General Compute has raised $100 million in a Series A round, TechCrunch reported on October 27, valuing the neocloud startup at $800 million after the investment. The round was led by Blackbird Ventures and Abstract Ventures, and it follows a deal with a large customer worth $100 million in annual recurring revenue. The startup builds its cloud from specialised inference chips made by Nvidia rivals, a deliberate break from the GPU-centred model most neoclouds follow.
The Round And The Valuation
According to TechCrunch, General Compute was valued at $60 million after a seed round in May. The new Series A lifts that figure to $800 million post-money, a more than thirteenfold rise in about five months. The company says a single large customer deal, worth $100 million in annual recurring revenue, paved the way for the raise.
James Palmer, a principal at Blackbird who co-led the round, told TechCrunch that his firm’s confidence comes partly from its investment in Baseten, an inference cloud company. Palmer said Blackbird has watched Baseten’s growth over the past ten months, alongside rising demand for open-weight inference and for companies owning their own intelligence. He added that chipmakers such as Cerebras, Positron and SambaNova need distribution partners that can take them to scale and navigate the complexity of the physical world.
The timing is awkward, TechCrunch noted. CoreWeave, the publicly traded neocloud, has seen its stock fall to a 12-month low. Elon Musk is flexing $40 billion in debt to buy computing capacity and has predicted that more than 200 neoclouds will face death and consolidation. General Compute co-founder and chief executive Finn Puklowski dismissed that warning, telling TechCrunch that fears of a chip-led collapse are massively overstated. He said Anthropic’s bullish S-1 filing is settling fears for now.
A Cloud Built On Rival Chips
General Compute’s niche is inference, the work of running trained models rather than training them. It builds its cloud from specialised inference chips made by Nvidia rivals. The company struck a $400 million deal to buy chips from SambaNova, which TechCrunch described as the first debt financing of inference chips, and has since added Positron and Cerebras to its supplier portfolio.
Its strategy is to avoid building huge data centres from scratch. Instead, General Compute uses smaller power sources, rents space inside other companies’ data centres, and converts former bitcoin-mining sites into places that run AI models. Puklowski expects the company to deploy 75 megawatts of computing capacity in 2027, and believes it will have access to 200 megawatts of power, more than it can probably fund and buy hardware for.
The startup’s customers are cloud companies that specialise in offering developers high-speed inference. General Compute supplies the bare metal, the raw servers with no software layered on top, for those companies to bring their offerings to market. Puklowski was frank that the financing work is not finished, telling TechCrunch that capital is going to be a full-time job for him, both on debt and equity.
Why It Matters
The round is a vote for a different shape of AI infrastructure. Most neoclouds chase the same Nvidia hardware and the same giant data-centre builds, which leaves them exposed when chip prices, power availability or investor sentiment turn. General Compute is betting that inference, the everyday work of serving models to users, rewards a cheaper and more flexible footprint than training does. If that bet pays off, smaller chipmakers gain a route to market that does not depend on displacing Nvidia head-on.
The deal also shows venture money still flowing into the compute layer, even as the biggest neoclouds wobble. Capital is following the same logic visible in other recent financings covered here, including Nvidia’s backing of Simplismart’s inference software and Guidewire’s $2.1 billion AI underwriting acquisition: buyers and investors are paying for tools and capacity that make AI cheaper to run, not just bigger to build. The contrast with SpaceX’s $40 billion leased-chip deal is sharp, since General Compute is raising equity to own a niche while SpaceX leases capacity at vastly larger scale.
Rodrigo Liang, SambaNova’s chief executive and co-founder, put the challenge plainly to TechCrunch. Everyone in this market is selling speed, he said, and the real question is whether you can make money doing it. General Compute’s answer, so far, is to keep its sites small, its power flexible and its chips cheap enough to earn their keep.
Background
General Compute was founded to serve the inference market with hardware from chipmakers outside Nvidia’s orbit, and its May seed round valued it at $60 million. The neocloud sector boomed as AI labs scrambled for computing capacity, then cooled as investors began questioning whether that capacity would earn its cost. General Compute’s Series A tests whether a smaller, inference-first operator can raise and deploy capital while its larger rivals retrench.
Source: TechCrunch: General Compute Series A
0 Comments