UK puts £100M into an AI biosecurity lab, and OpenAI and Anthropic add up to £200M

The UK government will set up an AI Biosecurity Lab (AIBL), an independent non-profit, with £100 million of public money and up to £200 million in commitments from OpenAI and Anthropic, according to a report relayed by Techmeme from the BBC and the government’s own announcement.

Public money plus up to £200M from two labs

The Cabinet Office’s announcement says the lab will use frontier AI to help the UK understand, prevent, detect and respond to biological threats, including AI-enabled ones. Early work will focus on spotting emerging threats sooner and developing vaccines, treatments and other countermeasures faster. The government calls it the world’s first dedicated AI biosecurity organisation.

The headline figure is up to £300 million: £100 million from government and up to £200 million from the two launch funders. The company money is not cash. OpenAI is offering £100 million in compute and model credits, plus trusted access to its most capable relevant models through its Rosalind Biodefense programme. Anthropic is offering up to £100 million in model credits, trusted model access and dedicated technical support. The government adds that public and philanthropic funding is meant to support the lab over time.

Google DeepMind joins as a research collaborator and says it intends to offer trusted access to Gemini and specialized science models. Basecamp Research will provide access to its AI-biology models and expertise.

Structure, partners and what is still blank

The lab is designed to run with the autonomy and pace of a frontier AI-biology company, and is modeled partly on the AI Security Institute. It launches under a National AI Biosecurity Initiative, alongside the UK Health Security Agency, the Animal and Plant Health Agency and the Defence Science and Technology Laboratory. The medicines regulator MHRA is named for health emergency preparation and for regulating AI-designed medicines. A scientific advisory committee drawing on academia, industry and the third sector will be detailed later.

Much is not yet established. The announcement names no director or staff, gives no location, and sets no launch or delivery dates. It does not say how the credits will be valued, how long they last, or what the “trusted access” tiers allow. Nothing in it independently verifies what the models can do for biodefense; those are the companies’ and government’s own framing. Bloomberg and the Financial Times also reported the plan on October 23, ahead of the official page.

Credits as currency in the AI labs’ public-policy push

The deal shows how frontier labs now pay for public-interest work: in model access rather than checks. For OpenAI, whose valuation talks we covered in its nearing $30B bridge round at a $1.25T valuation, compute credits are cheap to give relative to their list value and tie its models into a government biodefense program. Anthropic’s matching offer fits a pattern: this week it also handed its Agent Skills standard to a neutral foundation, as we reported in Anthropic hands Agent Skills to a neutral foundation.

The funding also lands amid political attention on the labs. Filings reported by The Verge, summarized by Techmeme, show Nvidia, Anthropic and OpenAI each had their highest-spending lobbying quarter between July and September, at $2.2M, $2.1M and $1.5M respectively. Those are US filings and unrelated to the UK plan, but they show how much the same companies now invest in dealing with governments.

The same companies build models that officials worry could be misused in biology, so the lab asks them to help defend against a risk their own products raise. Whether an independent, non-profit body can keep that arrangement credible depends on governance details the government has yet to publish, starting with the advisory committee.

Who can use it, and when

Nobody outside the named partners can use the lab yet. The announcement describes work with UK biosecurity centres, the life sciences sector, academic institutions and public bodies in the UK and abroad, but publishes no application route. Readers in biotech and public health should watch for the advisory committee’s makeup, the first research priorities and any rules on sharing results.