Oracle and AMD are collaborating to deliver a next-generation AI supercluster that for enterprises to train and deploy increasingly large-scale AI models.
Announced at Oracle AI World, the collaboration will make Oracle the first major cloud provider to offer a public AI supercluster powered by 50,000 AMD Instinct MI450 Series GPUs, beginning in Q3 of 2026.
The new system will combine AMD’s upcoming Helios rack design — featuring MI450 GPUs, EPYC Venice CPUs, and Pensando Vulcano networking — with Oracle Cloud Infrastructure’s (OCI) scalable architecture.
OCI will also roll out general availability for compute instances based on AMD’s Instinct MI355X GPUs as part of its zettascale supercluster.
The Oracle-AMD deal adds to AMD’s growing list of marquee AI partnerships. Just last week, the chipmaker announced a multi-year agreement with OpenAI to deploy six gigawatts of AMD GPUs, marking one of the largest GPU supply deals in the industry.
For Oracle, the announcement builds on a strong fiscal start to 2026 with Q1 revenue up 12 percent year-over-year to US$14.9 billion, driven by 28 percent growth in cloud sales to US$7.2 billion.
“By bringing together the latest AMD processor innovations with OCI’s secure, flexible platform and advanced networking powered by Oracle Acceleron, customers can push the boundaries with confidence,” said Mahesh Thiagarajan (top), Executive Vice President of Oracle Cloud Infrastructure.
“With our AMD Instinct GPUs, EPYC CPUs and advanced AMD Pensando networking, Oracle customers gain powerful new capabilities for training, fine-tuning and deploying the next generation of AI,” said Forrest Norrod, Executive Vice President and General Manager of Data Center Solutions Business Group at AMD.
The expanded partnership underscores the accelerating race among tech giants to scale AI compute resources, as next-generation models stretch the limits of existing infrastructure capacity.
Entelechy Asia 