OpenAI Inks $10 Billion Deal for Cerebras AI Chips to Power Real-Time Inference

The AI compute war just reached a $10 billion boiling point. OpenAI is betting big on Cerebras to break the NVIDIA bottleneck. On Wednesday, the AI giant announced a multi-year agreement to secure massive compute resources, signaling a shift in its hardware strategy.

OpenAI Cerebras $10 Billion Compute Deal: Scaling Real-Time AI

According to sources familiar with the matter, the deal is valued at over $10 billion. Cerebras will deliver 750 megawatts of compute power to OpenAI starting this year through 2028. The primary focus is speed. OpenAI stated these systems'd accelerate responses that currently face processing bottlenecks, making AI interactions feel more natural and instantaneous.

Sponsored

Advertise with Us

[email protected]

Sponsored

Beyond the GPU: Why Cerebras Matters

Cerebras co-founder and CEO Andrew Feldman noted that just as broadband transformed the internet, "real-time inference will transform AI." The company's Wafer-Scale Engine is designed for low-latency tasks, claiming superior performance over standard GPU-based clusters. This partnership is particularly notable as Sam Altman is an early investor in the firm. While Cerebras has pushed back its IPO multiple times, it's reportedly in talks to raise another $1 billion at a $22 billion valuation.