HawkInsight

  • Contact Us
  • App
  • English

Oracle Expands AI Ambitions with 50,000 AMD Chips Coming in 2026

On Tuesday, Oracle announced plans to deploy 50,000 units of AMD’s upcoming MI450 AI chip across its data centers starting in the third quarter of 2026."We feel like customers are going to take up AMD

On Tuesday, Oracle announced plans to deploy 50,000 units of AMD’s upcoming MI450 AI chip across its data centers starting in the third quarter of 2026.

"We feel like customers are going to take up AMD very, very well — especially in the inferencing space," said Karan Batta, Senior Vice President of Oracle Cloud Infrastructure, in an interview.

Following the announcement, AMD’s shares rose in pre-market trading.

The MI450 marks AMD’s first AI accelerator capable of being assembled into rack-scale systems, allowing 72 chips to operate in unison to power advanced AI training and inference workloads. OpenAI CEO Sam Altman appeared alongside AMD CEO Lisa Su at the product’s launch event.

Batta emphasized that AMD’s software stack is “critical,” and reiterated that “customers will be very proactive in adopting AMD, particularly for inferencing.”

The Oracle deployment follows OpenAI’s recent multi-year partnership with AMD.

Earlier this month, OpenAI announced a multi-year chip supply agreement with AMD to deploy a total of 6 gigawatts (GW) of processing capacity over the coming years, with the first 1 GW going online in 2026. If the project proceeds smoothly, OpenAI could eventually hold up to 160 million AMD shares—around 10% of the company’s total equity.

In addition, OpenAI signed a five-year cloud computing contract with Oracle in September, valued at up to $300 billion, further cementing their collaboration.

Industry data shows that AMD shipped roughly 100,000 AI chips in the second quarter of this year, compared with NVIDIA’s 1.5 million units over the same period. While NVIDIA still dominates more than 90% of the data center GPU market, Oracle’s partnership with AMD signals a growing move among enterprise customers toward multi-supplier strategies to secure computing capacity.

OpenAI, historically close to NVIDIA, has relied heavily on its GPUs for model training. However, its leadership has repeatedly stated that the company needs “as much compute as possible.” As a result, OpenAI is working with multiple chipmakers and has also partnered with Broadcom to co-design custom AI chips.

Oracle founder and chairman Larry Ellison is expected to deliver a keynote at the Oracle AI World conference on Tuesday, where he will discuss the company’s latest collaboration with OpenAI and outline Oracle’s strategic positioning against major cloud rivals Microsoft, Amazon, and Google.

Daniel Newman, CEO of Futurum Group, commented, “Oracle has already shown it is willing to place big bets and go all in to meet the AI moment. The company must now prove that beyond capacity, it can capitalize on its massive underlying data and enterprise capabilities … to add meaningful value to the enterprise AI wave.”

Disclaimer: The views in this article are from the original Creator and do not represent the views or position of Hawk Insight. The content of the article is for reference, communication and learning only, and does not constitute investment advice. If it involves copyright issues, please contact us for deletion.