General Compute raised $100 million and is now valued at $800 million, a sign that demand for AI computing is holding up even as some investors worry.
A young company that rents out computing power for AI has raised $100 million, in a week when many people are asking whether the AI boom is about to bust. General Compute is now valued at $800 million. Its CEO says fears of a collapse are "massively overstated."
General Compute announced a $100 million Series A round, led by Blackbird Ventures and Abstract Ventures. A Series A is an early funding round for a young company. In May, after its seed round, the company was valued at $60 million. The new valuation is $800 million.
A deal with a large customer worth $100 million in annual recurring revenue helped pave the way for the round.
General Compute is a "neocloud," a newer kind of company that rents computing power for AI. CoreWeave, a publicly traded neocloud, has seen its stock fall to a 12-month low. Elon Musk pointed to $40 billion in debt for buying computing capacity and predicted that "200+ neoclouds are going to see guppy-level death and consolidation."
CEO Finn Puklowski is unfazed. He thinks Musk is talking his book, meaning talking up his own interests, to push down the prices of the chips Musk wants to buy. Puklowski says people call him every week asking for more computing power. He also says Anthropic's bullish S-1, a filing for a public stock offering, is "settling fears for today."
Most of the market chased GPUs, the chips Nvidia is best known for. General Compute went another way. It builds its cloud from chips designed for inference, which means running an AI model to produce answers. These chips come from Nvidia rivals.
The company does not build giant data centers. It uses smaller power sources, rents space in other companies' data centers, and converts former bitcoin-mining sites to run AI models. That means it will not rise to the top ranks of neoclouds or rival the biggest cloud providers soon. But when it has chips, it will have somewhere to plug them in.
Its customers are cloud companies that offer developers fast inference. General Compute supplies the raw servers, with no software on top.
Puklowski expects to run 75 megawatts of computing capacity in 2027. A megawatt is a measure of electrical power. He suspects he will have access to 200 megawatts, "more than we can probably fund and get hardware for." He says his financing work is not done: "capital is going to be a full-time job for me, both on debt and equity."
SambaNova CEO Rodrigo Liang put the challenge plainly: "Everyone in this market is selling speed. The real question is whether you can make money doing it." The article does not say whether that question has been answered yet.
Investors are still backing AI computing despite bust talk, but whether these companies can make money at it is the open question.
Visa tested 22 AI models on realistic shopping tasks. They passed just over half, and often bought the wrong thing anyway.
Read more: Visa tested 22 AI shoppers, and most still make costly mistakesDARPA wants AI to hunt for design flaws in stock, bond, energy and cloud markets, so enemies cannot use them to cause economic damage that reaches households.
Read more: Pentagon wants AI to find weak spots in stock and energy marketsTuring Award winner Richard Sutton says more computing power alone will not make AI truly smart. Big AI companies disagree, which matters for anyone betting on AI.
Read more: AI pioneer says piling on computing power won't make machines smart
