Of all the companies I’ve covered in my career, none have ever sparked the same level of takeover discussion as AMD. Now, with CES in the rearview mirror, the rumors have raised their heads again, sparked apparently by Intel’s lack of a CEO, a very strong overall show for Team Green, and perceived differences in energy level between Intel’s presentation at the tech show and AMD’s overall position in the market.
EETimes spoke to a number of analysts about the possibility, most of whom dismissed it out of hand. “I think this was a landmark event,” Kevin Krewell, a principal analyst at Tirias Research, told EE Times. “It was quite dramatic for AMD to usurp the leadership role from Intel.” At the same time, Krewell told EET to file the idea of an Intel-AMD merger under “fiction.” Almost every analyst took a similar tone. So does ExtremeTech.
Why an Intel-AMD Merger Is a Non-Starter
An Intel-AMD merger would give one company complete control of the entire PC market. ARM is simply not a credible player in either the conventional server or consumer PC space to be considered an actual threat to Intel’s market dominance, and there are substantial barriers to entry (like full software stack compatibility) that prevent it from assuming that role in the near term. In order to justify a merger, AMD and Intel would have to argue that their fusion would encourage companies that currently have no plans to compete aggressively in the PC industry to begin doing so against an even more entrenched, more dominant player. As arguments go, this is a weak one. Regulators around the world would have an absolute field day with the concept.
Moreover, it’s not particularly clear what either company has that the other wants. Even if you believe that AMD’s CPU technology is superior to Intel’s, integrating AMD’s tech into Intel manufacturing wouldn’t be a simple process. Intel might prefer to buy AMD’s GPU IP to integrate into its own, but the larger firm has already sniped more than a few members of AMD’s GPU team to start with (and AMD isn’t going to sell off the IP that keeps it inside the Xbox and PlayStation). Intel, of course, is the one company that could buy AMD’s x86 business without any of the legal challenges that Chipzilla would promptly launch if anyone else bought AMD’s x86 business — and also the one company that all the governments and regulators would sue on account of it now having a complete monopoly on desktop and mobile computing.
The impetus for these rumors, more than anything, seems to be Lisa Su’s excellent overall CES performance and keynote presentation. Whether Intel might try to snipe AMD’s CEO is, of course, a very different question from whether it would try to buy AMD. And it’s not hard to miss the difference between Intel, which has had an acting CEO for months now, and AMD, whose CEO is a bit of a breakout star.
But in this case, it’s not even clear how much weight the overarching narrative can hold. Intel missed on 10nm and that miss has had a significant impact on the company’s roadmap for new products. This is a fact. It has also spent billions of dollars positioning itself to respond more adroitly to new emerging markets in 5G, IoT, self-driving cars, AI, and a host of computing-adjacent spaces. AMD, in contrast, has stuck mostly to its core competencies, with a moderate push into AI/ML markets with Radeon Instinct. Most of AMD’s efforts have gone into improving the competitive standing of its Ryzen, Ryzen Mobile, and Epyc CPUs.
Both of these strategies make good sense when you consider the relative earnings of the two companies, their overall market positions, and their near-term financial prospects. Intel, which already dominates the PC and server CPU space, is looking for new markets. AMD, which had been largely driven out of the PC and server space where it once competed, is busy winning back new share for itself. AMD’s 7nm Ryzen design is expected to win it new sales in consumer markets and servers, while Intel’s push into IoT, AI, and 5G are all less certain because these markets are themselves much newer.
Regardless of which company you favor or whose CPUs you use, there’s no arguing that the firms are now somewhat differently positioned when it comes to where they believe their own future revenue growth will come from. It’s not even clear how weak Intel’s overall position actually is. If the company’s 10nm ramp whiffs and its new product introductions falter, we’ll all be talking about how 10nm’s delays were the writing on the wall for yet further problems. If the 10nm ramp and launch go off without a hitch, competitive positioning in the foundry market could look very different in 12-18 months. For now, we don’t expect any significant shake-ups between the major players or big merger moves between one-time competitors.
- AMD Comments on Threadripper 2990WX Scheduling Issues
- Intel Optane Memory H10: Cache, NAND Flash on Single M.2 Device
- Intel at CES 2019: Ice Lake, 10nm Parts, New Network SoCs