Digital transformation: Inside AMDs AI factory vision – News

Artificial intelligence continues to accelerate digital transformation across industries, reshaping how organizations build and scale modern infrastructure. Advanced Micro Devices Inc. is positioning its technology to play a central role in this shift.

A key element in that evolution is the concept of the AI factory — data centers designed to handle and expand AI workloads. AMD’s strategy focuses on building a full-stack compute platform that bridges hardware and software to support this next phase of AI infrastructure, according to Anush Elangovan (pictured), vice president of AI software at AMD.

“To enact that, it all comes down to how you can get the compute required to power this AI infrastructure,” he said. “And in compute, AMD has had good, successive generational delivery of hardware, and now we are focusing on the software layer, so that you have a pervasive software layer that goes on top of the hardware layer that can then enable all of these AI factories to be built.”

Elangovan spoke with theCUBE’s John Furrier for theCUBE + NYSE Wired: AI Factories – Data Centers of the Future event series, during an exclusive broadcast on theCUBE, News Media’s livestreaming studio. They discussed how AMD is redefining data center architecture and building large-scale, software-defined supercomputers engineered to power the next wave of AI innovation.

Digital transformation driven by software-defined AI

AMD’s ROCm open-source software stack sits at the center of its AI factory strategy, creating a standardized layer that runs across EPYC processors, Instinct graphics processing units and Pensando network interface cards. The goal, according to Elangovan, is a “pervasive software layer” that allows developers and enterprises to tap compute power at any scale — from hyperscalers to enterprise data centers to edge devices.

“The first one is how we do an open ecosystem, how we participate in open source, open ecosystems,” he said. “We started with a clean slate of having a platform that anyone can contribute towards, and so all of our ROCm software is open source. We also have an open ecosystem, so that it’s not just the software; we want other companies and constituents to come into our ecosystem and move the ball forward in terms of innovation.”

One of AMD’s newest milestones is Helios, a rack-scale supercomputer introduced at the Open Compute Project Summit. Each Helios rack features 72 MI450 GPUs and delivers 1.4 exaFLOPs of FP8 performance, supported by 31 terabytes of HBM4 memory and up to 19 terabits per second of bandwidth, according to Elangovan.

“It’s compact enough for a data center, so that you can rack a bunch of these and go to gigawatt-scale,” he said. “We are now talking about gigawatt-scale deployments; one of these racks is literally a supercomputer. We also want to build that hardware layer, with a very robust software layer, in which you can build your AI factories, your industries and your software innovations on top of AMD hardware and software.”

AMD’s approach to integrating hardware and software emphasizes reliability, security and performance “built into the platform, not bolted on at the end,” according to Elangovan. The company’s open-source model reinforces that strategy, inviting developers and partners to expand the ROCm software ecosystem and advance AMD’s role in digital transformation.

“We have AMD EPYC processors, we have the AMD Instinct GPUs, we have the AMD Pensando NICs and most importantly, we have the AMD software, the ROCm software, on top of it to stitch all of these pieces together to build a supercomputer,” Elangovan said.

Here’s the complete video interview, part of News’s and theCUBE’s coverage of theCUBE + NYSE Wired: AI Factories – Data Centers of the Future event series:

Photo: News