Center for AI Safety (CAIS) — infrastructure: Compute cluster with 256 NVIDIA A100 GPUs available for AI safety researchers

NoteThe source directly confirms the claim's core assertion: CAIS operates a compute cluster with 256 A100 GPUs. The source text explicitly states 'We have set up a compute cluster of 256 A-100 GPUs with 80GB memory and 1,600 Gbit/s inter-node network speeds.' The source is from a Wayback Machine capture dated March 15, 2023, which is prior to the claim's 'as of 2024' timeframe. However, the claim does not specify that the infrastructure changed between 2023 and 2024, and the source confirms the infrastructure existed and was available for AI safety researchers through an open application process. The temporal mismatch (source from 2023, claim from 2024) does not contradict the claim — it simply means the source predates the claim's reference year, but the infrastructure described in 2023 would reasonably still exist in 2024 unless explicitly stated otherwise.