be the problematic area for Nvidia is that its APU is Arm-based, so will run Windows on Arm with its inherent drawbacks, as seen with current Snapdragon SoCs. Whereas AMD’s Strix Halo is a traditional x86 chip and won’t have any compatibility (or performance overhead) worries – though come 2026, the picture for Arm chips in that respect may have changed a great deal. Microsoft is certainly betting on that, and Nvidia too, no doubt – if this rumor is right.

More competition in the mobile CPU arena will be a prospect that Intel isn’t keen on. Team Blue has a big business in producing laptop CPUs, an area it continues to be strong in despite all the woes around its desktop processors. Is that going to change with Nvidia’s entry into the laptop CPU market? It’s not what Intel needs right now, that’s for sure.

Via TweakTown

You might also like...

• Windows 11 remains an unloved OS – but why won't people upgrade?
• Report suggests 2025 will be the year Windows on Arm growth goes in overdrive
• There’s never been a better time to buy AMD’s flagship graphics card – as prices plummet ahead of inevitable Nvidia RTX 5000 reveal

• Fujitsu and AMD partner to develop energy-efficient HPC/AI platforms
• Partnership aims to broaden access to AI, support open-source
• Monaka chip features 288 cores, 2nm process, Armv9-A architecture

Fujitsu and AMD have announced a new strategic partnership focused on developing HPC and AI platforms.

This collaboration will combine Fujitsu’s ARM-based processor technology with AMD’s GPU expertise, aiming to build energy-efficient and open-source solutions addressing the growing demand for diverse, cost-effective computing architectures.

The partnership, formalized in a memorandum of understanding, covers joint efforts in technology development, commercialization, and ecosystem expansion, with a goal of creating powerful computing platforms by 2027.

Instinct accelerators

Central to this partnership is Fujitsu’s next-generation Arm-based processor, the Monaka chip, set for release in 2027.

As we first reported in July 2024, Monaka will feature Armv9-A architecture, a 2nm process for high performance and power efficiency, and a 288-core structure (144 cores per socket). Notably, Monaka moves away from high-bandwidth memory in favor of PCIe 6.0 (CXL3.0), enhancing scalability and connectivity. This processor is expected to build upon Fujitsu’s work with the A64FX chip, used in the Fugaku supercomputer, and could support the upcoming FugakuNEXT project planned for 2030.

AMD will support Monaka with its Instinct accelerators, providing customers with flexible options for handling massive AI workloads while optimizing data center costs. The collaboration will also make use of AMD’s ROCm software stack and Fujitsu’s proprietary software, developing an open source ecosystem to accelerate the development of AI and HPC applications.

This joint venture will also involve global marketing efforts and customer engagement, as well as a shared customer center to support the development and implementation of AI technologies.

“By combining AMD’s innovative GPU technology with Fujitsu’s low-power/high-performance processor Fujitsu-Monaka, we seek to create an environment in which more companies will be able to utilize AI while reducing the power consumed by data centers," noted Vivek Mahajan, CTO of Fujitsu.

You might also like

• Fujitsu is planning the most powerful CPU ever for its supercomputer
• AMD head outlines the challenges for data centers
• AMD launches a new single-slot GPU for data centers

It’s no secret that 2024’s batch of AI PCs owes much of their success to the excellent batch of Qualcomm Snapdragon chips, which has the best version of Windows Arm running as the OS. However, a nasty split between the two could jeopardize everything, especially for consumers.

According to a breaking news report from Bloomberg, Arm canceled a “license that allowed longtime partner Qualcomm Inc. to use Arm intellectual property to design chips.” The company issued Qualcomm a 60-day notice of the cancellation of said license as required by law. The report stated that the split was caused by a breach of contract back in 2022, which escalated a legal dispute over vital smartphone technology.

Qualcomm responded in a statement, writing that Arm is trying to interfere with its CPUs and increase its royalty rates and that its claims are “baseless.”

This news comes as Qualcomm announced new automotive and mobile chips at the yearly Snapdragon Summit in Hawaii, and in the same year, the brand Qualcomm Snapdragon Elite chips launched with the latest batch of AI laptops.

Why this legal battle is so disastrous

But this legal battle isn’t just a lovers’ quarrel between those major tech companies; it could have widespread ramifications for consumer tech at large. Windows laptops like the Microsoft Surface Laptop 7 received massive rejuvenation through the Snapdragon Elite, which is made from tech provided by Arm.

