Oracle Confirms Use of Cerebras AI Chips in Cloud Infrastructure

In a move that could bolster its standing ahead of a future public offering, AI chip specialist Cerebras Systems has secured a notable endorsement. During Oracle's quarterly earnings call, co-CEO Clay Magouyrk confirmed the tech giant's infrastructure utilizes Cerebras hardware, placing it alongside established accelerators from Nvidia and AMD.

Magouyrk described Oracle's approach as building flexible infrastructure capable of handling a wide range of workloads. "We continually offer the latest in accelerators, from the most recent Nvidia and AMD options to emerging designs from companies like Cerebras and Positron," he stated. This public acknowledgment by a major cloud provider marks a significant win for Cerebras, which designs large-scale WSE-3 chips for AI training.

The company, which withdrew an IPO filing last October before raising $1.1 billion in private funding, has been working to diversify its customer base. Its 2024 prospectus revealed a heavy reliance on a single Middle Eastern partner, G42, which accounted for 87% of revenue. Adding Oracle to its client roster follows another major announcement: a $10 billion commitment from OpenAI, which itself uses Oracle's cloud services. OpenAI and Cerebras also recently collaborated on a research preview of a fast-acting AI model for software development.

Oracle's call highlighted surging demand for AI infrastructure, with the company reporting a more than fourfold increase in remaining performance obligations to $553 billion. Magouyrk emphasized that strategic investments in data centers and compute capacity are expected to grow in value. He pointed to the rapid innovation in AI accelerators, noting that customers are focused on reducing both the cost and latency of AI inference, the process of running trained models.

While Cerebras did not immediately comment, and Oracle's public price list does not yet feature a Cerebras option, the co-CEO's remarks provide a clear signal of the chipmaker's growing reach within the competitive arena for high-performance AI computing.

Source: CNBC