OpenAI has entered into a multi-year agreement with Cerebras, an AI chipmaker, to secure 750 megawatts of compute power, beginning this year and extending through 2028. A source familiar with the details told TechCrunch that the deal is valued at over $10 billion; Reuters also reported the deal size.
The collaboration aims to accelerate the speed at which OpenAI delivers outputs to its customers. According to a blog post by OpenAI, the increased compute capacity will reduce processing times for responses that currently require more time. Both companies emphasized that the deal would lead to faster outputs for OpenAI's customers.
Cerebras, founded over a decade ago, has seen increased attention since the rise of ChatGPT in 2022 and the subsequent surge in AI development. The company asserts that its systems, powered by chips specifically designed for AI tasks, outperform GPU-based systems, such as those offered by Nvidia. Andrew Feldman, co-founder and CEO of Cerebras, likened the impact of real-time inference to the transformative effect of broadband on the internet. "Just as broadband transformed the internet, real-time inference will transform AI," he said.
The deal underscores the growing demand for computational resources in the AI sector, driven by the increasing complexity and scale of AI models. The need for faster processing and real-time inference is becoming critical as AI applications expand across various industries, from customer service to scientific research.
Cerebras had filed for an IPO in 2024 but has postponed it multiple times. In the interim, the company has continued to raise significant capital. It was reported on Tuesday that the company was in talks.
Discussion
Join the conversation
Be the first to comment