OpenAI (The owners of the renowned AI platform, ChatGPT) signed a multi-year partnership with chipmaker Cerebras on January 14, 2026, valued at approximately $10 billion.
This deal is specifically designed to bolster OpenAI’s high-speed AI inference capabilities as it scales to reach over a billion users.

Key details of the agreement include:

🔹️Compute Capacity: Cerebras will deliver 750 megawatts of power to OpenAI’s inference-as-a-service platform through 2028.

🔹️Ultra-Low Latency: The partnership utilizes Cerebras’ wafer-scale “dinner-plate-sized” chips, which are optimized for real-time AI tasks such as voice interactions, code generation, and image creation.

🔹️Phased Rollout: Deployment will begin in 2026 and continue in tranches through 2028, representing the largest high-speed AI inference deployment globally to date.

🔹️Strategic Diversification: This deal significantly reduces OpenAI’s reliance on Nvidia and Microsoft for compute infrastructure.

For Cerebras, the agreement provides a massive revenue base to diversify away from its previous primary client, UAE’s G42, ahead of its planned 2026 IPO.