General Compute raises $100 million Series A at $800 million valuation, betting on non-Nvidia inference chips
A $100 million annual-revenue customer deal helped drive General Compute’s October 2026 raise, as the startup expands beyond Nvidia with chips from SambaNova, Positron and Cerebras. It plans to deploy 75 megawatts of capacity in 2027.
General Compute, a neocloud startup focused on AI inference chips built outside the Nvidia ecosystem, announced a $100 million Series A funding round in October 2026, valuing the company at $800 million post-money. The round was led by Blackbird Ventures and Abstract Ventures.
The fundraise comes at a difficult moment for neoclouds broadly. CoreWeave, a publicly traded competitor, has seen its stock fall to a 12-month low, and Elon Musk has predicted that “200+ neoclouds are going to see guppy-level death and consolidation” while leveraging $40 billion in debt for computing capacity. General Compute co-founder and CEO Finn Puklowski dismissed those concerns, saying fears of a chip-led collapse are “massively overstated” and that Musk is “talking his book to drive down the prices of the GPUs he wants to buy.”
A key driver of the raise was a deal with a large customer worth $100 million in annual recurring revenue. The company was valued at $60 million after a seed round in May 2026. General Compute had previously struck a $400 million deal to purchase inference chips from SambaNova — described as the first debt financing of inference chips — and has since added Positron and Cerebras as chip suppliers.
Rather than building large data centers from scratch, General Compute rents space in existing facilities and converts former bitcoin-mining sites to run AI models. Puklowski said the company expects to deploy 75 megawatts of computing capacity in 2027, with access to potentially 200 megawatts of power available. He acknowledged that capital raising will remain a priority: “Capital is going to be a full-time job for me, both on debt and equity.”
The startup’s primary customers are cloud companies that offer developers high-speed inference services; General Compute supplies the bare-metal servers those companies build on. James Palmer, a principal at Blackbird who co-led the round, cited the firm’s investment in inference cloud company Baseten as context for its confidence, noting that Cerebras, Positron, and SambaNova “all need distribution partners who can take them to scale.” SambaNova CEO Rodrigo Liang framed the broader challenge simply: “Everyone in this market is selling speed. The real question is whether you can make money doing it.”