This chip took one of the most inconsistently performing laptop lines and turned it into an absolute powerhouse in terms of performance and battery life. Further, its Windows Arm operating system, while still not at the level of Windows 11, has improved by leaps and bounds to the point where it’s near perfect for productivity work and reasonably competent for creative work and gaming.

Arm and Qualcomm are also vital as they’ve generated competition with Intel, AMD, and Microsoft in the AI PC and OS market. This has given consumers more variety while also keeping laptop prices more affordable than their competitors. If it can continue to gain momentum in the market, both the chips and OS could become a true threat to the big dogs as the tech behind them is refined.

Why would Arm move to cancel Qualcomm’s license?

While it’s impossible to know the full story behind this legal matter, Qualcomm’s statement in response and relevant Arms-related news suggest possible hints.

Qualcomm stated, "This is more of the same from ARM—more unfounded threats designed to strong-arm a longtime partner, interfere with our performance-leading CPUs, and increase royalty rates regardless of the broad rights under our architecture license.” According to Qualcomm, Arm is trying to acquire more money from the former by forcing an increase in royalty rates.

This also coincides with the news that Intel sold its 1.18 million-share stake in Arm back in August 2024. While it’s considered unlikely that the sale was attributed to any issues with Arm (it’s more related to streamlining operations amidst financial trouble), it’s still a possibility that there could have been something involving Arm’s finances. Otherwise, it should be a valuable asset worth keeping hold of.

Regardless of the reason, this legal dispute is akin to playing dice with the future of Windows laptops. What once seemed like a brighter future for laptops could stand to crumble if this ends with the permanent split between two tech companies that together have been successfully challenging the industry giants.

You might also like

• Qualcomm reveals major layoffs, hundreds of workers let go
• Forget Snapdragon 8 Gen 4 – Qualcomm’s Apple A18 Pro rival is the Snapdragon 8 Elite
• I saw a Snapdragon X Elite laptop run Baldur’s Gate in 4K, yet Qualcomm insists it’s 'not a gaming laptop'

Semiconductor giant Arm is reportedly looking to bypass its Chinese subsidiary, ArmChina, by offering IP licenses directly to Chinese customers,

According to DigiTimes reporting, Arm China, established in 2002, remained a foreign-owned enterprise in the nation until 2016 when a majority (51%) stake was sold to a Chinese investor consortium.

At the time, it was hoped the sale would help Arm comply with local regulations better as well as adapt to China’s unique needs. Since 2016, Arm China has acted as an intermediary, earning itself a margin by licensing Arm’s IP to Chinese customers, however tensions between Arm and its Chinese subsidiary have been ongoing for several years, centered mostly around who controls the division.

Arm wants to abolish Arm China

The British company now appears to be exploring ways to circumvent the joint venture model by selling directly to China. By doing this, it would be able to increase its gross margins by eliminating the need to share projects with Arm China.

Despite tensions, Arm China continues to contribute pretty significantly to the overall business. It generated $122.6 million in the second quarter of 2024, accounting for around 13% of the overall company’s revenue.

Moreover, Arm China has been busy developing its own processor IP. In September 2024, it launched its domestically-approved Linglong graphics and video processors, and has also reportedly been exploring general-purpose GPUs for AI and HPC, reflecting China’s push for semiconductor self-sufficiency in light of ongoing geopolitical tensions and subsequent restrictions and sanctions on imports and exports.

In a letter to shareholders following its most recent quarterly earnings report, Arm CEO Rene Haas and CFO Jason Child wrote: “We expect the migration of AI from the cloud to edge devices to drive even higher compute demand per device over time and Arm’s unprecedented share across all end markets means that Arm is uniquely positioned to benefit from this migration, which will increase royalty revenue per chip.”

TechRadar Pro has asked the company about plans to sell direct to the Chinese market, but the company did not immediately respond.

More from TechRadar Pro

• These are the best cloud storage management services
• Fancy an upgrade? Here are the best business laptops and best mobile workstations
• Huawei is dodging advanced chip sanctions through suppliers, claims US Select Committee

US trade restrictions have created significant obstacles for Chinese companies, limiting their access to advanced AI hardware needed to remain competitive globally.

Nvidia’s H20 GPUs, scaled-down versions of the powerful H100, were developed to meet export control requirements but still come with a hefty price tag of around $10,000 per unit.

