UK AI Security Funding: Inside the £600 Million Plan to Defend Against AI Threats

Key Takeaways

• Britain’s government set out around £600 million, about $800 million, of funding on Saturday, October 24, 2026, largely geared towards defending against health and security threats from artificial intelligence, according to Reuters.
• The largest share goes to a new biosecurity initiative, including £100 million for an AI Biosecurity Lab to detect biological threats and £65 million to develop countermeasures against pathogens.
• OpenAI and Anthropic will offer up to £200 million of model credits to the biosecurity initiative, which is in-kind access to models rather than government cash, and Google DeepMind and Basecamp Research will offer other assistance.
• Britain’s AI Security Institute receives an extra £120 million, while £130 million goes to cyber defence projects called the National Cyber Shield and the National Exposure and Real-Time Vulnerability Centre.
• Smaller allocations cover robotics, AI research and procurement, economic impact research, a chip design academy, a public policy prize and £30 million for regional mayors to support local AI work.
• This article covers the new funding package, not an earlier warning from AI minister Kanishka Narayan about AI risks, and readers should treat the headline total as approximate, since it mixes government spending with company credits.

What was announced

Britain’s government set out around £600 million of funding on Saturday, largely geared towards defending against health and security threats from artificial intelligence, Reuters reported from London on October 24, 2026. At the exchange rate cited in that reporting, $1 equals 0.7472 pounds, the package is worth about $800 million.

AI minister Kanishka Narayan set out the ambition for the biosecurity work in plain terms, saying the aim was for the AI Biosecurity Lab to become a major global institution bringing together world-leading scientific expertise to protect people from emerging biological threats. That sentence is the clearest statement of intent in the announcement. The lab is not presented as a small research grant, but as an institution intended to carry weight internationally.

Where the money goes: biosecurity first

The biggest chunk of funding goes to a new biosecurity initiative. Within it, £100 million funds a laboratory using AI to detect biological threats, and £65 million goes to developing countermeasures against pathogens. Together, those two lines alone account for £165 million, which explains why biosecurity dominates the coverage of the package.

The private sector contribution is a distinctive part of this initiative. Tech companies OpenAI and Anthropic will offer up to £200 million of model credits to the biosecurity initiative, while Google DeepMind and London-based AI life sciences startup Basecamp Research will offer other assistance. Model credits are access to a company’s AI models, not cash handed to the government, so readers should not count the £200 million as public spending. Its value depends on how the credits are used, which models they cover, and what the lab can actually do with them. Still, for a new lab, access to frontier models from two leading labs is a practical resource, and the involvement of DeepMind and Basecamp Research adds life sciences expertise alongside raw model access.

The AI Security Institute and cyber defence

Beyond biosecurity, Britain’s government will give its existing AI Security Institute an extra £120 million. The Institute is the government’s body for testing and understanding advanced AI systems, and additional funding suggests a larger testing workload ahead, as models grow more capable and more widely deployed. For readers following AI safety, this is the part of the package that most directly concerns how frontier models themselves are evaluated, rather than how their outputs are used in biology or cyberattacks.

A further £130 million goes towards new projects to improve cyber defences for critical infrastructure and the wider economy, known as the National Cyber Shield and the National Exposure and Real-Time Vulnerability Centre. The names point to two different jobs. A shield suggests active protection for essential services, while an exposure and vulnerability centre suggests finding weaknesses before attackers do, and doing so in something close to real time. Critical infrastructure includes energy, water, transport, health services and finance, all of which now depend on connected systems, so the practical question for organisations in those sectors is whether these projects produce tools, warnings and standards they can actually use, rather than reports alone.

Britain’s National Cyber Security Centre will also receive an unspecified sum to expand in Manchester, according to the reporting. The amount is not stated, so it cannot be added to any total, but the location is notable. Expanding cyber capability outside London fits a broader pattern of spreading technical public sector roles across the UK, and Manchester already has a substantial technology and university presence to draw on.

The smaller funds, and what they are for

The rest of the package is made up of smaller allocations, each aimed at a different gap:

• £60 million goes towards a fund that will invest in British companies using AI to advance robotics, autonomous systems and manufacturing.
• £60 million will be used for AI research and procurement contracts.
• £23 million goes to research on the economic impact of AI.
• £18 million goes to a new chip design academy.
• £15 million will fund a prize for AI solutions to public policy problems, initially focused on healthcare.
• £30 million goes to regional mayors to support local AI work.

Read together, these lines sketch a supply chain view of AI capability. Robotics and manufacturing cover physical applications, research and procurement cover the public sector buying and using AI well, economic research covers understanding the effects on jobs and growth, chip design covers the hardware skills underneath it all, and the prize and mayoral funding cover public services and local delivery. None of these sums is large by national infrastructure standards, but they are targeted at specific bottlenecks rather than spread thinly, and the chip academy in particular addresses a long-term skills gap that cannot be fixed quickly.

A note on the headline total

Readers adding up the announced lines may notice that the individual figures do not reconcile neatly to around £600 million, and that is worth stating plainly rather than smoothing over. The £200 million in model credits from OpenAI and Anthropic is company-provided access, not government expenditure, and the National Cyber Security Centre expansion is an unspecified sum on top of the listed items. Reuters describes the package as around £600 million, and this article follows that wording. The precise budgetary treatment of each line, including what counts as new money and what is reallocated or contributed in kind, is a detail to watch for in the government’s own publication of the package, and readers should not treat any unofficial addition of the lines as an audited total.

What this means for different readers

For cyber security teams, the National Cyber Shield and the vulnerability centre are the lines to watch, along with the National Cyber Security Centre expansion in Manchester. Organisations running critical infrastructure should look for guidance, threat information and testing support emerging from these projects, and for any new standards that affect how they report or patch vulnerabilities.

For startups and robotics companies, the £60 million investment fund and the £60 million in research and procurement contracts are the most relevant parts. Public procurement can be a difficult first customer for a small company, so the design of those contracts, including how quickly they are awarded and how much risk a small supplier is asked to carry, will matter more than the headline sum.

For everyone else, the broader signal is that the UK government is treating AI security as a public safety issue on the same footing as health security and cyber defence, rather than as a technology policy debate alone. That framing connects this funding package to the wider AI safety conversation, including the risks posed by autonomous agents operating on the live internet. For a recent example of how agent safety can fail in practice, see our report on Anthropic’s Claude fake police tip, and for how AI is already being deployed in UK public safety work, see our explainer on AI cameras and UK drivers.

What to watch next

The announcement leaves several questions open, and they are the right ones to track. First, timelines: the reporting reviewed for this article does not set out opening dates for the lab, delivery dates for the cyber projects, or application windows for the funds and the prize, so none are stated here. Second, governance: how the AI Biosecurity Lab balances open scientific collaboration with controls on sensitive biological information will shape its credibility. Third, measurement: the government will eventually need to show what the lab detected, what the Institute tested, and what the cyber projects protected, in terms that can be checked rather than asserted.

Bottom line

The UK AI security funding package puts biosecurity first, strengthens the AI Security Institute, invests in cyber defence for critical infrastructure, and adds targeted support for robotics, research, chips, public policy prizes and local AI work. Its headline figure of around £600 million mixes government spending with up to £200 million in company model credits, so it should be read as an approximate package total rather than a precise budget. What matters now is delivery: a lab that can detect threats, an institute that can test frontier models credibly, and cyber projects that make essential services harder to attack.

Sources: Reuters reporting on the announcement, as carried by Reuters and CNA.