How I Made $5000 in the Stock Market

Oracle to Deploy 50,000 AI Chips From AMD. The Stock Is Rising.

Oct 14, 2025 08:27:00 -0400 by Tae Kim | #AI

Advanced Micro Devices and Oracle are expanding their partnership. (Photograph by Chris Stowers/Bloomberg)

Key Points

Oracle and Advanced Micro Devices announced an expanded artificial-intelligence partnership Tuesday that will see Oracle deploy tens of thousands of AMD’s most advanced AI chips next year.

Oracle Cloud Infrastructure, the company’s cloud-computing arm, will be a launch partner for the first publicly available AI “supercluster” powered by AMD’s next-generation Instinct MI450 graphics processing units, or GPUs. The pair will deploy 50,000 chips starting in the third quarter of 2026 and expand the partnership in 2027 and beyond, the companies said.

“Our customers are building some of the world’s most ambitious AI applications, and that requires robust, scalable, and high-performance infrastructure,” said Mahesh Thiagarajan, executive vice president, Oracle Cloud Infrastructure.

AMD stock rose 0.8% in early trading Tuesday as the stock market fell. Oracle shares were down 3.8%. Nvidia , the leading maker of advanced AI chips, fell 4%.

Oracle Cloud Infrastructure has used earlier generations of AMD’s Instinct chips since 2024. The upcoming MI450 GPUs feature greater memory bandwidth for AI training models and are able to process more complex workloads, the companies said.

The future clusters will also incorporate AMD’s “Helios” rack server design, which incorporates 72 MI450 GPUs, AMD EPYC CPUs, and networking solutions, in a server rack.

Last week, AMD announced a long-term deal to become a key supplier to OpenAI’s artificial-intelligence infrastructure buildout. The agreement entails OpenAI deploying 6 gigawatts worth of AMD GPUs, starting with 1 gigawatt worth of rack-scale MI450 GPU servers in the second half of 2026.

OpenAI is the leader in AI technology with market-leading products such as ChatGPT, AI models for enterprises, and the recently released social video app Sora.

Nvidia’s currently available GB200 NVL72 is a rack-scale AI server incorporating 72 GPUs. The NVL72 handily outperformed AMD’s current non-rack-scale offerings for performance and efficiency in the newly-released SemiAnalysis InferenceMAX benchmarks. Inference is the process of generating answers from AI models.

But the OpenAI and Oracle deals may serve as some validation that AMD’s upcoming rack-scale servers will be more competitive with Nvidia late next year.

Write to Tae Kim at tae.kim@barrons.com and Nate Wolf at nate.wolf@barrons.com