Even at that price, the availability of these GPUs is limited, compounding the difficulties Chinese companies face. This shortage has fueled a thriving black market for Nvidia’s high-end chips, such as the H100 and A100, where prices continue to rise due to overwhelming demand. However, global companies, particularly ByteDance - the parent company of TikTok, already under intense scrutiny in the US - cannot afford the legal and reputational risks associated with engaging in such illicit markets.

Two AI chips

ByteDance has made significant investments in AI, reportedly spending over $2 billion on Nvidia's H20 GPUs in 2024, and now according to The Information, the company is looking to develop its own AI GPUs to reduce dependency on Nvidia.

The report adds these chips will include one designed for AI training and another for AI inference, and both will be produced using TSMC’s advanced N4/N5 process, the same technology used for Nvidia’s Blackwell GPUs.

Broadcom, recognized for its AI chip designs for Google, will reportedly lead the development of these GPUs, which are expected to enter mass production by 2026. While several Chinese companies have developed their own AI GPUs to reduce reliance on Nvidia, most still depend on Nvidia’s hardware for more demanding tasks. Whether ByteDance can fully transition to its own hardware - and whether it would want to - remains to be seen.

The move will certainly not be without challenges. As Tom's Hardware notes, “The company now relies on Nvidia's CUDA and supporting software stack for AI training and inference. Once it goes with its AI GPUs, it must develop its software platform and ensure its software stack is fully compatible with its hardware."

More from TechRadar Pro

• International standard for LLMs to be developed by US-China tech coalition
• China leads in up to 89% of tech research, study shows
• China leads in chipmaking equipment spending

Valve appears to be testing a version of its Proton translation layer – which facilitates playing Windows games on Linux – built for Arm processors (as opposed to x86 CPUs).

Tom’s Hardware spotted that in Valve’s recent round of testing a large number of PC games with Proton, there seems to be an Arm64 version of the tool (called ‘proton-arm64ec-4’) which is also being tried out with a whole bunch of titles.

Tom’s notes that this includes the likes of Left 4 Dead 2 and Shadows of Mordor, to pick a couple of more high-profile examples.

For some, this hints at the possibility Valve could be considering releasing a version of SteamOS for Arm64, meaning a potential Steam Deck running with an Arm chip as its engine – like the silicon that powers Copilot+ PCs, the new Snapdragon X processors – as opposed to an x86 AMD (custom Zen 2) CPU, which powers the Deck currently.

Analysis: Eyeing the potentially explosive growth in Copilot+ PCs?

What would the point be of such an Arm-toting Steam Deck? Well, we could be looking at a more wallet-friendly Deck with better battery life, in theory. Although that said, x86 is making big strides in terms of improving battery life as it is (just look at Intel’s new Lunar Lake mobile chips). Also, as Tom’s points out, the entry-level Steam Deck is already pretty affordable as it is – particularly when it’s on sale.

Rather than this being part of the future plans for Valve’s gaming handheld, what’s probably more likely is that Valve is simply looking more broadly to make a version of Steam for Arm-based Windows 11 laptops like the aforementioned Copilot+ PCs.

After all, we’ve been hearing quite a lot about how the sales spike in these Copilot+ PCs will be huge in 2025, if analysts are to be believed (add seasoning, naturally). So, perhaps Valve is eyeing that potential casual gaming market.

Of course, Arm-based Snapdragon X-powered Copilot+ laptops already do emulation of x86 PC games, using Microsoft’s Prism technology – but while this works well enough for a good number of titles, there have been a surprisingly large amount of games left out in the cold. Gamers feel that Microsoft hasn’t fully delivered on its big gaming promises with Snapdragon chips.

At any rate, what’s happened here only represents an early hint that Valve is working on developing Proton for Arm chips, and nothing may come of it. In short, we need to watch this space, but there’s bound to be plenty of speculation about where Valve’s going with this – if anywhere.

You might also like...

• Affordable Copilot+ PCs might be coming – but leak shows their 8-core Snapdragon CPU won’t cope with gaming
• Windows 11 remains an unloved OS – but why won't people upgrade?
• I've been using a work laptop with a dedicated Microsoft Copilot key for months now, and it's made me realize one thing

SolarWinds has recently patched a critical severity vulnerability in its Access Rights Manager (ARM) program, which allowed threat actors to run malicious code remotely.

Access Rights Manager (ARM) is a piece of software designed to help organizations manage, monitor, and audit user access rights across their IT systems.

