General Compute, a startup specializing in AI inference cloud services, has secured a $400 million loan from Upper90, a technology investment firm. This deal may mark the first instance where inference-specific chips—designed to run pre-trained AI models quickly and efficiently—are used as collateral, as opposed to the more costly chips employed for model development. The financing underscores a growing market response to concerns about AI tool and token pricing, shifting toward infrastructure that operates open-source models at a lower cost than the newest large language models from leading labs. Founded by CEO Finn Puklowski, General Compute raised a $15 million seed round in May to establish an inference-focused neocloud using silicon from SambaNova, an Intel-backed chipmaker. Unlike traditional hyperscalers like AWS or Azure, which offer general-purpose infrastructure, neoclouds are tailored specifically for AI workloads.

The company’s SN50 chips are optimized for inference, offering power efficiency and eliminating the need for expensive water-cooling systems, enabling faster deployment across a broader range of data centers compared to GPUs. General Compute claims these new chips deliver 16 times faster inference than GPU-based clouds. The main challenge lies in acquiring a sufficient number of these chips, particularly for a nascent company. Upper90 co-founder and CEO Billy Libby, a former Goldman Sachs quantitative trader, applied a proven strategy: in 2021, his firm financed GPU purchases by Crusoe, an energy-focused data center startup, in what he believes was the first loan backed by the value of advanced chips. Traditional lenders had previously avoided such deals due to risks and uncertainties around GPU depreciation. However, as CoreWeave transformed chip-backed loans into a viable business model—culminating in a blockbuster IPO—this financing approach has become widespread. “We could really put together something as an early participant, and kind of get compensated for the risk,” Libby noted.

With GPUs now relatively well-understood and potentially overbought, Upper90 is turning to companies like General Compute to capitalize on the next phase of the AI boom. “We think open source models are going to be important, and we went and looked for a player last year that was in inference,” Libby said. “Everyone doesn’t need a supercomputer, but they do need inference and AI.” This thesis is gaining traction, as firms offering access to open models—such as OpenRouter and Fireworks—secure new funding rounds at substantial valuations. Recent models like Kimi’s K3, released just this week, have proven competitive with the latest from Anthropic and OpenAI on coding benchmarks. Additionally, emerging chipmakers like Groq and Cerebras are attracting interest from both acquirers and public markets. General Compute’s ability to source chips outside Nvidia’s ecosystem is crucial for the same reason. TensorWave, another AI infrastructure company, is similarly betting on a partnership with AMD. As more Nvidia alternatives emerge, compute providers not tied to Nvidia deals may gain an edge in delivering cost-efficient inference. “There are a bunch of chips that are starting to scale that have amazing [total cost of ownership], or that can operate much faster than Nvidia, but there’s not too many buyers for them,” Puklowski said. “By getting together with Upper90, this is not just, ‘a cool startup got some money to buy some compute.’ Like, this is the first signal of capital organizing itself and the fragmenting of Nvidia’s monopolistic dominance.”