Partners
Who we work with.
Who needs compute
Open model labs
Compute for launch evaluations before release, free to the lab.
Third-party evaluators
Compute for independent evaluations of frontier open models.
Safety hubs
Compute for the researchers and programmes they support.
Academic institutions
Compute for interpretability, red-teaming and replication research.
Independent researchers
Compute for independent safety research, at or below cost.
Incident investigators
GPUs to support AI incident investigations.
Who helps supply compute
Funders
Fund the compute, engineers and operations behind AI safety work, with commitments that outlast a grant cycle. We don’t take philanthropic dollars from governments or groups deeply involved with politics.
Chipmakers
Hardware donations, substantial discounts, dependable supply and engineering support.
Data centers and neoclouds
Discounted long-term capacity today, and stalled data center projects that could become safety-only clusters.
New data centers
A share of capacity set aside for AI safety research at below-market rates.
Put compute to work.
Bring a model, an evaluation, or the funding to make it happen.