In a security advisory, SolarWinds said ARM was vulnerable to a “deserialization of untrusted data remote code execution flaw,” which essentially means that the software was not validating user-supplied data properly. That data, if malicious, could be abused in cyberattacks.

Patch available

The bug is tracked as CVE-2024-28991, and has a severity rating of 9.0/10 (critical). It was first spotted by security researchers from Trend Micro’s Zero Day Initiative (ZDI), The Hacker News reports, who gave it a 9.9 severity rating. One of the reasons is because of poor authentication practices: "Although authentication is required to exploit this vulnerability, the existing authentication mechanism can be bypassed," the ZDI said.

If you are using ARM, you should make sure you’re running version 2024.3.1. SolarWinds says there is no evidence that the flaw is being abused in the wild, but is still advocating caution and advising users to patch without delay.

Although SolarWinds is a major IT services provider, popular among enterprises, it rose to infamy back in 2020 when an upcoming patch for one of its products was compromised by ransomware hackers. Unknowingly, the company pushed a tainted update to countless customers, compromising many of them in the process, and resulting in sensitive data being stolen from dozens of high-profile organizations.

Last year, the US Securities and Exchange Commission (SEC) decided to sue SolarWinds for the incident, claiming its executives knew about its security shortcomings. Instead of notifying investors and users, the lawsuit alleges, the execs kept the information for themselves and even tried to convince everyone of the opposite.

Via The Hacker News

More from TechRadar Pro

• The inside story of the infamous SolarWinds hack
• Here's a list of the best firewall software around today
• These are the best endpoint security tools right now

One of the best VPN services on the market has just become an early mover in delivering a VPN app compatible with next-generation Windows devices.

ExpressVPN has closely collaborated with Qualcomm (developers of the Snapdragon X Elite processor) and Microsoft to accelerate the delivery of a fully ARM-compatible Windows VPN app. The team eventually found a revolutionary solution to avoid building a native ARM application from scratch.

As of September 4, the ExpressVPN ARM app is now available in beta, with a full production launch set for the middle of October.

A hybrid solution 

While new Microsoft Copilot+ PCs powered by Snapdragon can deliver better battery life and super-fast performances, they have been worryingly lacking in VPN compatibility. This is why most VPN apps still aren't ready to support Qualcomm's high-performance ARM-based processor. This was something that the ExpressVPN team knew it needed to fix

"Snapdragon is one of the most exciting things to happen to the Windows platform in recent times, and we saw this as an opportunity to showcase a high-performance VPN solution tailored for ARM's capabilities," said Peter Membrey, Chief Engineering Officer at ExpressVPN.

The main issue with VPN compatibility on ARM Windows machines was that emulation posed serious performance limitations that could affect VPN apps' functionality. Now, ExpressVPN has found a way to fix this without the need to build a native ARM application from scratch.

As the provider explains in a blog post, the release of Microsoft’s new Prism emulator with the Windows 11 24H2 upgrade has made it possible for emulated apps to match the quality of a native build. The team then decided to harness this new opportunity by shaping its in-house built WireGuard-inspired VPN protocol, Lightway, accordingly.

"By leveraging Microsoft's enhanced emulation and Lightway's adaptable architecture, we were able to solve initial limitations and deliver this build efficiently," Membrey explains.

The provider promises such a hybrid approach can offer users "the best of both worlds," namely the full security of a premium VPN service and accelerated development time.

As we mentioned earlier, just a handful of top VPN services are available for ARM Windows devices at the time of writing. Private Internet Access (PIA) launched its ARM-native app about a month ago. Windscribe and Surfshark also have ARM-native Windows VPN apps, while the likes of NordVPN, and ProtonVPN are currently working on their versions.

ExpressVPN now believes that its innovative approach to ARM compatibility marks a significant milestone in VPN technology. This could then support other providers fasten their transition to ARM.

Membrey said: "We’re proud to be one of the first major VPN providers to demonstrate our commitment to this platform, and we will continue to optimize Lightway and our software to ensure that users can enjoy the best that Snapdragon, Microsoft, and ExpressVPN have to offer."

We test and review VPN services in the context of legal recreational uses. For example:

1. Accessing a service from another country (subject to the terms and conditions of that service).

2. Protecting your online security and strengthening your online privacy when abroad.

We do not support or condone the illegal or malicious use of VPN services. Consuming pirated content that is paid-for is neither endorsed nor approved by Future Publishing.

