General Compute lands $400M loan backed by inference chips

General Compute, an AI inference cloud startup, has landed a $400 million loan from tech investment firm Upper90 — potentially the first deal to put up inference-specific chips as collateral, rather than the pricier chips used to train AI models in the first place. The financing reflects a broader market response to concerns over the cost of AI tools and tokens, as capital shifts toward infrastructure that runs open-source models more cheaply than the newest frontier-lab LLMs. General Compute, founded by CEO Finn Puklowski and CTO Jason Goodison, raised a $15 million seed round in May to build an inference neocloud — infrastructure purpose-built for AI workloads, unlike the general-purpose offerings of hyperscalers such as AWS or Azure — around chips from SambaNova, an Intel-backed chipmaker.

General Compute’s SN50 chips are built for inference. They are power-efficient, skip the expensive water-cooling systems GPUs require, and so can be deployed faster across a wider range of data centers. The company says the chips deliver 16 times faster inference than GPU-based clouds — though sourcing enough of them is a challenge for a brand-new company.

Upper90 co-founder and CEO Billy Libby, a former Goldman Sachs quantitative trader, had a playbook: in 2021 his firm financed GPU purchases by energy-focused data center startup Crusoe, in what he believes was the first loan made against the value of advanced chips. Traditional lenders had avoided such deals at the time over GPU-depreciation risk, but that changed once CoreWeave turned chip-backed loans into a business model and, later, the basis for a blockbuster IPO. Now, with GPUs comparatively well understood — and perhaps over-bought — Upper90 is betting on companies like General Compute to ride the next wave of the AI boom. “Everyone doesn’t need a supercomputer, but they do need inference and AI,” Libby said.

That thesis is gaining traction elsewhere: providers of open-model access like OpenRouter and Fireworks have raised new rounds at large valuations, Kimi’s K3 model has proven competitive with the latest Anthropic and OpenAI releases on coding benchmarks, and chipmakers Groq and Cerebras have drawn interest from acquirers and public markets alike. General Compute’s access to chips outside Nvidia’s ecosystem — echoed by AI infrastructure company TensorWave’s partnership with AMD — may give Nvidia-independent compute providers an edge on cost-efficient inference as alternatives to Nvidia scale up.

Why it matters

Puklowski frames the deal as more than a startup raising cash to buy compute: it signals capital organizing itself around the idea that Nvidia’s dominance over AI infrastructure is starting to fragment.