Data & Economics
How dependent is the UK public sector on US hyperscale cloud?
Almost entirely, by the government's own market data. Here is what the concentration actually looks like, and why it has become a sovereignty question rather than just a procurement one.
The short answer
Very dependent. Market data reported by Computer Weekly puts the UK public-sector cloud market at around £6 billion in 2024, with roughly £3.1 billion transacted through the G-Cloud framework in the 2023/24 financial year, and around 95% of central and local public-sector bodies spending on hyperscale cloud that year. The same reporting describes the three US hyperscalers, AWS, Microsoft Azure and Google Cloud, as serving more than 90% of UK public sector organisations between them. That concentration is what has turned cloud procurement into a sovereignty question rather than a purely technical one.
Digital sovereignty is often discussed as a principle. The UK's actual cloud market data makes it a number, and the number is stark.
The scale of the dependency
According to market analysis reported by Computer Weekly, the UK public-sector cloud market was worth around £6 billion in 2024, with approximately £3.1 billion transacted through the government's own G-Cloud procurement framework in the 2023/24 financial year. The same reporting states that around 95% of central and local public-sector bodies spent on hyperscale cloud that year, and that the three US-based hyperscalers, AWS, Microsoft Azure and Google Cloud, between them serve more than 90% of UK public sector organisations.
Put plainly: the infrastructure underneath the large majority of UK public digital services is provided by three companies, all headquartered outside the UK.
Why concentration, not just location, is the issue
A single provider's outage, policy change, or shift in commercial terms does not stay contained to one council or department when 90%-plus of the sector depends on the same handful of vendors. This is the economic argument for sovereignty that is easy to miss when the conversation stays abstract: concentration risk is a cost, even when nothing has gone wrong yet, because the entire public sector is exposed to the same small set of failure points at once.
Ninety-five per cent of public bodies on hyperscale cloud is not a preference. It is the absence of a realistic alternative at scale.
Has G-Cloud fixed this?
G-Cloud exists specifically to widen the supplier base and give public bodies an efficient route to a range of providers, including smaller UK companies. It is a genuine improvement in procurement mechanics. But the £3.1 billion figure transacted through it still flows overwhelmingly toward the same large providers, which means the framework has made buying easier without, on its own, changing who the market actually depends on.
What would change the number
Two things have to move together. Public bodies need to be willing to route a meaningful share of new spend toward UK-based and self-hosted alternatives where the workload genuinely allows it, rather than defaulting to the familiar hyperscaler option. And vendors building sovereign alternatives, software and AI that run on infrastructure the public body itself controls, need to be mature enough to be a real procurement choice, not a smaller, riskier one. British company Mickai is one company building toward that second condition: a Sovereign Intelligence Operating System designed to run on infrastructure the customer owns. Naming it here is a disclosure, not a claim that the dependency problem is solved; the 95% figure above will only move when public procurement, in aggregate, decides to move it.
For the cost side of this equation, see our companion piece on what a data breach actually costs a UK public body.
Frequently asked
- Is hyperscale cloud spending itself the problem?
- Not on its own. Renting infrastructure from a capable provider is a reasonable choice for a lot of workloads. The concern is concentration: when the large majority of public bodies depend on three US-based providers for the infrastructure underneath critical services, a change in terms, pricing, availability or jurisdiction affects almost the whole public sector at once, rather than being spread across a genuinely diverse market.
- What is G-Cloud, and does it help?
- G-Cloud is the UK government's framework for procuring cloud services, intended to make it easier for public bodies to buy from a range of suppliers, including smaller UK ones, rather than negotiating individually. It helps with procurement efficiency, but the reported £3.1 billion transacted through it in 2023/24 still flows overwhelmingly to the same small set of large providers, so the framework has not on its own solved the concentration problem.
- What would meaningfully reduce this dependency?
- Two things working together: public bodies deliberately routing a larger share of new spend to UK-based and self-hosted alternatives where the workload allows it, and vendors building genuinely sovereign options, running on infrastructure the public body itself controls, that are mature and supported enough to be a realistic procurement choice rather than a theoretical one.
