Executive VP & GM of the Data Center Solutions Business Group, Advanced Micro Devices Inc.
The AI Ecosystem: AMD's Full Stack Data Center Strategy - Six Five Media at AMD Advancing AI
Six Five Media hosts Daniel Newman and Patrick Moorhead are at AMD Advancing AI where they’re joined by AMD's Forrest Norrod, Executive Vice President and General Manager, Data Center Solutions Business Group, for a conversation on AMD's strategic evolution toward becoming a data center solutions provider and the latest collaboration driving forward AI and chip ecosystems.
Their discussion covers:
An overview of AMD's evolved strategy from compute engines to comprehensive data center solutions
AMD's integration of AI, Instinct Accelerators, and EPYC CPUs, and the role of ecosystem collabora...
About Forrest Norrod
Forrest Norrod, executive vice president and general manager of AMD’s Data Center Solutions Business Group, presented the company’s first rack-scale AI system, Helios, in a July 2026 CNBC exclusive. Norrod described Helios as a 7,000-pound, $5 million system containing 72 GPUs and 18 CPUs per rack, and stated it is shipping to customers including Microsoft, Meta, OpenAI, and Oracle later this year. He characterized Helios as an open system built on industry standards, contrasting it with Nvidia’s proprietary approach, and said the system performs better, is more efficient, and is more customizable than Nvidia’s Grace Blackwell and Vera Rubin systems. Norrod also noted that optical interconnects will become prevalent within racks in the next two to three years, and said AMD would comply with regulatory regimes regarding sales to China. Norrod expressed confidence in AMD’s competitive position, stating that “there’s a serious case in which AMD does great and can get to 20 and 25%” market share, and that “it’s absolutely our aspiration to be able to gain market share.” He also observed a shift in the AI market away from “token maxing” toward practical token utilization, adding that if AMD architectures deliver better throughput and higher memory at a rationalized cost per token, “that stuff will start to win attention.”
Transcript (21 segments)
I
Interviewer0:11
Francisco at AMD's Advancing AI event, second annual event. I mean, guess what, it's all about AI: data center AI, client AI, and the supporting cast to make it happen: CPU, GPU, networking, NPUs. Great tech. We covered the gambit. It really did show AMD's wares across data center and client, and I think there were even a few surprises in there. It was just an overall really positive event for the overall perspective on AI and of course for the company.
Yeah, the event was primarily data center, and I need to pull in Forrest Norrod, who runs that business here at AMD, to talk about it. Forrest, welcome back.
F
Forrest Norrod1:10
Thank you so much. Super proud of what the team has done.
I
Interviewer1:15
Yeah, it was a really compelling mix. And I think you heard in a bit of the buildup there what AMD is, you know, in terms of data center and client, and there's this movement with AI PC and even networking. And so you have data center under your purview, right? So you had some pretty comprehensive announcements around EPYC, you had some pretty big announcements around Instinct, and of course a number of sizable announcements in networking, which I think was a little new, maybe even a little surprising to people. Talk about how you're sort of evolving from a company that's focused on moving many, many parts to a company that's really building systems and an entire AI stack for companies.
F
Forrest Norrod1:52
Yeah, well, we've been on a journey in the data center really for the data center into the rest of the gear. And so networking was the obvious next step in terms of the silicon that you need. You need a CPU, you need a GPU, you need a way to put it all together. But when you look beyond that, the complexity of the systems that we're building nowadays is getting so high. Just the power, the density, the challenges around building a 200-kilowatt rack with 70-plus GPUs in it and CPUs, etc., you have to start really thinking about it as you're not designing chips anymore. You have to think from the beginning that you're designing systems, because if you don't, you're going to screw it up.
I
Interviewer3:11
Yeah, it is interesting how it has evolved. Every huge inflection point in technology, whether it was mainframes to minis, minis to client-server, and then you add on different workloads. I remember when even database was putting stresses on memory, and by the way it still is, and storage. But having to have an alignment: CPU, and today GPU, storage, memory, and networking is important. And you talked about the simplicity of that. Can you talk a little bit about any other reasons than simplicity that people want to buy a complete solution?
F
Forrest Norrod4:10
They want to buy a solution. They don't want to assemble it themselves. Particularly for enterprises or tier-2 cloud players, they may not have the interest or the capability to assemble best-of-breed components into an integrated solution and make it work. And so they look at companies like AMD working with our partners Dell, HP, Lenovo, etc., to put together full-up solutions. That's what they want to buy. It's funny, I sometimes tell my team: nobody wants to buy a server CPU much. They don't want to buy a server. They want a solution to a problem. So I think that's the dominant reason. But beyond that, you really have to, if you put these things together, you can gain significant performance and efficiency advantages.
I
Interviewer5:12
It also sounds like it's really about have it your way, right? Because you're partnering with other people in the ecosystem. They might want to piece-part it, and you're open, so you're giving them open pathways to be able to do that. But you're also, it's almost like some people want the easy button, or the easier button. And at least what our research suggests is that even most training runs bomb out because of the network, and second is because there's maybe an issue with the GPU. And it sounds like what you're doing on the networking side is addressing that.
