CoreWeave, a publicly traded neocloud that rents out computing power for AI, has seen its stock fall to a 12-month low, and Elon Musk is predicting that “200+ neoclouds are going to see guppy-level death and consolidation” while flexing $40 billion in debt to buy computing capacity. It is not an ideal week to announce a neocloud Series A, but General Compute co-founder and CEO Finn Puklowski is unfazed. He is bullish on CoreWeave’s chances and thinks Musk is talking his book to drive down the prices of the GPUs he wants to buy. Fears of a chip-led collapse are, he says, “massively overstated”: he has people calling every week asking for TPUs, GPUs, or any chip to get more compute. Puklowski also says Anthropic’s bullish S-1 is “settling fears for today.”

General Compute’s niche is building its cloud out of specialized inference chips made by Nvidia rivals. The company struck a $400 million deal to purchase new chips from SambaNova—the first debt financing of inference chips, per TechCrunch—and has added Positron and Cerebras to its portfolio of chip suppliers. A deal with a large customer worth $100 million in annual recurring revenue paved the way for the company to raise $100 million in Series A funding led by Blackbird Ventures and Abstract Ventures. After a May seed round that valued it at $60 million, General Compute is now worth $800 million post-money to its investors.

The startup’s strategy is to zig where others zag. It focused on chips built for inference while everyone else chased GPUs. Instead of taking on the complex effort of building huge data centers, it uses smaller power sources, rents space in other companies’ data centers, and converts former bitcoin-mining sites into places that run AI models. That approach means General Compute isn’t going to catapult into the top ranks of neoclouds or compete with hyperscalers any time soon, but it also isn’t subject to the constraints that come with building data centers from scratch. When General Compute has chips, it will have someplace to plug them in. Puklowski expects the company to deploy 75 megawatts of computing capacity in 2027, and suspects he will have access to 200 megawatts of power—“more than we can probably fund and get hardware for.”

The plan has an explicit limit. Puklowski’s financing work is not done; as he told TechCrunch, “capital is going to be a full-time job for me, both on debt and equity.” The startup’s key customers are cloud companies that specialize in offering developers high-speed inference; General Compute provides the bare metal, or raw servers with no software layered on top, for those companies to bring their offerings to market. James Palmer, a principal at Blackbird who co-led the round, said his firm’s confidence in General Compute comes partly from an investment in Baseten, an inference cloud company. “We’ve been proximate to their journey over the last 10 months and observing the wild growth story that [Baseten] has gone on, and the demand for open weight inference, and for companies owning their own intelligence,” Palmer said. When it comes to General Compute, he said, Cerebras, Positron, and SambaNova “all need distribution partners who can take them to scale, who can navigate the complexity of the physical world.” Executives at all three chipmakers praised General Compute to TechCrunch, each noting the importance of making the economics of AI cloud compute work. “Everyone in this market is selling speed,” SambaNova CEO and co-founder Rodrigo Liang said. “The real question is whether you can make money doing it.”