Neocloud General Compute, which specializes in bare-metal infrastructure, raised a $100M Series A led by Blackbird and Abstract at an $800M post-money valuation

Bare-metal inference compute costs decrease for AI developers.

General Compute, a company specializing in bare-metal infrastructure for AI inference, has secured $100 million in Series A funding. The round was led by Blackbird Ventures and Abstract Ventures and values the company at $800 million post-money. This funding comes despite broader market concerns about a potential neocloud collapse, with some predicting significant consolidation. General Compute's strategy focuses on utilizing specialized inference chips from Nvidia rivals, including SambaNova, Positron, and Cerebras, and employs a lean operational model. This approach involves renting data center space and repurposing former Bitcoin mining sites, rather than building massive data centers from scratch. The company also announced a $400 million deal for inference chips from SambaNova and a $100 million deal with a large customer, contributing to its rapid growth and valuation increase from a $60 million post-money valuation in its May seed round.

General Compute's success signals a potential shift in AI infrastructure strategy, moving away from a sole reliance on GPUs and hyperscalers towards specialized inference hardware and more flexible, distributed data center models. This diversification in chip supply and infrastructure approach could foster greater competition and innovation in the AI cloud market, potentially lowering costs for AI development and deployment.

The company's focus on inference chips and efficient, non-traditional data center builds addresses a growing demand for specialized AI compute that is often more cost-effective than general-purpose GPUs. This approach may also appeal to companies seeking greater control over their AI hardware and data, potentially leading to new partnership opportunities between chip manufacturers, infrastructure providers, and AI model developers.