General Compute announced a $100 million Series A co-led by Blackbird and Abstract on October 27, backing an AI-inference cloud whose first major deployments are planned for early 2027. TechCrunch reported an $800 million post-money valuation.
The startup supplies the computing capacity AI companies use to run models, combining GPUs with chips from Cerebras, SambaNova and Positron AI. It buys the hardware and installs it in rented data-center space; customers can rent dedicated servers or contract for a guaranteed rate of AI output.
The round adds equity alongside a committed debt facility of up to $400 million from Upper90, announced July 17. The original release set its starting size at $100 million, with expansion tied to customer demand.
In a September 29 announcement, General Compute described a multiyear Cerebras deal as its largest hardware commitment and the first deployment financed through the Upper90 facility. That capacity is scheduled to open in the first quarter of 2027.
The company plans to start the Cerebras service with coding-agent workloads. Its rationale is that those programs repeatedly call AI models while working on code, allowing small response delays to accumulate into longer waits for developers.