Skip to content
TechCrunch·

💡General Compute Secures $400M Loan for AI Inference Chips

AI inference chips get a new financial lifeline

TL;DR

General Compute has secured a $400 million loan from Upper90 to finance its AI inference chip business. This marks the first deal using such specialized chips as collateral and signals a shift in the market.

General Compute just landed a $400M loan from Upper90, making it the first AI inference startup to use its specialized chips as collateral. Why does this matter? It's a major step towards breaking Nvidia's stranglehold on the market. General Compute's SN50 chips offer 16x faster inference than GPU-based clouds and are designed for running already-trained models efficiently. The loan is not just financial; it’s a statement that alternative chipmakers can compete with Nvidia, potentially lowering costs for AI inference.

General Compute Secures $400M Loan for AI Inference Chips — TechCrunch

Key Points

1

General Compute secured a $400 million loan from Upper90 for its AI inference business, marking a first in the industry.

2

The SN50 chips by SambaNova used as collateral provide 16 times faster inference than GPU-based clouds.

3

Upper90’s CEO previously financed GPU purchases for Crusoe, an energy-focused data center startup.

4

CoreWeave made chip-backed loans a business model and then went public, showing the potential of this approach.

5

General Compute's chips offer alternatives to Nvidia's ecosystem, potentially lowering costs for AI inference.

Why It Matters

If you're running AI models on cloud services like AWS or Azure, General Compute’s loan signals cheaper options may be coming. The company's SN50 chips provide 16x faster inference than GPU-based clouds and are designed to run trained models efficiently. This could mean significant cost savings for large-scale deployments but only if your workload scales past the current Nvidia lock-in.

AIInference ChipsNvidiaGeneral ComputeUpper90

Frequently Asked Questions

Why does this matter?

If you're running AI models on cloud services like AWS or Azure, General Compute’s loan signals cheaper options may be coming. The company's SN50 chips provide 16x faster inference than GPU-based clouds and are designed to run trained models efficiently. This could mean significant cost savings for large-scale deployments but only if your workload scales past the current Nvidia lock-in.

What happened?

General Compute has secured a $400 million loan from Upper90 to finance its AI inference chip business. This marks the first deal using such specialized chips as collateral and signals a shift in the market.

Comments

Subscribe to join the conversation...

Be the first to comment

Enjoyed this article?

Get it daily. 7am. Free. Reads in 5 minutes.

Also get