General Compute Raises $100 Million to Rent Out Inference Chips From Nvidia's Rivals; TechCrunch Puts Its Valuation at $800 Million
General Compute, which runs an inference cloud on GPUs and on chips from Cerebras, SambaNova and Positron AI, announced a $100 million Series A on October 27, co-led by Blackbird and Abstract Ventures. TechCrunch reports that the round values the company at $800 million post-money, up from $60 million after a May seed round, and that it followed a deal with a large customer worth $100 million in annual recurring revenue. The company says its first major deployments go live in the first quarter of 2027; the valuation and customer figures are not in its own announcement.
General Compute has raised $100 million for a cloud built partly on inference chips from Nvidia's rivals, and TechCrunch frames the round as a rebuttal to a week of doubts about neoclouds.
The round#
General Compute says its $100 million Series A was co-led by Blackbird and Abstract Ventures, with participation from Fuse Venture Partners, Village Global and others. TechCrunch reports that the company was valued at $60 million after a May seed round and is now worth $800 million post-money to its investors. TechCrunch says a deal with a large customer worth $100 million in annual recurring revenue paved the way for the raise; it does not name the customer.
Inference chips in rented space#
The company says it finances, deploys and operates GPUs and purpose-built inference chips from Cerebras, SambaNova and Positron AI, and sells either dedicated bare metal or a guaranteed tokens-per-second service level. TechCrunch reports that its key customers are cloud companies offering developers high-speed inference, for which General Compute supplies the raw servers. Rather than building large data centers, TechCrunch says, it uses smaller power sources, rents space in other companies' data centers and converts former bitcoin-mining sites.
Two accounts of the $400 million and the megawatts#
General Compute's announcement says the round builds on a previously announced $400 million debt facility from Upper90. TechCrunch instead describes a $400 million deal to purchase new chips from SambaNova, which it calls the first debt financing of inference chips. The company says it has more than 50 megawatts of capacity in its pipeline. TechCrunch reports that CEO Finn Puklowski expects 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." Puklowski told TechCrunch that "capital is going to be a full-time job for me, both on debt and equity."
The neocloud argument#
TechCrunch set the announcement against a 12-month low in CoreWeave's share price and Elon Musk's prediction that "200+ neoclouds are going to see guppy-level death and consolidation," which TechCrunch says Musk made as he touted $40 billion of debt for buying computing capacity. Our earlier report covers Musk's post and SpaceX's $40 billion GPU lease commitment with an Apollo-managed entity. According to TechCrunch, Puklowski thinks Musk is talking his book to drive down the prices of GPUs he wants to buy, and he called fears of a chip-led collapse "massively overstated." He also told TechCrunch that Anthropic's public S-1 is "settling fears for today." James Palmer, a Blackbird principal who co-led the round, told TechCrunch that his firm's confidence comes partly from its investment in inference cloud company Baseten.
Limits#
- The $800 million valuation, the $60 million seed valuation and the $100 million ARR customer deal come from TechCrunch; General Compute's announcement does not state them, and TechCrunch does not name the customer.
- The company and TechCrunch describe the $400 million differently: a debt facility from Upper90 in the company's account, a SambaNova chip purchase in TechCrunch's. Neither source we read reconciles the two.
- The capacity figures are of different kinds: more than 50 megawatts in the pipeline (company) and 75 megawatts expected to be deployed in 2027 (Puklowski to TechCrunch). Neither is installed capacity.
- Musk's forecast and Puklowski's response are opinions, not data on neocloud failures.