General Compute raised US$100 million in a Series A funding round. Blackbird and Abstract co-led the round. The San Francisco company builds an inference cloud for ultra-fast tokens.
Funding and investors
Fuse Venture Partners, Village Global, and others also joined the round. Additional investors include Bold Capital, QED Fund, and Endeavor Catalyst. General Compute previously secured a US$400 million debt facility from Upper90.
Company context
General Compute focuses on alternative chips for AI inference. It finances, deploys, and operates GPUs and purpose-built inference chips. Its partners include Cerebras, SambaNova, and Positron AI.
What customers get
Customers can choose dedicated bare metal or a guaranteed tokens-per-second SLA. The company matches multi-stage workloads to the architecture that runs them best. It supports nearly any major frontier or open-source model.
First major deployments are planned for Q1 2027. General Compute has more than 50 MW of capacity in the pipeline. Customers can access that capacity without procuring or managing the hardware themselves.