UK Backs AI Biosecurity Lab With £100 Million as OpenAI and Anthropic Add Credits

£100 million in public money, and up to twice that again in donated AI computing time. That is the shape of the UK's new AI biosecurity lab, which the British government announced on October 23 as part of a wider AI security package set out ahead of this week's Budget.

The lab is meant to use advanced AI to spot biological threats early and to speed up work on vaccines and treatments. According to TNW, the government calls it the world's first dedicated AI biosecurity organization. OpenAI, Anthropic and Google DeepMind have all signed on in some form, though the companies are giving computing access rather than cash.

The figures behind Britain's biosecurity bet

• £100 million ($134 million): government funding for the AI Biosecurity Lab, according to Bloomberg.
• Up to £200 million: AI model credits pledged by OpenAI and Anthropic combined, per a government statement cited by Bloomberg.
• £100 million each: TNW reports OpenAI is offering £100 million in compute and model credits, and Anthropic up to £100 million in credits, model access and technical support.
• £65 million: a separate pot for developing countermeasures against pathogens.
• £120 million: extra money for the AI Security Institute.
• £130 million: for using AI to find cyber vulnerabilities in critical infrastructure and to shore up national cyber defenses.
• £23 million: to set up a new AI Economics Institute.
• More than £600 million: the size of the full AI security package, according to TNW.

Where the £200 million in credits actually goes

The company numbers are not checks. OpenAI's contribution is compute and model credits, along with access to its models through its Rosalind Biodefense program, TNW reported. Anthropic is offering credits, model access and technical help. Google DeepMind will join research projects and may open up its Gemini and science models, while the British firm Basecamp Research will share its AI biology models.

Credits matter because frontier AI is expensive to run at scale. Still, they only have value if the lab has staff and projects ready to use them. The pattern is becoming familiar: TNW notes that Nvidia pledged $250 million in compute for UK AI safety research earlier in October, also as computing power rather than money.

There is a second reason the AI companies want a seat. Matt Clifford, who leads international affairs at Anthropic and advised the prime minister on AI until last June, told Bloomberg that "AI is changing both the biological risks we face and the nature of biological discovery." The same models that could help design a vaccine could, in the wrong hands, help someone misuse biology. The big labs have been tightening controls on their most capable systems, as shown by OpenAI's move to cut frontier models off from the live internet.

Two jobs for a lab with no address yet

The lab will be an independent nonprofit modeled on the AI Security Institute, TNW reported. It starts with two tasks:

1. Detecting new biological threats early.
2. Speeding up the development of vaccines and treatments.

Public partners include the UK Health Security Agency, the Defence Science and Technology Laboratory and the medicines regulator MHRA, which will work with the lab on how to regulate AI designed medicines.

What the announcement leaves out is just as telling. No director has been named. No location or start date has been given, and the government did not say over what period the £100 million will be spent. AI minister Kanishka Narayan called the lab "a major new global institution" and said that "as AI becomes more powerful, Britain must have the capabilities to protect its people and shape its own future," according to Bloomberg.

What £600 million says about the UK's AI strategy

Bloomberg frames the package as part of a plan to build AI institutions that sit outside regular government departments. The UK trails the US and China on frontier models, so it is trying to matter in a different way: as a place that tests, monitors and studies the technology, and as an "honest broker" on regulation, in contrast to the stricter approach taken by the EU.

The £130 million cyber allocation fits that plan. Bloomberg links it to the recent rush by governments and companies to patch two critical flaws in widely used technology that AI helped identify, a reminder of how quickly AI found bugs, like the critical PHP flaws patched this month, can turn into an emergency.

For US readers, the contrast is with the voluntary approach Washington has leaned on, such as the White House accord on super intelligence safety, or Anthropic's own draft safety bill for Congress. Britain is putting public money into permanent bodies instead. Whether the AI biosecurity lab turns that money and those credits into working tools will depend on details the government has not released yet: who runs it, where it sits and when it opens its doors. More AI policy coverage is on our AI page.