While there are many interesting companies earnings to discuss, and we will hopefully get to all of them, Intel’s stood out to me the most. It is no secret my concern for Intel has been growing, but I remained cautiously optimistic up until this point. On the surface, most everything looks great. They just hired one of the best chip designers in the world in Jim Keller, who came from Tesla, but before that was responsible for amazing work at AMD. Intel also hired Raja Koduri who is one of the leading engineers in graphics design and worked with Jim at AMD as the head of Radeon’s engineering team. These two are rockstars in the silicon industry on every level.
Intel also posted a great second quarter result. Largely driven by the data center group which posted the highest growth of any division at 24%. This group will inevitably become Intel’s largest group from a revenue perspective, overtaking the current leader which is the client computing group which consists mostly of PCs but also smartphones. Intel’s future is in data centers, which can be a lucrative and profitable business. But there was one thing to come out of Intel’s announcements last week that is the most concerning, and I believe, the indicator that disruption has set its sights on Intel.
Lost the Lead
My thesis on Intel, which I developed over the years talking with management, was that as long as Intel could maintain technology leadership from an architecture and process technology standpoint, then they would be fine. Intel, like Samsung, TSMC, Samsung, and Global Foundries, all create proprietary process designs, and each is working to making their process more efficient by pushing the envelope in quantum physics to get their technology smaller and smaller. This is measured in nanometers, and Intel has currently been operating at the 14nm process while Samsung and TSMC are moving to 10nm and 7nm processes.
Intel used to be the undisputed leader in moving to new process technology roughly every two years. Our first signs of worry were when Intel moved from a two-year cadence to a 2.5-year cadence in transitioning to new process technologies. But even then I remained confident the quality of Intel’s proprietary designs were better than the competitions. Third-party experts were beginning to claim Samsung, and TSMC to a degree were starting to close the gap on Intel regarding quality and transistor design and density. But the common wisdom was as long as Intel didn’t have more dramatic delays then they would still be in a leadership position.
In last week’s earnings call, Intel CEO Brian Krzanich announced that Intel would be further delaying their ETA for volume production of chips relying on its next-gen 10-nanometer (10nm) manufacturing process from the second half of 2018 to 2019. If I had to bet I would bet more toward 2019 than 2018. For those of us paying attention, Intel had trouble getting to 14nm and indicated to both investors and us analysts that getting to 14nm was very challenging but they projected confidence those same struggles would not apply to 10nm. On the earnings call, Krzanich said almost the same thing verbatim insisting getting to the challenges they are facing getting to 10nm wouldn’t be there when they try to go to 7nm. Intel may believe this, and they need to believe it, but I’m extremely certain Intel will have the same challenges getting 7nm as they did getting to 14nm and getting to 10nm. And if I had to bet, I’d bet their 7nm transition takes 3.5 to 4 years. Which means, in a nutshell, Intel has lost process technology leadership. The only thing they had going for their future.
Arm’s Disruption is Setting In
Arm, as a technology, is disruptive by nature. It is an enabling disruptor to Intel and always has been. However, Intel has fought it off by maintaining their lead in process technology leadership. With that lead all but gone, the Arm ecosystem will inevitably pounce and hit Intel from all directions. Qualcomm is already making strides with a product portfolio of chips that will go into PCs, Chromebooks, tablets, and servers that can hit Intel where it hurts. New silicon startups are popping up in China, Israel (many ex-Intel folks in Israel who left to start these companies), and the even US are using Arm, and other technology licenses to fill gaps in the market and pick off parts of Intel’s core value propositions. NVIDIA similarly is eating Intel’s lunch when it comes to graphics in data centers and new industries like automotive. Large in part to Nvidia’s laser focus on machine learning and AI, spending tens of billions of dollars in R&D on this focus alone in ways no other company can.
Arm has enabled and will continue to enable an army of competitors to Intel they simply can’t compete with. And having lost process technology leadership, making chips at Samsung fabs or TSMC, even Global Foundries, will mean the competition’s designs will be on par and in some cases better than Intel’s. The competition doesn’t have tens of billions of dollars in manufacturing facilities so by being a fabless designer means many of these companies can charge less than Intel for their products. Meaning as good, or better quality, at lower prices. Quite the recipe for disruption.
I have gone from optimistic to extremely concerned about Intel’s future. I hope they can change my mind, but with what, remains to be seen. Intel’s unraveling won’t happen overnight. Disruption is a process not, an event. But I don’t see the fundamentals that kept Intel in a leadership position changing and ultimately that is not a good position for them to be in.