Oracle offers cloud services with AI chips from AMD
Chip manufacturer AMD and software company Oracle are expanding their cooperation. Oracle will operate AI superclusters with AMD's MI450 GPUs from 2026.
(Image: JHVEPhoto/Shutterstock.com)
The US software manufacturer Oracle will offer publicly accessible cloud services from the second half of 2026, which will initially be powered by 50,000 AMD processors of the new MI450 Series and are to be expanded later. Both companies announced the expansion of their long-standing cooperation on Tuesday. Through the contract with Oracle, AMD gains a major customer for its upcoming chips, while Oracle can expand its processor offering.
"The demand for large-scale AI capacities is increasing rapidly, as next-generation AI models are pushing the boundaries of current AI clusters", it states in the companies' announcement. "Our customers are developing some of the world's most ambitious AI applications, and for that they need a robust, scalable, and high-performance infrastructure," said Mahesh Thiagarajan, Executive Vice President of Oracle Cloud Infrastructure (OCI).
OCI's planned new AI superclusters will therefore be operated with the AMD "Helios" rack design, which includes AMD Instinct MI450 Series GPUs, next-generation AMD EPYC CPUs codenamed "Venice", and next-generation AMD Pensando Advanced Networking codenamed "Vulcano". This rack-scale architecture offers "maximum performance, scalability, and energy efficiency for large-scale AI training and inference", according to the announcement.
Videos by heise
Numerous new deals
Last week, AMD already announced a multi-year agreement for the delivery of AI chips with a total capacity of six gigawatts to ChatGPT manufacturer OpenAI worth several billion US dollars. The agreement also includes the option for OpenAI to take an equity stake in AMD if various target milestones are reached. According to the news agency Reuters AMD had already been working with OpenAI to improve the design of its MI450 chips for AI applications. OpenAI, in turn, is reportedly building a one-gigawatt facility based on the processor next year.
With Oracle and other partners, OpenAI has in turn initiated the multi-billion dollar joint venture "Stargate Project". In this framework, the companies plan to invest 500 billion US dollars in new data centers for artificial intelligence in the coming years.
(akn)
