SF Compute Secures $40 Million Series A to Expand AI Compute Marketplace
December 2, 2025
As the global demand for artificial intelligence compute power continues to outstrip supply, a new wave of companies is emerging to create more efficient markets for this critical resource. San Francisco-based startup SF Compute has positioned itself at the forefront of this trend, securing significant new funding to scale its operations.
The company announced it has raised $40 million in a Series A equity financing round, which valued the business at $300 million. The funding was co-led by venture capital firms DCVC and Wing Venture Capital, with participation from existing investors Electric Capital and Alt Capital. The capital injection is earmarked for expanding the company's core offering: a marketplace that connects buyers and sellers of GPU computing capacity.
Founded in 2023, SF Compute operates a platform that allows enterprises and data center operators to resell their unused GPU capacity. This model aims to address two key industry challenges: mitigating the financial risk of overbuilding infrastructure and providing flexible, short-term access to high-performance chips for companies that wish to avoid long-term commitments. The startup reports that it currently manages more than $100 million worth of hardware on its platform, despite not owning any GPUs itself.
Ali Tamaseb, General Partner at lead investor DCVC, highlighted the platform's potential impact. "SF Compute is building a true marketplace for compute – allowing companies to buy GPU/compute capacity on flexible, short-term contracts while enabling data centers and enterprises to sell excess capacity," Tamaseb stated. "This unlocks liquidity, efficiency, and price discovery in a market that has historically been rigid and hard to access."
The company, which now employs around 30 people, has been bolstering its executive team with hires from the tech sector, including appointing Eric Park, former CEO of AI cloud provider Voltage Park, as its Chief Technology Officer. Its platform currently lists pricing for leading Nvidia H100 and H200 GPUs and has indicated plans to soon offer access to the newer B300 chips.
This funding round underscores a growing investor appetite for solutions that optimize the allocation of expensive and scarce AI infrastructure. By creating a secondary market for compute, SF Compute's model could potentially increase overall utilization rates across the industry and provide a more agile procurement path for developers and enterprises racing to build and deploy AI models.
Source: datacenterdynamics
SF Compute Secures $40 Million Series A to Expand AI Compute Marketplace