F
Forrest Norrod6:11
Yeah, that's right. I mean, networking is becoming an appreciated part of the overall problem of these big GPU clusters. And you're exactly right. When you think about these things that scale to 10,000, 20,000, 30,000 GPUs, you're going to have a failure every few hours. Just the laws of physics dictate that, no matter how resilient you are. So number one, you want to have high-reliability solutions to minimize those times. But then two, you have to have the ability to recover, roll back to a checkpoint if you absolutely have to. But we start to see errors accumulating. And so the Pensando technology, to be able to offer high-speed networking but in a fully programmable way that you can add high-value services on top of it, is critical.
I
Interviewer7:29
It also seems like a really substantial TAM opportunity as you look to expand, right? Because every one of these, the fastest ramp part with the new Instinct and the GPU product, but we've seen the value of systems and when it's created for this market, this industry, and the opportunity for AMD to provide more of the total solution is certainly going to increase the company's opportunity to drive more revenue and of course better outcomes for customers. I also think it's important to mention you made some announcements on EPYC.
F
Forrest Norrod8:12
I'd be remiss to not talk about EPYC. I know everyone wants to spend all the time talking about the AI chip, but CPU and GPU are another symbiotic thing.
I
Interviewer8:22
Yeah, and you've done a remarkably good job of winning the cloud providers. I think 50%, 60%, I've heard numbers as high as 80% with certain cloud providers on EPYC. You made some really great announcements there. Talk just a little bit about how that roadmap evolves, what you attribute so much success to, and do you think you can keep going? Is there more market to gain?
F
Forrest Norrod8:48
Yeah, I think, well first off, I'm incredibly proud of what the team has done. EPYC is an incredible product, a great whole series of products. It's a great roadmap. For the cloud providers, the data center is not a cost center, it's their factory. Their products are produced in the data center, and so making that data center more efficient, higher performance, translates immediately into cost of goods sold, cost per service, cost per user, cost per query, cost per YouTube video. And so the folks making selection on the server CPUs for the data center customers have a direct connection to the CEO. They're critical in driving business results, and so they're going to embrace the superior solution, devoid of any other considerations.
And I think there, the CIOs are much more concerned about risk, about business interruption. They don't want to disrupt the business. Their calculus in making a decision of whether to embrace something new is different. I think that over time, as we've continued to deliver, the acceptance of EPYC and the fear of the new is growing, the fear is diminishing. So I think we're starting to see more and more customers embrace EPYC in enterprise on-prem as well. And the other interesting thing is we've noticed, and maybe we should have seen it, that many enterprises are now in a hybrid environment. They all have some cloud deployment, and it makes it much easier to go back and talk to them about on-prem as well.
I
Interviewer11:22
Yeah, it's interesting behavior in the enterprise. You have some enterprises that say, 'Oh my gosh, some hot new technology is great in the cloud, I need to adopt that,' or 'I will adopt that in the future.' Others might think, 'Oh, that's not for me.' But ultimately they do end up at what the hyperscalers want to use. It might not look the same. I mean, even networking offload that you have with Pensando today had been in hyperscalers for years before it came to the enterprise. So I fully believe that you will have an even bigger opportunity there.
F
Forrest Norrod12:11
It's a headend, and then you've got a bunch of headless GPU servers, and it's a nice traffic cop out there. On the other side, history says that when things get mainstream, you want to integrate a lot of that onto the CPU for lower latency and lower power, either by putting blocks on there or algorithms that support it.
I
Interviewer12:37
What's your view of the future of the CPU in AI and its importance?
F
Forrest Norrod12:42
Yeah, I think the CPU can play a pretty important role in determining the performance of an AI system. Let's keep this simple. Let's just talk about the CPU feeding a GPU. There's a common misconception that the CPU doesn't matter. It's actually easy to disprove. We've shown that moving from, say, a Sapphire Rapids head node to an EPYC head node can give you 10 to 15% more performance on inference, by the way, for MI300, for AMD Instinct accelerators or Nvidia, and for some training applications we've shown 20% performance uplift, everything else held constant. All we've done is move a higher performance CPU in there. Why is that? It's because of Amdahl's Law. By speeding up that part of the algorithm that the CPU is executing, you get a significant overall gain. And as we play this forward, we'll design CPUs that are even better at feeding, accelerating, and orchestrating GPUs.
I
Interviewer14:19
Interesting. Not a lot of people know that. I think you sort of see it when you see these big systems that are combined with a CPU and GPU, and that's why I use the word symbiotic. Because I think we've actually seen it in the market. There are a number of companies, and it became a little taboo to talk about the CPU for AI, but we still know a lot of influencing is done on CPUs, including on EPYC. So that's true as well. Forrest, I just want to thank you so much for joining us. I know it's been a really busy day. Congratulations on all the announcements, and we look forward to hopefully having you back soon.
F
Forrest Norrod15:10
Thank you.
I
Interviewer15:10
For AMD in San Francisco, hit subscribe, join us for all of our content and coverage from the event. There was a lot of it, and be part of our community. We appreciate you, but we got to go for now. See you later.