Most startups borrow money against things like buildings or equipment. General Compute is borrowing against something different: chips that are built for only one job.
What makes this deal unusual is what backs the loan. It is not a pile of Nvidia GPUs.
General Compute's cloud services are purpose-built for inference workloads. Its co-founder and CEO, Finn Puklowski, called the loan "the first signal of capital organizing itself and the fragmenting of Nvidia's monopolistic dominance."
Get the market news that matters in a five-minute read with Market Briefs, our free daily newsletter
Upper90 has financed chips before. Back in 2021, it gave loans to the data-center startup Crusoe for buying Nvidia GPUs.
Now the firm is betting on something else. Billy Libby, Upper90's co-founder and CEO, said the market for GPU loans was inefficient when they started. "We could really put together something as an early participant, and kind of get compensated for the risk," he said.
With this new loan, the firm is betting that the AI industry wants cheaper ways to run already trained models. Not every company needs a supercomputer to ask a model a question. Open-source models that run on specialized, power-efficient chips are becoming a popular alternative to the most expensive large language models.
Libby put it this way: "We think open source models are going to be important, and we went and looked for a player last year that was in inference. Everyone doesn't need a supercomputer, but they do need inference and AI."
The AI industry has historically focused on training large models using expensive GPU clusters. However, as models become widely deployed, inference - the process of running them - is where costs add up. Specialized inference chips like SambaNova's are designed to do this more efficiently, making them attractive for companies watching their budgets. This loan signals that lenders see value in these alternative architectures, potentially accelerating the shift away from Nvidia's near-monopoly in AI compute.
The loan marks an unusual financial move in the AI hardware space. While GPU-backed loans have been used before, this is one of the first major debt deals to leverage specialized inference chips. As AI models proliferate, the demand for cost-effective inference is rising, and lenders like Upper90 are seeking new ways to capitalize on the trend. The deal also underscores the growing importance of open-source models, which can run on less expensive hardware, potentially reshaping the AI supply chain.
This loan highlights a growing trend where investors are reacting to cost pressures in AI by backing infrastructure designed to run open-source models at lower expense. General Compute's ability to access chips outside of Nvidia's ecosystem matters for the same reason. As Nvidia faces more competition, providers not tied to its hardware could benefit when it comes to cheap inference.
Join Market Briefs, our free daily newsletter, for a quick daily rundown of the markets
