🔍 Read the full analysis: Does Britain Spend More Than The US On Pre-Release AI Testing? on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Britain’s Budget confirmed £120 million in additional funding for its AI Security Institute over four financial years, on top of the institute’s stated £66 million annual funding. Using the source’s estimates, that puts UK funding at about £96 million a year, compared with roughly $15 million for the US center in fiscal 2026; the comparison changes if proposed US increases become law.
Britain’s October 28 Budget confirmed an extra £120 million over four financial years for the AI Security Institute, widening the funding gap with its US counterpart when the figures are compared annually. The source’s calculation puts the UK institute at about £96 million a year including existing funding, against roughly $15 million a year for the US center in fiscal 2026; the exact comparison depends on exchange rates, funding definitions and whether proposed US increases are approved.
The UK allocation covers 2026-27 to 2029-30. Divided evenly across those years, the new money averages about £30 million annually, though the Budget does not specify how it will be distributed year by year. The AI Security Institute says its existing funding is £66 million per financial year. Adding the figures gives about £96 million annually, a calculation rather than a separately stated Budget total.
The US comparison is the testing center within NIST, referred to in the source as the Center for Advancing Innovation and Standards for Super Intelligence, or CAISSI. The Institute for Progress estimates its fiscal 2026 resources at about $15 million, including appropriated funding and a Technology Modernization Fund loan spread over two years. The White House has requested $27 million for fiscal 2027, while a House committee report provides for up to $15 million. Those figures are not equivalent to an enacted, permanent increase.
The source describes UK and US testing as voluntary. It also cites AISI evaluations involving pre-release models and agent tests, but the reported results do not establish how a model will behave in every real-world setting. Government testing is evidence of an evaluation effort, not a certification that a system is safe.
Infographic 01 · The money, made comparable
Britain spends about 8 to 9 times more per year. Not “10 times” from the top-up alone.
The new £120 million for the UK AI Security Institute covers four financial years. Put both countries on a per-year basis and the gap is still large, just smaller than the headline.
Testing budget per year
Million US dollars · UK at about $1.33 per pound, the rate Semafor used
Why “10 times” needs a footnote
- Semafor: the £120m (~$160m) top-up “alone is more than 10 times” the US center’s annual budget
- But: the £120m is spread over four financial years, the US figure is one year
- Per year: the top-up is about $40m, roughly 3 to 4 times the US budget
- With AISI’s existing money: about 8 to 9 times
- Not counted: free compute and model access from companies, and testing by other agencies such as the NSA or the UK’s NCSC
Institute for Progress estimates: a “limited” US center needs $26m a year, an “equipped” one $84m. Matching the UK’s share of total government spending would mean about $850m a year.
Sources: HM Treasury, Budget 2026, 28 Oct 2026; AISI About page, accessed 29 Oct 2026; Institute for Progress, 13 May 2026; Semafor, 27 Oct 2026; House CJS report FY2027 (up to $15m). Per-year averages, ratios and dollar conversions are my arithmetic.
Powered by Thorsten Meyer AI
What the Funding Gap Changes
More public testing capacity can give governments a stronger basis for identifying risks in advanced systems before or after release. The UK figure, if spent as described, would support an institute that says it has more than 100 technical staff and priority access to substantial public computing resources. Funding alone, however, does not show how many evaluations will be completed or whether developers will act on the findings.
For companies deploying frontier models or AI agents, the difference matters because government testing can surface risks involving cybersecurity, biosecurity and unsanctioned actions. The source advises companies to check which rules apply to them and use public evaluation reports to inform their own safeguards. For users, a government test should not be read as a general approval label: the scope, conditions and date of an evaluation matter.
The comparison is also relevant to policy debates about whether voluntary arrangements can keep pace with rapidly changing systems. A larger budget may allow more testing, but it does not by itself give an institute authority to compel access or block a release. The source says the UK and US approaches rely on voluntary participation, while the EU has a separate legal framework for evaluating large models.
Products Worth Considering
Amazon Kindle Scribe Colorsoft 64GB (1st Generation)
Apple iPad Air (2025) 13‑inch Wi‑Fi 256 GB – Space Gray (M3 chip)
Anker SOLIX S2000 Portable Power Station, 2,010Wh, 1,500W
As an affiliate, we earn on qualifying purchases.
How the Annual Figures Compare
The headline claim that Britain’s new funding is “more than 10 times” the US center’s annual budget compares four years of UK funding with one year of US funding. That is not an annual-to-annual comparison. The source’s arithmetic spreads the UK’s new £120 million evenly over four years, producing about £30 million a year, then adds the institute’s stated £66 million annual funding.
On those assumptions, UK annual funding is about £96 million, or roughly $128 million using the source’s conversion, compared with about $15 million for the US center. That is around eight to nine times the fiscal 2026 estimate. Against the White House’s proposed $27 million for fiscal 2027, the ratio would be under five times, but that request is not settled funding.
The UK allocation is part of a broader £700 million package described by the Treasury. The source says the package also includes £165 million for a National AI Biosecurity Initiative and £40 million for a National Agentic Security Unit. AISI began in 2023 and sits within the Department for Science, Innovation and Technology. The US center also began in 2023 and operates within NIST; its name and proposed funding have changed over time.
What the Budget Figures Leave Open
The UK Budget sets out the four-year total, but the source says it does not specify the annual schedule or precisely what activities the new money will fund. The £96 million annual figure is an estimate that assumes an even spread of the new allocation and adds it to the institute’s stated existing funding. The source does not establish whether all of that funding is directly comparable with the US center’s operating resources.
US funding remains subject to the budget process. The $27 million request and House committee provision are different stages of federal budgeting, and neither should be described as a final increase on the information provided. The source also reports that the US center lacked a permanent director at the time of publication and that NIST had not published its headcount.
Evaluation claims also need careful boundaries. The source reports specific model-test outcomes, including simulations and live internet tests, but does not provide enough detail here to independently assess the test protocols, representativeness or how results compare across models. It says Anthropic released Mythos 5.1 in September without giving any organisation outside the US access; the reasons and full circumstances are not established in the material provided.
Budget Decisions and Test Access
The next immediate question is how the UK allocates the £120 million across 2026-27 to 2029-30 and what testing capacity that spending supports. In the United States, lawmakers still have to resolve the fiscal 2027 funding process. The source also describes a bipartisan proposal for $100 million a year through 2029, but that proposal is not enacted funding.
For AI developers and users, the practical next step is to follow published evaluation reports and check the limits of each assessment: which model version was tested, under what conditions, and whether the developer participated voluntarily. The funding comparison is likely to change as budgets are finalized, while access to systems for testing remains a separate issue from the size of the testing teams’ budgets.
Key Questions
Does Britain spend more than the US on pre-release AI testing?
Using the source’s estimates, yes: annual UK funding is about £96 million, compared with roughly $15 million for the US center in fiscal 2026. The figures depend on the UK’s four-year allocation being spread evenly and on which US funding estimate is used.
Why is the “more than 10 times” comparison disputed?
It compares the UK’s £120 million over four years with one year of US funding. Annualized, the new UK money averages about £30 million; adding the institute’s stated £66 million annual funding gives an estimated £96 million a year.
Has the US approved the proposed $27 million?
The source identifies $27 million as the White House’s fiscal 2027 budget request, not a final appropriation. It also cites a House committee provision of up to $15 million; the material does not establish the final enacted amount.
Does government testing mean an AI system is safe?
No. Testing describes an evaluation under particular conditions; it is not a general safety seal. The source says UK and US evaluations are voluntary and stresses that test findings should be understood in light of their scope and limits.
Source: ThorstenMeyerAI.com
