🔍 Read the full analysis: Is Britain’s Pre-Release AI Testing Spend Ten Times Higher? on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
The UK’s 28 October Budget confirmed £120 million in additional funding for the AI Security Institute over four financial years. Adding the annualized top-up to the institute’s stated existing funding gives about £96 million a year, compared with an estimated $15 million for the US testing center; the comparison is not like-for-like, and the US has requested more for 2027.
The UK’s 28 October Budget confirmed an additional £120 million for the AI Security Institute (AISI) over four financial years, prompting a claim that Britain’s pre-release AI testing spend is more than ten times its US counterpart. On an annualized basis, however, the source material estimates UK funding at about £96 million a year, against roughly $15 million a year for the US testing center — a comparison that depends on exchange rates, what each figure includes and whether proposed US funding is counted.
The Budget sets the new AISI funding across 2026-27 to 2029-30. That averages about £30 million a year, although the Budget does not specify how the money will be divided among those years. AISI’s website lists existing funding of £66 million per financial year. Adding the new funding’s average annual value produces an estimated £96 million a year; it is an arithmetic comparison, not a published annual budget figure.
The US counterpart is a testing center at NIST whose current webpage uses the name Center for Advancing Innovation and Standards for Super Intelligence (CAISSI). The source material cites an estimate of about $15 million for fiscal 2026, while the White House has requested $27 million for fiscal 2027. A House committee report provides for up to $15 million. The request and committee figure are not the same as enacted funding, and a separate proposal for $100 million annually through 2029 is not law.
The widely repeated “more than 10 times” comparison refers to the UK’s £120 million multi-year top-up against a single year of US funding. The source’s annualized calculation puts the UK figure at roughly $128 million, or about eight to nine times the $15 million US estimate. Against the US administration’s $27 million request, the ratio would be under five times. These are approximate comparisons: the cited figures differ in period and status, and the source does not provide a full, independently reconciled account of each institution’s spending.
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 Can Mean
These institutes test advanced AI systems for risks such as cybersecurity, biosecurity and misuse, including through evaluations before release. Their resources can affect how many systems they examine, how deeply they test them and how quickly they can publish findings. A larger stated budget does not, by itself, prove that one country’s testing is more effective or that its assessments cover the same activities as another’s.
For developers and organizations deploying AI agents, the practical issue is that testing and oversight arrangements differ across jurisdictions. The source says both UK and US evaluations are voluntary, so a government assessment is not a general certification or guarantee of safety. Public evaluation reports can inform organizations’ own risk controls, but they do not replace testing against each system’s specific use and operating conditions.
The spending also sits within a wider UK security package. The Treasury puts that package at £700 million, including £165 million for a National AI Biosecurity Initiative and £40 million for a proposed National Agentic Security Unit. Those allocations suggest a broader focus on AI-related security, but the source material does not establish how the total package will be spent in practice or what results it will produce.
Products Worth Considering
LG UltraGear 27-Inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDMI 2.1
Amazon Kindle Paperwhite Signature Edition (12th Generation, 2024), 32 GB, Metallic Black
Logitech MX Keys S Wireless Keyboard, Graphite
As an affiliate, we earn on qualifying purchases.
How UK and US Testing Differ
The UK established AISI in 2023, initially under the name AI Safety Institute, as part of the Department for Science, Innovation and Technology. Its website lists more than 100 technical staff and priority access to over £1.5 billion in public compute. The 2025 Spending Review had already allocated £240 million to the institute; the source does not set out whether that earlier sum and the stated £66 million per-year funding figure cover identical periods or budget categories.
The US testing body also began in 2023. The source traces a name change to the Center for AI Standards and Innovation in 2025, while NIST’s current page uses CAISSI. NIST describes it as a government contact point for industry and says evaluations focus on demonstrable risks through voluntary agreements with developers. The source reports that the center has no permanent director following Chris Fall’s July resignation; it also says NIST does not publish a headcount for the center.
The source cites a separate European model: the EU AI Office, which it says has more than 125 staff and gained legal authority on 2 August 2026 to evaluate large models and fine providers up to 3% of global turnover. That makes the comparison more than a simple UK-US budget contest: staffing, legal powers, access to systems and the scope of evaluation also shape what governments can do.
What the Budget Figures Leave Open
The Budget does not disclose the year-by-year schedule for the £120 million, so £30 million annually is an average rather than a confirmed allocation. The source also does not reconcile whether all UK figures include the same cost categories, or whether compute access and staff resources should be included in a direct financial comparison.
The US figures remain unsettled: the $27 million request is a proposal, the House committee’s “up to $15 million” provision is not described here as final enacted funding, and the $100 million annual proposal is also not law. The source gives no final US fiscal 2027 appropriation. Exchange rates and the precise basis of the estimated $15 million figure also limit the precision of the ratio.
The source describes both countries’ testing as voluntary, but does not detail the terms of every agreement or provide evidence that all leading models are submitted for evaluation. It reports that Anthropic released Mythos 5.1 in September without giving access to any organization outside the US; the scope and circumstances of that reported access arrangement are not independently established in the material supplied.
Funding Decisions and Test Access
The next clear budget milestones are the UK’s annual spending allocations and the US Congress’s decisions on fiscal 2027 funding. Those will show how the UK distributes its four-year top-up and whether the US request or committee provision becomes law. Any later comparison should use confirmed annual figures and identify whether staffing, computing resources or other related programs are included.
For AI developers and deployers, the immediate steps are to check which national or regional rules apply, review public evaluation reports and conduct their own testing for the systems they use. The source’s central caution is that government testing is not a safety seal: it is an attempt to examine risks, not a guarantee that a model will behave safely in every setting. Further details on test access, methods and published findings will be needed to judge how the additional UK funding changes evaluation coverage.
Key Questions
Is Britain spending more than ten times as much as the US on AI testing?
Not on the annualized comparison presented in the source. It estimates about £96 million a year for the UK after averaging the top-up, versus roughly $15 million for the US in fiscal 2026. That is estimated at eight to nine times, not more than ten. The ratio is approximate and depends on which US figure is used.
How long does the UK’s £120 million top-up cover?
The Budget says the additional funding runs across 2026-27 to 2029-30, or four financial years. Dividing it evenly gives an average of £30 million a year, but the Budget does not publish a year-by-year breakdown.
Is the US center receiving $27 million?
The source describes $27 million as the White House’s fiscal 2027 budget request, not as confirmed funding. It also cites a House committee provision of up to $15 million; the final enacted amount is not established in the material provided.
Does a government AI test mean a model is safe?
No. Testing can identify risks under particular evaluation conditions, but it is not a guarantee of safe behavior in every use. The source says UK and US evaluations are voluntary and cautions that government testing should not be treated as a general approval seal.
Source: ThorstenMeyerAI.com