Microsoft’s line-up of Copilot+ PCs is designed to run AI tasks in a superfast manner, along with providing welcome features like all-day battery life, but what about gaming?

Okay, so these are AI laptops, and not designed for gaming, that’s true - but what about dabbling in some PC games casually on your notebook? While it’s not unheard of for Microsoft to make questionable software design decisions, a new report gave me pause for thought, asserting that the first wave of Copilot+ PCs - powered by Arm-based Snapdragon X chips - don’t perform well even for more casual gaming in some cases.

Indeed, according to the report from The Wall Street Journal (flagged up by MacRumors), those Snapdragon X-powered Copilot+ laptops might struggle to run a large number of PC games. In extensive testing, the WSJ found that of 1,300 PC games, only half of them ran smoothly without noticeable issues or glitches.

While problems with running PC games on Arm-based Copilot+ laptops have been well-documented since the launch of these devices, the failure rate mentioned here is an eye-opening one.

Compatibility challenges for Arm devices

The problem, as you may be aware, is that Arm is a different architecture from traditional x86 chips that PC games are written for. So, to run those games on an Arm device, Microsoft needs to employ a software layer to translate x86 games (and apps) to get them working using a system called Prism.

Prism works as an emulation layer, constantly converting instructions designed for Windows devices with x86 chips into instructions for devices with Arm chips. Apple has a similar tool for its Macs that you’ve probably heard of, called Rosetta (we’re now on Rosetta 2, in fact), allowing Macs with M-series (Arm-based) chips to run x86 apps.

The trouble is that emulation inevitably introduces performance overheads, and the whole process can lead to compatibility problems in some cases, with Prism apparently not faring that well on this side of the equation as the WSJ report makes clear. On Arm-based Copilot+ PCs, as mentioned only half of tested games ran without nasty glitches or problems launching. Some popular games like Fortnite and League of Legends can’t be made to work on Arm-based devices at all right now (anti-cheat systems are a particular bugbear, as we’ve heard in the past).

Clearly, this is an issue that Microsoft must continue to work on. I’m stating the obvious here about apps and games, but being able to choose freely among the existing releases out there is a huge part of what makes a computing device usable for any given person.

Copilot+ PC expectations and the future

For now, Microsoft has told users that if they want to play games with demanding graphics, a Copilot+ PC may not be for them. And to be fair to Microsoft, this is not one of the core aims of a Copilot+ laptop, and it was never intended to be - what is slightly worrying is the gravity of the problem as painted by this new report from the WSJ.

I hope Microsoft will address these issues with gaming on Arm silicon and continue to improve its Prism emulation tech going forward, which to be fair, is still in its early days (Rosetta is now on version 2, as noted). In the world of Copilot+ PCs, though, we are going to start seeing laptops with x86 chips inside that won’t have these compatibility issues - notebooks built on Intel’s Lunar Lake processors, and AMD’s Ryzen AI 300 silicon. This is to say this is not an issue with Copilot+ PCs as such, but with Arm chips - at least for now.

YOU MIGHT ALSO LIKE...

• Affordable Copilot+ PCs might be coming – but leak shows their 8-core Snapdragon CPU won’t cope with gaming
• CoPilot+ PCs could finally get more affordable soon – here’s what's coming
• I've been using a work laptop with a dedicated Microsoft Copilot key for months now, and it's made me realize one thing

The rumor mill reckons that more affordable Copilot+ PCs are set to debut imminently, but the cheaper Snapdragon X Plus 8-core chip they are built around can’t do much on the gaming front in comparison to the beefier Arm-based silicon we’ve already had from Qualcomm, going by a fresh leak.

Wccftech (via Techspot) highlighted leaked benchmarks for the Snapdragon X Plus chip with 8-cores which are somewhat disappointing, particularly for PC gaming.

The benchmarks were carried out on an Asus laptop (ProArt PZ13) and show that the integrated GPU in the 8-core Snapdragon hit a score of 1,025 in 3DMark Time Spy, and a result of 916 for graphics – compared to 1,900 overall for the Snapdragon X Elite.

Measuring frame rates in Shadow of the Tomb Raider, the 8-core Snapdragon could only manage 18 frames per second (fps) on average at 1080p with low graphics settings, and just 12 fps with high graphics details. That’s too slow to be palatable, whereas the Snapdragon X Elite can average 33 fps with high details, a playable frame rate.

