Organizations must own the right to their cloud compute, or innovation will suffer

Data is the beating heart of any modern enterprise, forming the foundations of digital sovereignty, scalability and customer trust.

Businesses around the world know this – what they don’t know is that this data could be taken out of their control at any moment.

When a company signs the terms and conditions of their cloud provider, they often unknowingly sign away true autonomy over their data. Businesses might be able to access it, but they can’t control the ecosystem it lives in or manage how it evolves.

The landmark ruling by the UK’s Competition and Markets Authority (CMA) deemed that the public cloud market is no longer competitive. In Europe alone, 70% of the market is dominated by three players, all of which impose expensive obstacles that prevent customers from shopping around.

The Synergy Research Group found that less than 1% of businesses switch cloud providers each year. This isn’t because they are thrilled with the status quo – the journey to move from one cloud provider to another is prohibitively expensive and technically painful.

The empty promises of multi-cloud

Immobility has led organizations to adopt multi-cloud strategies in the hope that more suppliers would enable the flexibility, cost reduction and portability they need.

There was a marketed promise that businesses could migrate all workloads from the data center as they are, and later on, rearchitect them so cloud efficiencies can be achieved. In reality, this never happened.

Enterprises today are burdened with the task of shifting on-premise workloads into the cloud, whilst simultaneously navigating the growth of cloud-first applications designed by a newer generation of developers. This creates fragmented data, with sprawling data clusters and vendor-locked information.

The underlying problem is more than just market competitiveness, it’s about customer choice, or lack thereof. It’s the erosion of cloud sovereignty, and the stifling of innovation. Startups are penalized for exploration.

University research is beholden to the commercial limitations of single providers. Developers are locked out of the data and compute needed to build responsible AI.

The industry has begun to resemble a past that we collectively tried to move away from, where monopolies dictate entire ecosystems with opaque logic and needless fees. The CMA’s conclusions may have put this reality into sharp focus, but it’s a truth many have felt for years.

This lack of choice matters far beyond the isolated concerns of infrastructure and IT teams.

It’s becoming a matter of national security. Regulators and organizations across Europe are starting to realize the data sovereignty risks of relying on global hyperscalers for AI model training and deployment.

If the geopolitical winds change direction, a government’s access to its own citizen data could become throttled because of the legal obligations of another country.

It also matters for ethics, because you can’t audit what you can’t see. A major fallacy in AI governance is the idea that ethics can be layered on top of closed systems.

The reality is extremely difficult for organizations to regulate model bias, algorithmic fairness and explainability, if they are deploying the most sensitive models through third-part APIs running in black box environments.

Beyond dependency to diversity

While there is no silver bullet, true AI governance starts with infrastructure governance. This means visibility into where every model is hosted. It means having audits and logs that ensure data rules and foreign regulations aren’t being breached.

The next evolution is a platform that doesn’t just promise portability, it guarantees it. Where workloads can be shifted seamlessly, whether it’s from a cloud region to an on-prem rack, or from one provider to another. Where compute engines travel with the data, and where secure, private and transparent AI can run from anywhere.

This isn’t about abandoning cloud. The scalability and elasticity of the cloud still remains revolutionary, but it must be balanced with autonomy.

Organizations can either continue down the path of deeper dependency, with rising costs and diminished control, or embrace a model where infrastructure is fluid, AI sovereignty is preserved, and where cloud is owned, not rented.

From lock-in to liberation

While the CMA ruling has shown that the current structure of the cloud market is unsuitable for the AI future, it’s clear that regulation alone won’t fix it.

We need architecture. We need code. We need systems that empower innovation and don’t punish exploration.

This architecture exists, and for the first time in years, there is a real momentum to make it a reality – unbundling an empire and bringing cloud simplicity to every CPU and GPU on earth.

Ultimately, the right to compute should remain with those who own the data.

We've featured the best database software.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro