SF Compute is an AI / GPU Cloud company founded in 2023 and based in San Francisco, United States. It has raised $52M in total funding, most recently a Series A in 2025 at a $300M valuation.
| Date | Stage | Amount | Valuation | Lead investors |
|---|---|---|---|---|
| Nov 1, 2025 | Series A | $40M | $300M | DCVC, Wing Venture Capital |
A market-based platform for buying and selling large-scale GPU capacity on H100, H200, and B300 hardware. Buyers purchase nodes on demand through a CLI or web interface at market-clearing prices, with clusters provisioning in roughly five minutes and no long-term contract required. The key differentiator is liquidity: customers can cancel or resell unused capacity back into the market at any time, letting them sign large or longer commitments they can still exit, while never paying egress fees to move their data.
An enterprise tier offering dedicated bare-metal GPU clusters and managed Slurm configurations for teams running demanding training and large-scale inference. Customers can opt for InfiniBand networking and receive dedicated engineering support, with SF Compute managing the stack from UEFI on up. The offering targets labs and companies that need specialized, high-performance infrastructure and hands-on support, drawing on a team that manages more than $100 million of GPU hardware.
We don't have a live feed for this company's ATS. Their careers page has every open role.
View all careers ↗