AMD and Oracle Expand Cloud Partnership with E7 Instances and Pensando DPUs
The plan puts AMD CPUs and programmable networking hardware at the center of Oracle's cloud, with release timing still left to Oracle.
At Oracle AI World 2026, AMD outlined an expanded partnership under which Oracle Cloud Infrastructure will offer OCI Compute E7 instances built on 6th Gen AMD EPYC server processors, aimed at enterprise applications and the CPU workloads that sit alongside AI models. Oracle's Acceleron converged network interface cards will also carry AMD Pensando data processing units, which handle networking, storage and security tasks in hardware so that the host CPU stays free for customer applications. The E7 Acceleron Standard virtual machines will offer adjustable compute and memory configurations. AMD argues that scaling AI depends as much on CPUs that coordinate agents and retrieve data as on accelerators, and that programmable DPUs let Oracle change its networking and observability services as training, inference and agentic workloads shift. The announcement is a product direction statement, and Oracle notes that release timing remains at its discretion.
Why it matters
The announcement signals that the processors running alongside AI accelerators are getting as much attention as the accelerators themselves. For cloud customers, the practical shift is toward instances where networking, storage and security run in dedicated hardware rather than on the main processor. Because Oracle has said the timing of these products is at its discretion and this is a statement of direction, the change should be read as a stated intention rather than a delivered service.
Matters or hype?
Does this story matter, or is it hype? Vote before you see what everyone else said.