AI Investment Shockwave: GPU Financiers Pivot to Inference Chips in $400M Deal

General Compute, an innovative AI inference cloud startup, has successfully secured a substantial $400 million loan from Upper90, a prominent tech investment firm. This financing landmark is notable for potentially being the first deal to utilize inference-specific chips as collateral. These specialized chips are meticulously designed to efficiently execute already trained AI models, distinguishing them from the more costly chips primarily used for developing AI models.

This significant financial injection signals a growing market response to escalating concerns regarding the high price of AI tools and tokens. The industry is increasingly pivoting towards infrastructure solutions that can run open-source AI models more economically than the latest large language models (LLMs) from leading AI research labs. Founded by CEO Finn Puklowski and CTO Jason Goodison, General Compute previously raised a $15 million seed round in May, aimed at establishing an inference neocloud built around silicon from SambaNova, an Intel-backed chipmaker. Unlike the general-purpose infrastructure offered by traditional hyperscalers such as AWS or Azure, neoclouds are purpose-built for specialized AI workloads.

The core of General Compute's offering lies in SambaNova's SN50 chips, which are explicitly engineered for inference tasks. These chips boast exceptional power efficiency and do not necessitate expensive water-cooling systems, enabling quicker deployment across a wider array of data centers compared to traditional GPUs. General Compute asserts that these new chips can deliver inference speeds up to 16 times faster than GPU-based clouds, offering a significant performance advantage.

A critical challenge for a nascent company like General Compute is securing a large volume of these advanced chips. Upper90's co-founder and CEO, Billy Libby, a former quantitative trader at Goldman Sachs, brought a proven strategy to the table. In 2021, Upper90 pioneered financing GPU purchases for Crusoe, an energy-focused data center startup, in what Libby believes was the inaugural loan backed by the value of advanced chips. At the time, conventional lenders were hesitant due to the inherent risks and uncertainties associated with GPU depreciation. However, as CoreWeave subsequently transformed chips-backed loans into a viable business model and eventually the foundation for a blockbuster IPO, this form of financing has become increasingly common within the tech landscape.

Libby explained to TechCrunch that when Upper90 initially financed Nvidia GPUs, the market was inefficient, allowing them to structure unique deals and be compensated for the early participant risk. With Nvidia GPUs now being relatively well-understood and perhaps oversaturated in the market, Upper90 is strategically shifting its focus to companies like General Compute, aiming to capitalize on the next phase of the AI boom. Libby emphasized Upper90's belief in the importance of open-source models and their active search for an inference-focused player last year, stating,