In other benchmarks away from gaming, the inbound Snapdragon X Plus looks a little slow in some respects, too. Cinebench 24 witnessed the CPU hit 102 points and 649 points for single and multi-core performance respectively, which is 5% and 20% slower than the Snapdragon X Plus 10-core model in Microsoft’s Surface Copilot+ PCs.

Analysis: Gaming is not the focus here

This 8-core model is expected to power new laptops that are rumored to drop at the IFA 2024 show that’ll kick off later this week. The CPU will usher in cheaper Copilot+ laptops apparently starting from $800 in the US (and presumably in line with that in other regions), which is obviously a good thing – these AI laptops are pricey as it stands.

As we’ve seen, gaming performance compared to the top Snapdragon X chips is disappointing, but then to be fair, Qualcomm's own execs say that these laptops aren’t made for gaming. Being able to play some games at decent frame rates on a Copilot+ laptop is a bonus, really – though a nice one, admittedly – and if compromising on that front for a more affordable device is the price you must pay, that’s not going to bother a good many folks.

In short, if you want a side-serving of solid gaming performance with your Qualcomm-powered Copilot+ PC, then you’re going to have to pay for it. But those looking for a productivity workhorse or everyday notebook for daily computing use doubtless won’t be phased by any shortfall on the gaming front, as long as the 8-core Snapdragon provides serviceable performance with their preferred apps – which it likely will do, even if it’s lagging somewhat in the multi-core stakes.

Those looking for more powerful devices should also bear firmly in mind that we’ll see Intel Lunar Lake-powered Copilot+ PCs at IFA 2024, and these could be pretty peppy indeed (with a secret weapon for gamers, incidentally). Intel has certainly been boasting about them from the rooftops, and this range will be Team Blue’s first set of mobile chips powerful enough to qualify for the Copilot+ laptop designation, going up against Qualcomm’s Snapdragon chips and AMD’s Ryzen AI 300 series. The AI laptop scene is about to get considerably more competitive.

You might also like...

• Here's a list of all the apps that can run on the Qualcomm Snapdragon X Elite
• Microsoft's Copilot+ AI PCs aren't all that special right now, but there's one major reason why that's about to change
• You can now run Windows 11 seamlessly on Apple silicon Macs

Microsoft’s Copilot+ PCs are set for booming sales if some fresh predictions from an analyst firm are correct – and these laptops could threaten Apple’s MacBooks in the future.

Obviously, these figures are just an estimation, involving a crystal ball as well as educated guesswork, and the fact that the boom in AI PC sales is already underway. In case you missed it, in Q2 of 2024, AI PCs – defined as any computer with an NPU (Neural Processing Unit to accelerate AI workloads), not a Copilot+ PC, which needs a more powerful NPU, we should clarify – hit sales of 8.8 million according to Canalys, another analyst outfit, doubling up on Q1.

As Wccftech reports, the new prediction from Omdia suggests Microsoft’s ARM-based Copilot+ PC sales will reach 800,000 units this year, and that will shoot up by over 500% to 5.2 million units in 2025. That’ll accelerate further to a total of 47.7 million units of these laptops in 2028, we’re told.

When considering all AI PCs, meaning ARM-based devices and Intel or AMD AI laptops, we’re looking at a rather staggering 34.2 million units shipped in 2025, which with the ARM-based machines on top, gives a total of 39.4 million units. By 2028, that total of ARM and Intel/AMD AI PCs will be 155.4 million units – if this educated guesswork pans out, of course.

Analysis: Smack-Book talk

The long and short of it is that Copilot+ PCs are expected to be very popular – to an eye-opening extent – going forward. Of course, after this year, most AI laptops of the Windows 11 variety are going to be Copilot+ PCs (as the majority of the chips with NPUs from Intel, AMD and Qualcomm will qualify for that designation, having a suitably powerful NPU).

Omdia gives us some reasons why we might see such an explosion in popularity, namely that AI will help people be creative, as AI can take on image or video creation tasks, giving anyone the ability to be a content creator. Whether that’s a good thing, or not, regarding a flood of AI-generated content and all the various worries therein, very much remains to be seen – but undeniably it should be a major selling point for Copilot+ PCs, we’d agree.

AI PCs are also expected to make a big impact in the field of education, in terms of helping not just students, but teachers in preparing materials for their pupils (all kinds of coursework media will be a snap to create, of course).

Omdia further points out that if Microsoft equips ARM-based Copilot+ PCs with a peppy discrete GPU, these devices could be a major problem for Apple’s MacBooks.

