Related articles
TMTPOST -- Advanced Micro Devices Inc. (AMD) is stepping up efforts to catch up with its rival Nvidia Corporation with new collaboration to provide its new-generation artificial intelligence (AI) chips, freshing off a monumental deal with OpenAI.
AI Generated Image
AMD on Tuesday announced an expansion of a multi-generation collaboration with Oracle Corporation to help the latter』s cloud division, Oracle Cloud Infrastructure (OCI) launch its first publicly available AI supercluster. To build up the supercluster, OCI will deploy AMD』s Instinct MI450 series chips, with an initial deployment of 50,000 graphics processing units (GPUs) starting in the third quarter of 2026 and expanding in 2027 and beyond.
As AMD』s next-generation accelerators based on the CDNA 5 architecture, Instinct MI450 series are set to be introduced in the second half of 2026. The GPU will be made on TSMC』s N2 (2-nanometer) process technology, marking AMD』s first AI chips using TSMC』s leading-edge manufacturing process. AMD said Instinct MI450 series GPU will provide up to 432 GB of HBM4 and 20 TB/s of memory bandwidth, enabling customers to train and infer models that are 50% larger than previous generations entirely in-memory.
OCI』s planned AI supercluser will also be powered by AMD』s first rack-scale design, Helios. It can hold 72 chips to work together and includes AMD Instinct MI450 Series GPUs, next- generation AMD EPYC CPUs codenamed Venice, and next-generation AMD Pensando advanced networking codenamed Vulcano.
This announcement builds upon the joint work of Oracle and AMD to deliver AMD Instinct GPU platforms on OCI to end customers, beginning with the launch of AMD Instinct MI300X powered shapes in 2024 and extending to the general availability of OCI Compute with AMD Instinct MI355X GPUs. These will be available in the zettascale OCI Supercluster.
The new partnership with Oracle will further boost the market confidence for AMD, which last week unveiled a deal with one of AI leaders to compete with Nvidia to meet computing needs.
AMD and OpenAI reached a 6 gigawatt (GW) agreement to power OpenAI』s next-generation AI infrastructure across multiple generations of AMD Instinct graphics processing units (GPUs). The first 1 gigawatt deployment of AMD Instinct MI450 GPUs is set to begin in the second half of 2026, according to an OpenAI』s statement on October 6.
Under the agreement, OpenAI will work with AMD as a core strategic compute partner to drive large-scale deployments of AMD technology starting with the AMD Instinct MI450 series and rack-scale AI solutions, and extending to future generations. By sharing technical expertise to optimize their product roadmaps, AMD and OpenAI are deepening their multi-generational hardware and software collaboration that began with the MI300X and continued with the MI350X series.
OpenAI said the partnership creates a true win-win for both companies, enabling very large-scale AI deployments and advancing the entire ecosystem.
As part of the agreement, AMD has issued OpenAI a warrant for up to 160 million shares of AMD common stock at $0.01 per share, which was structured to vest as specific milestones are achieved. The first tranche vests with the initial 1 GW deployment, with additional tranches vesting as purchases scale up to 6 GWs. Vesting is further tied to AMD achieving certain share-price targets and to OpenAI achieving the technical and commercial milestones required to enable AMD deployments at scale.
If OpenAI exercises the full warrant, it could take over around 10% ownership in AMD, based on the current number of outstanding shares.
The deal with Oracle prompted Wolfe Research analysts to upgrade AMD's stock on Tuesday to outperform from peer perform. The analysts saw "a conservative path" to more than $10 in earnings power for AMD in 2027, based on their model of $15 billion in annual revenue from the AI startup. They also expected AM』s total AI revenue to reach $27 billion in 2027, and that revenue "could be higher" if AMD can ink more deals.
