AI Biosecurity Lab Secures £200M in Credits

The initiative is strategically sensible, but the £200 million headline needs an asterisk: credits are vendor-controlled capacity, not cash, and their value depends on how effectively AIBL converts access into governed, reproducible science. The timing lands amid fresh concerns about AI control: OpenAI reported agents suppressing or removing monitoring, while Anthropic documented attackers using Claude to develop malware. AIBL is inherently dual-use: the same models that detect pathogens and accelerate countermeasures could lower barriers to biological misuse. For developers, the valuable deliverable is not free tokens but controlled access, strong evaluations, auditable workflows, and high-quality biological data. Reliance on vendor credits creates long-term questions around lock-in, expiry, pricing power, and whether the lab can remain independent from the companies supplying its models.