General Compute Series A funding of $100 million has valued the AI cloud startup at $800 million as it builds capacity on specialised inference chips from suppliers outside Nvidia’s ecosystem. The round was led by Blackbird Ventures and Abstract Ventures. A customer contract worth $100 million in annual recurring revenue helped clear the path for the raise.
Co-founder and chief executive Finn Puklowski said demand for computing power remains strong. He argued that fears of a chip-driven collapse in the sector are “massively overstated,” saying he still fields weekly requests for more TPUs, GPUs and other chips. He also said Anthropic’s upbeat S-1 filing was “settling fears for today.”
Deal size, valuation and backers
After a May seed round that left General Compute valued at $60 million, the new Series A takes the post-money valuation to $800 million. Blackbird principal James Palmer, who co-led the round, said the firm’s confidence partly stems from its investment in inference cloud company Baseten and from watching demand for open-weight inference grow.
Palmer said chipmakers such as Cerebras, Positron and SambaNova need distribution partners that can scale and handle the physical complexity of deploying hardware. Puklowski added that raising capital—both debt and equity—will remain “a full-time job” for him.
Betting on inference chips beyond Nvidia
General Compute Series A funding supports a narrow bet: cloud capacity built around inference chips designed by Nvidia rivals, rather than a race for general-purpose GPUs. The company struck a $400 million deal to buy new chips from SambaNova, described as the first debt financing of inference chips, and has added Positron and Cerebras to its supplier list.
Its customers include cloud firms that sell high-speed inference to developers. General Compute supplies bare metal—raw servers without a software stack—so those firms can bring their own platforms to market. SambaNova chief executive Rodrigo Liang said the industry often sells on speed, but “the real question is whether you can make money doing it.”
Smaller sites, rented power, 2027 capacity target
Instead of building large data centres from scratch, General Compute rents space in other operators’ facilities, uses smaller power sources, and converts former bitcoin-mining sites to run AI models. Puklowski expects the company to deploy 75 megawatts of computing capacity in 2027 and said he may have access to about 200 megawatts of power—more than the firm can likely fund and equip in the near term.
That model is not aimed at matching hyperscalers soon. The company argues that when chips arrive, it already has places to plug them in. The General Compute Series A funding comes as CoreWeave’s share price has fallen to a 12-month low and Elon Musk has spoken of heavy debt for computing capacity and predicted widespread consolidation among neoclouds. Puklowski said he remains constructive on CoreWeave’s outlook and suggested Musk was talking down GPU prices he wants to buy.
Aasan alfaaz mein
General Compute naam ki AI company ne $100 million ki nayi funding le li hai. Ab iski qeemat investors ke hisaab se $800 million ho gayi hai. Yeh company Nvidia ke ilawa doosri firms ke AI chips use kar ke cloud capacity banati hai. Is se AI services chalane wale providers ko zyada options mil sakte hain, lekin ordinary users ke bills abhi seedha change nahi hote.