The analyst firm observes: “If Microsoft introduces an AI gaming or AI creator laptop with a 45-80 watt thermal design power (ARM) chip with a discrete graphics card (Discrete-GPU), and compresses the thickness of the laptop to close to the institutional thickness of Apple’s MacBook Pro (within 16 mm) through the thin and light function of OLED displays, it will cause a crisis in Apple’s MacBook Pro sales.”

AI on the rise

Microsoft going this route wouldn’t be a surprise. In fact, it was a stated goal when Microsoft first launched Copilot+ PCs, when the company told us: “We will bring new Copilot+ PC experiences at a later date. In the future we expect to see devices with this silicon [meaning Qualcomm Snapdragon ARM-based chips, plus Intel and AMD CPUs] paired with powerful graphics cards like Nvidia GeForce RTX and AMD Radeon, bringing Copilot+ PC experiences to reach even broader audiences like advanced gamers and creators.”

Well, fair enough – Microsoft is doubtless planning more powerhouse ARM-based Copilot+ PCs (and x86 models, meaning Intel and AMD silicon) down the line, and they will surely be a challenge to the MacBook Pro. However, Apple is not going to stand still with the MacBook Pro either, and in a few years’ time, we’ll have even thinner (OLED-toting) models of the Pro too – maybe with their own equally highly performant graphics solutions (we certainly wouldn’t bet against it).

This isn’t just about the hardware, though, and a lot depends on exactly how well Microsoft does with pushing the software side of the equation, and getting Windows on ARM working seamlessly with emulation – given that most apps aren’t natively coded for ARM (but x86 instead). Microsoft’s Prism emulation layer is the key to achieving this – and the firm will need to progress with software compatibility, and smooth running, as we’re not quite there yet with Prism.

It's still very early days for Prism, and Copilot+ PCs in general, though – and attention-grabbing predictions like Omdia’s are kind of par for the course. However, major growth in AI PCs isn’t something that can be ruled out, and indeed it feels like a distinct possibility – but a lot of that may also be wrapped up in what AI features Microsoft can grace Copilot+ PCs with. On that front, things aren’t off to a great start with the Recall debacle, of course, and lately, everything has gone very quiet with that particular feature – perhaps ominously so.

You might also like...

• Microsoft’s embarrassment over Recall fiasco gets worse as Windows 11 feature becomes the butt of Apple exec’s joke
• Windows 11 remains an unloved OS – but why won't people upgrade?
• Don’t make these 5 big mistakes when using Windows 11

Intel has divested its 1.18 million-share stake in British chip firm Arm, according to a regulatory filing, as it looks to recover some momentum.

The sale, which likely netted Intel around $147 million based on the average price of Arm’s stock in the second quarter of 2024, follows a troubling period for the chipmaker which is undergoing significant restructuring and cost-cutting measures.

Intel’s decision to sell its stake in Arm is considered to be part of its strategy to streamline operations, thus it’s unlikely to be a reflection of lost confidence in the British firm.

Intel sells Arm shares

Intel is currently facing intense competition, particularly in the AI chip sector, where it lags significantly behind industry leader Nvidia, which was the world’s most valuable company for a period earlier in 2024.

In response to increased pressure from competition, Intel announced a $10 billion cost-reduction plan including reducing its headcount by 15% – one of the biggest tech layoffs in terms of percentage since the pandemic-induced layoff season that started in 2022.

The news saw stock plummet by 26% on August 2. Intel has a market cap of $87.52 billion, compared with AMD’s $228.41 billion and Nvidia’s $2.856 trillion.

Intel CEO Pat Gelsinger’s strategy is to reposition the company as a leader in advanced AI chip development and for-hire manufacturing, but this has led to rising costs.

In June, Intel reported cash and cash equivalents of $11.3 billion, with total current liabilities of around $32 billion. Intel stock has been down by around 57% since the beginning of the year, following a period of steady decline followed by a sharp drop that occurred after the company’s statement.

Upon the breaking of Intel’s Arm-related news, Arm stock was not negatively impacted, however what the move means for Intel is yet to be seen.

Intel declined to comment.

More from TechRadar Pro

• Check out the best processors from AMD and Intel
• We’ve rounded up a list of the best AI tools and best AI writers
• AMD teams up with Arm to unveil AI chip family that does preprocessing, inference and postprocessing on one silicon