Key takeaways
The partnership represents one of the largest computing infrastructure commitments in the AI industry as OpenAI seeks to address its ongoing shortage of computing resources.
The ChatGPT maker will use Cerebras' specialized wafer-scale processors to power its AI models, with the infrastructure rolling out in multiple stages through 2028.
Both companies emphasized the deal's focus on delivering faster response times for OpenAI's rapidly growing user base of more than 900 million weekly users.
Sachin Katti, who works on compute infrastructure at OpenAI, explained the strategic rationale behind the partnership in a company blog post.
"OpenAI's compute strategy is to build a resilient portfolio that matches the right systems to the right workloads," Katti said. "Cerebras adds a dedicated low-latency inference solution to our platform.
That means faster responses, more natural interactions, and a stronger foundation to scale real-time AI to many more people."
Speed and diversification drive partnership
The deal addresses two critical needs for OpenAI: faster AI inference speeds and reduced dependency on single suppliers.
Cerebras builds dinner plate-sized wafer-scale processors that integrate massive computing capacity, memory, and bandwidth on a single chip to eliminate bottlenecks in conventional GPU-based systems.
The company claims its systems can run AI models up to 15 times faster than traditional hardware.
Andrew Feldman, co-founder and CEO of Cerebras, framed the partnership's potential impact in transformative terms.
"Just as broadband transformed the internet, real-time inference will transform AI, enabling entirely new ways to build and interact with AI models," Feldman said.
The agreement helps OpenAI diversify its computing infrastructure beyond Nvidia, which currently dominates the AI chip market.
OpenAI has previously signed computing deals with AMD and maintains a preliminary agreement with Nvidia for up to 10 gigawatts of chips. The company is also developing custom chips with Broadcom.
Financial implications and industry context
The $10 billion commitment comes as OpenAI pursues aggressive infrastructure expansion while navigating concerns about its ability to finance massive computing contracts.
The company generated approximately $13 billion in revenue last year but has signed cloud computing contracts worth $600 billion with Amazon, Microsoft, Oracle, and other providers.
CEO Sam Altman has stated these deals are phased over time and will be funded through future revenue growth.
The partnership provides significant benefits for Cerebras as well.
The Sunnyvale-based startup had been heavily dependent on a single customer, UAE-based AI firm G42, which accounted for 87% of its revenue in the first half of 2024.
Cerebras filed for an initial public offering in 2024 but withdrew its plans, instead raising $1.1 billion in private funding at an $8.1 billion valuation.
The company is now reportedly in discussions to raise an additional $1 billion at a $22 billion valuation.
Notably, Altman is personally invested in Cerebras, and OpenAI reportedly considered acquiring the chipmaker as early as 2017.
The two companies have maintained regular discussions about potential collaboration since then, with negotiations for the current deal beginning last fall and reaching a term sheet by Thanksgiving.
The deal underscores the unprecedented infrastructure spending across the AI industry as companies race to build computing capacity for increasingly sophisticated models.
Read more:
Character.AI And Google Settle Landmark Lawsuits Over Teen Deaths Linked To AI Chatbots
Europe’s Labor Market Faces Crisis As AI Reshapes Hiring Amid Rising Unemployment
Pentagon Adopts Elon Musk’s Grok AI Amid Global Backlash Over Explicit Deepfakes