Oracle Cloud Infrastructure (OCI) will be a launch partner for the first publicly available AI supercluster powered by AMD’s upcoming Instinct MI450 Series GPUs — with an initial order of 50,000 of the chips to be deployed starting in Q3 2026 and expanding in 2027. The resulting Oracle installations will feature Instinct MI450s configured with AMD-designed CPUs in AMD’s new Helios server rack systems, positioned to compete with Nvidia’s Vera Rubin NVL144 CPX racks when both platforms are mass-released next year. Oracle is challenged to rapidly scale its data center capacity due to massive compute commitments made this year to OpenAI.

“Both AMD and Oracle have been on dealmaking sprees as the AI race heats up,” writes The Wall Street Journal, adding that “demand for computing infrastructure continues to surge and software developers are rushing to lock down access to the latest AI processors.”

Nvidia is estimated to have a 90 percent market share of the 2025 global AI accelerator market, and OCI is among those using a lot of Nvidia hardware in its facilities. But with demand outstripping supply for Nvidia’s most in-demand AI chips, Oracle has been incentivized to look elsewhere and began using AMD Instinct chips since 2024. This year, AMD made headway in the high-performance GPU space with its Instinct MI355X chip, which appears to perform competitively with Nvidia’s 2025 offerings.

The upcoming MI450s “are AMD’s first AI chips that can be assembled into a larger rack-sized system that enables 72 of the chips to work as one, which is needed to create and deploy the most advanced AI algorithms,” reports CNBC.

The AMD Helios rack design pairs Instinct MI450 Series GPUs with next- generation AMD EPYC CPUs (codenamed Venice), and advanced AMD Pensando networking, explains Oracle’s announcement. The vertically-optimized, rack-scale architecture “is designed to deliver maximum performance, scalability, and energy efficiency for large-scale AI training and inference.”

WSJ says this new deployment “will use the equivalent of 200 megawatts of computing power.”

“Customers want to set up computing clusters using each new generation of data-center chip ‘the day after it rolls off your manufacturing line,’” OCI EVP Mahesh Thiagarajan told WSJ.

Earlier this month, The Wall Street Journal wrote that OpenAI and AMD struck a five-year deal that will see OpenAI getting a 10 percent stake in AMD “if it hits milestones for chip deployment.”

CNBC runs down $1 trillion worth of AI deals, writing that OpenAI racked up that amount on paper this year alone. Last month, OpenAI struck a $300 billion deal with Oracle, though it is unclear whether there is any overlap between that and either Oracle’s or OpenAI’s agreements with AMD.