What I am about to write in this article, isn't entirely new. Some of these fears, in isolation, have been around for a few years. Others have not. However, there's something unique in the present. The uniqueness is in that these fears are now all banding together.
Said another way, each of these threats might not actually materialize. But all of them not materializing is unlikely. The problem here, of course, is that with Intel (NASDAQ:INTC) being such a dominant force in the market, a single one of these threats materializing is enough to make it a bad investment over the coming years.
So what are these threats and what's new about them? Let's go over each one.
The Process Advantage Threat
For the longest time, Intel has had a significant manufacturing advantage over its competitors, both in and outside the x86 market. This manufacturing advantage came from Intel always being 1-2 process nodes ahead of everyone else, plus it having a technological advantage when it came to the basic unit of construction, the transistor itself.
Being 1-2 process nodes ahead, meant that Intel's transistors were smaller and more efficient. It meant that Intel could get more chips out of the same silicon wafer area. This manufacturing advantage translated into three things:
- Intel had a cost advantage, because of getting more chips from the same number of wafers.
- Intel had a power usage advantage, because smaller transistors took less current to drive, and this was also helped by its technology advantage at the transistor level.
- Intel had a performance advantage, since lower power usage per transistor meant there was less heat to be dissipated, and lower area usage meant Intel could stick more transistors into the same or less space. This meant Intel could drive the transistors at higher frequencies or have more of them, granting it an architectural advantage. Both routes led to greater performance.
The thing is, over the last couple of years Intel's advantage has slowly narrowed. Intel delayed its jump to the 10nm process, while others - forced by cut-throat mobile competition - had no luxury of delaying anything. The result? Well:
- Samsung has just said it is starting 10nm mass production.
- TSMC (NYSE:TSM) has just said it will be the first to 7nm production, expecting risk production in Q1 2017.
Intel, for its part, has put its tick-tock strategy on hold and just started shipping Kaby Lake CPUs. These are, uncharacteristically, the 3rd CPU generation on Intel's 14nm process. At the same time, Intel is said to have started 10nm trial production, but commercial products are only expected for H2 2017.
There are debates on whether Intel's 10nm process is actually comparable to Samsung's or TSMC's. Arguably, some would say, Samsung's 10nm is more like Intel's 14nm, and TSMC's 7nm will be more like Intel's 10nm. But as Intel shifts further and further back, we can now say that at best, Intel has been caught in process tech. At worst, it already lags.
This, obviously, will be significant. The three advantages listed above are about to be gone. Once those advantages go, it's only a matter of time until competition heats up and margins decline.
The Advanced Micro Devices (NASDAQ:AMD) Threat
I've already written on this threat in my article titled "You Can't Believe AMD And Intel At The Same Time". For the longest of times, AMD has been bleeding market share to Intel. It completely lost the server and high-end markets, and was stuck trying to sell 4 and 8 core chips against 2 and 4 core Intel chips on the low end.
AMD is especially important because it's the one other relevant player shipping x86 chips. The news regarding 2017, is that it will finally be taking an architectural leap with its new Zen cores. On top of this, the Zen cores will be on chips made using the 14nm process, and right after that, they'll be made on a 7nm process while Intel will still be at 10nm.
There is the chance that these Zen-based chips will be competitive with Intel's offering. Even if the first generation of Zen chips are not competitive (something which I find likely), there is a chance that architectural and process improvements will make the 2nd generation Zen+ chips competitive.
Any increase in competitiveness by AMD will detract directly from Intel's market share. Worse still, any increase in competitiveness will not just remove market share and associated sales quantities, but it will also affect the remaining share through pressure on prices and margins.
The ARM Threat
We've heard about how ARM-based CPUs could pose a threat for years now. It never materialized. However, in the last 2 years something did change materially. Apple (NASDAQ:AAPL) pushed its own SoC designs so hard, that they can no longer be ignored as a threat.
As a measure of this threat, I am going to point to Geekbench 4.0 benchmarks on Apple's A10 chip powering the iPhone 7. The A10 scores 3490 on the single-core benchmark. More impressively still, leaked A10X benchmarks point to the A10X bringing this up to 4236. This compares to an Intel Core m7-6Y75 @ 1.30 GHz at 3677, or an Intel Core i7-6700K @ 4.00 GHz at ~6603. What does this mean? It means:
- A smartphone-class ARM CPU already has the single-core performance of a top Intel notebook CPU. A tablet-class ARM CPU already exceeds it.
- Given the different power and cooling envelopes, it's likely that the same mobile core can be pushed to desktop-class performance.
- Hence, it's a matter of time until Apple substitutes its Mac Intel CPUs for its own, internal, SoCs. This is something I talked about in my article titled "Intel Should Worry A Lot If Apple Excludes It From Future Macs", but which is now much more likely both due to architectural and process improvements. This event alone would represent the removal of 8% of the entire PC market from Intel's hands. And it's going to happen.
- It gets worse. As Apple pushes the envelope, other smartphone and smartphone SoC makers are forced to follow. Samsung, for instance, is already going towards custom ARM cores, the same way Apple did. So is Qualcomm. They will quickly reach those performance levels as well. And when they do, it becomes a matter of time until things like Chromebooks or ARM-based servers start taking away what are today Intel workloads.
As an added threat, the ARM world lives on power optimization, due to it being a central demand of the mobile market. As it happens, today's PC world also requires power optimization. Notebooks require it for battery life, server farms require it for economic reasons. Thus, while AMD might have difficulty matching Intel when it comes to power efficiency, the ARM-based competitors likely won't have such trouble.
In the next 1-3 years, Intel is likely to see massively increased competition. Intel isn't particularly cheap going into this. It trades at ~13.5x 2017 earnings estimates and 8x TTM EV/EBITDA. Sure, there's a profitability bump coming from abandoning some of its mobile ambitions, but that's at least partially already in the 2017 numbers.
The threats above, however, should be able to reduce Intel's profitability in a material way. AMD becoming more competitive can impact quantities sold as well as prices and margins. Apple moving to its own CPUs in its Mac lines would have an instant impact. Losing the process advantage can make both threats more serious. For instance, one can only wonder what AMD will do with a half-competitive architecture running a process node ahead of Intel.
In short, Intel right now is far from a safe investment. Arguably, given the near-term nature of some of these threats, it might even be a straight short.
Oh, and by the way, the Apple move is already rumored to be in testing. If and when it comes, it will have to be by surprise, as such announcement would tank Apple's existing x86-based sales. This could be a 2017 event.
Disclosure: I/we have no positions in any stocks mentioned, but may initiate a short position in INTC over the next 72 hours.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.