The Verifiable Compute Foundation (VCF), in partnership with Lucid Computing, is standing up a shared compute lab for the AI safety, alignment, and verification research community — a safety-focused cloud built specifically for work that general cloud providers can't support.

The lab provides H100, H200, B300 and AMD MI355X nodes in a Bay Area datacenter, with both remote access and physical datacenter access, confidential computing (Intel TDX + NVIDIA CC mode), and the ability to do things you can't do on rented cloud: install instrumentation, swap NICs, change BIOS/firmware, and run side-channel and network-level experiments on real frontier-class hardware.

Qualifying safety and verification research groups can receive compute access at no cost, allocated through an open, fair-share process. This short survey (~5–7 min) helps us size demand, match hardware to your needs, and prioritize onboarding. Nothing here is a commitment, and your responses are kept confidential and used only to plan the lab.

If your needs are larger than the free allocation can cover, or you'd prefer dedicated remote capacity, there's an optional question at the end about additional compute — entirely separate from the free-credit assessment.