General Compute Secures $400 Million Loan Backed by Inference Chips as Nvidia Alternatives Gain Ground
Updated
Updated · TechCrunch · Jul 17
General Compute Secures $400 Million Loan Backed by Inference Chips as Nvidia Alternatives Gain Ground
1 articles · Updated · TechCrunch · Jul 17
Summary
$400 million from Upper90 will fund General Compute’s AI inference cloud in what may be the first loan collateralized by inference-specific chips rather than training GPUs.
General Compute is building around SambaNova’s SN50 chips, which it says deliver 16 times faster inference than GPU-based clouds while using less power and avoiding costly water-cooling.
The financing reflects a broader shift toward cheaper infrastructure for running open-source models as buyers push back on expensive AI tokens and look beyond Nvidia’s ecosystem.
Upper90 previously financed GPU purchases for Crusoe in 2021, but now sees inference hardware as the next lending frontier as chip-backed financing becomes more accepted after CoreWeave’s rise.
General Compute, which raised a $15 million seed round in May, argues access to non-Nvidia chips could give inference providers a cost edge as rivals like TensorWave make similar bets with AMD.
With inference chips now used as loan collateral, is AI hardware the new gold standard for tech financing?
As AI's 'Unit Economics Crisis' deepens, can specialized hardware truly deliver profitable inference at scale?
$400M Asset-Backed Loan Signals New Era in AI Infrastructure: How General Compute and SambaNova Are Reshaping AI Hardware Financing
Overview
General Compute has secured a landmark $400 million loan from Upper90, uniquely backed by specialized SambaNova SN50 inference chips. This innovative financing marks a shift from traditional models, showing growing confidence in the value of specialized AI hardware. By leveraging these chips, General Compute aims to challenge Nvidia’s dominance and deliver more cost-effective AI solutions. The deal highlights a broader industry move toward asset-backed lending, as demand for high-performance, scalable AI infrastructure grows. Success will depend on General Compute’s ability to maintain high performance and scale its neocloud platform to meet increasing needs in the evolving AI landscape.