It is amazing how things have changed over the last few years. Back when I first started as an independent analyst, Intel was dominant, and AMD was almost an afterthought. Now AMD appears to be on a ramp-up to dominance while Intel is struggling and doing massive layoffs.
AMD’s incredible rise is largely due to the focus it has had since Dr. Lisa Su took over running the company and started kicking butt. I’m at AMD’s Advancing AI event today where Intel used to have its IDF event which had once assured Intel’s dominance.
AMD has led in attaching NPUs to X86 hardware with one of the most effective responses to Qualcomm’s equally impressive Snapdragon PC effort which effectively started the AI-PC revolution with Microsoft. AMD announced that it has over 150 developers creating for this platform, which is important given the lack of AI apps to use these new NPU-configured laptops.
But one presentation stood out to me from Forrest Norrod who leads AMD’s data center unit. He talked about AMD’s three-legged open approach to the data center, and I think it is a uniquely competitive approach to this vastly lucrative market.
Let’s talk about that this week.
CPU + GPU + Networking
AMD has largely led the market with its balanced CPU/GPU approach. NVIDIA focuses on GPUs, Intel on CPUs, and while NVIDIA attempted to buy ARM to fill its CPU gap, and Intel hired a GPU leader from AMD to create ARC, neither effort created the same level of balance that AMD has enjoyed for years.
The latest addition to AMD’s approach is networking, and it forms the third leg of the stool. Even Cisco was on stage praising AMD’s networking technology. Given this is common in AMD’s segment, you’d expect that AMD would lock out Intel and NVIDIA from its solution, but this isn’t AMD’s way.
Because Lisa Su was, as I was, at IBM when lock-in nearly killed the company, AMD is open. This means if a customer wants to use a non-AMD technology for any part of a solution, AMD will still work with them to make the result as performant as possible. This is core to AMD’s customer-first approach to solutions and is what has differentiated the company since Su took over.
This provides a unique level of flexibility to AMD’s solutions, and it is likely partially why it has had so many large OEMs on stage during the event, each one outspoken in praise about AMD’s products and approach.
The Power of Open
The problem with lock-in, closed approaches to product strategies is that it can create a false sense of security. Customers don’t stay with you because they want to, but increasingly stay because they simply can’t afford to leave. And your product teams can lose complete track of what the customers’ changing needs are. When the cost of staying exceeds the cost of leaving, they leave en masse and will avoid your products for some time after.
This happened to IBM in the early 90s, and it likely was as traumatic for Su as it was for me. With an open approach, every part of your solution must be competitive. You can’t rely on lock-in to force the sale of inferior products. Thus, product management for each unit focuses like a laser on being competitive, and this approach is paying huge dividends at the AMD AI event this week as every part of its GPU/CPU/Networking solution set isn’t just competitive but arguably market-leading when it comes to price performance.
AMD’s level of focus and management execution is being showcased at this event, and the level of integration and synergy between each of the legs of AMD’s AI solution is arguably unmatched as a result of this approach.
Wrapping Up
While the computer market was largely founded on locked-in strategies created by IBM, it is fascinating to not only see IBM abandon this policy but see IBM-trained executives like Dr. Lisa Su abandon them as well.
AMD is well positioned to play a major role in the massive data center move to AI. This move is incredibly disruptive and a lot of the tech companies around today likely won’t be once this trend concludes. This week, AMD convinced me it not only will survive this wave but should emerge as an even more powerful AI player in the end.
This shouldn’t be a surprise because success is often the result of creating a solid market-focused strategy and executing sharply on it. Impressive work.
