General Compute has raised 100 million dollars in Series A funding co-led by Blackbird Ventures and Abstract Ventures, at an 800 million dollar post-money valuation, TechCrunch reported on October 27. The company's own announcement the same day confirms the round, with participation from Fuse Venture Partners, Village Global and others. It is a steep step up from a 60 million dollar valuation after its May seed round, and TechCrunch says a single deal with a large customer worth 100 million dollars in annual recurring revenue paved the way for it.
Where most neoclouds chased Nvidia GPUs for training, General Compute built its niche on specialized inference chips from Nvidia's rivals. It struck a 400 million dollar deal to purchase chips from SambaNova, which TechCrunch describes as the first debt financing of inference chips, and the company says that buildout is backed by a 400 million dollar debt facility from Upper90. It has since added Positron and Cerebras as suppliers. The company sells bare metal - raw servers with no software layer on top - to cloud companies that offer developers high-speed inference, and its announcement promises customers dedicated hardware or a guaranteed tokens-per-second service level without procuring or managing the infrastructure themselves.
The physical strategy is deliberately unglamorous. Instead of building huge data centers from scratch, General Compute rents space in other companies' data centers, taps smaller power sources, and converts former bitcoin-mining sites into places that run AI models. Co-founder and chief executive Finn Puklowski expects the company to deploy 75 megawatts of computing capacity in 2027, with access to about 200 megawatts of power, and the company says its first major deployments go live in the first quarter of 2027, with more than 50 megawatts in the pipeline.
The timing is the test of the thesis. CoreWeave, the publicly traded neocloud, has fallen to a 12-month stock low, and Elon Musk has predicted that more than 200 neoclouds face consolidation. Puklowski called fears of a chip-led collapse massively overstated, said demand for more compute arrives weekly, and pointed to Anthropic's bullish S-1 as settling fears for now. SambaNova chief executive Rodrigo Liang framed the real question plainly: everyone in this market is selling speed, and the question is whether you can make money doing it.
For people building with AI, the significance is distribution. Alternative chips have rarely lacked benchmarks; they have lacked an easy way to rent them at scale. Blackbird principal James Palmer, whose firm also backed inference platform Baseten, told TechCrunch that Cerebras, Positron and SambaNova all need distribution partners who can take them to scale and navigate the physical world. If General Compute makes non-Nvidia inference genuinely easy to buy, open-weight models get cheaper and faster homes, and token prices face real pressure. The caveat is capital: Puklowski admits financing will remain his full-time job, and a valuation built ahead of the first major deployments is a promise, not proof.
