AI infrastructure startup General Compute has raised $100 million in Series A funding, reaching an $800 million post-money valuation. Blackbird Ventures and Abstract Ventures led the round, which follows a major customer agreement valued at $100 million in annual recurring revenue.
The company is building its cloud around specialized chips designed for AI inference--the process of running trained models. Its suppliers include SambaNova, Positron and Cerebras. General Compute previously agreed to spend $400 million on chips from SambaNova.
A flexible approach to AI computing
Rather than constructing large data centers from the ground up, General Compute rents space in existing facilities, uses smaller power sources and is adapting former bitcoin-mining sites for AI workloads. It supplies raw servers to cloud providers that serve developers seeking fast inference capacity.
The approach gives the startup a way to bring different chip technologies to market without taking on the full infrastructure burden of building major data centers. General Compute expects to deploy 75 megawatts of computing capacity in 2027 and says it may have access to as much as 200 megawatts of power.
Demand for inference computing is driving interest in alternatives to conventional GPU-based systems. By connecting specialized chipmakers with cloud providers and customers, General Compute is positioning itself as an infrastructure partner in a rapidly evolving market.
If its model scales, a wider mix of chips and more adaptable data-center capacity could help broaden the options available for running AI systems.
