Yes, Intel was disrupted. It was just disrupted by ARM. Originally the chips are too slow for normal use, but find a niche in low power devices. Over time the rough edges are smoothed out, and the underdog can make inroads in the high end marked while the incumbent isn't structured to compete in the low end very well, and their attempts fail.
> M1 designs aren't available as general CPUs for alternative uses other than on a mac.
M1 is just the capstone of that long process. This is sort of a wrinkle, but Apple's strategy means they can build a high margin product ("We don't strive to make the most popular product, we strive to make the best.") and not have to hand them over to by suppliers. Given the high margins M1 chip command when placed in Macs, it doesn't seem likely to pressure Intel
But make no mistake, Intel's margins are strongest on servers. Prices on Xeons are like 10x that for Cores. This is where the disruption is happening. Running macOS on M1 in AWS is neat but is probably for running build and test farms; Graviton is presumably the ARM chipset AWS customers might ditch Intel for. I've met teams that saved substantial money by switching, and that has to feed into demand for Xeons at some point.
The typical way a firm might survive an "attack from the bottom" is to chase after the high value compute. In 2022 that's AI / Tensorflow, where Nvidia rules and frankly, Intel underperforms. Hopefully Nvidia pays Mark Harris well because they probably owe him a few billion.
Intel CCG is at 9.3b revenue/2.8b profit this quarter (30% operational margin).
Intel DCG is at 6b revenue/1.7b profit this quarter (28% operational margin).
I remember hearing that Steve Jobs had originally asked intel to develop a CPU for the iPhone (Apple had a close relationship with intel back in the days when they switched from PowerPC.) I remember intel being on stage at Apple keynotes and Apple also got "first dibs" on a lot of new intel CPUs back then. But intel dind't believe it would be profitable enough for them to pursue.
Apple had been screwed before back when IBM wouldn't make them a power efficient G5 chip for laptops. Then intel wouldn't make them a power efficient phone CPU. So here we are today, Apple does it themselves. Had intel made a phone CPU for the iPhone, the M1 might not have ever existed.
Yeah, Intel had the opportunity to avoid this outcome, and they fumbled on the mobile sector. Ben Thompson gives a deep analysis of the history here: https://stratechery.com/2021/intel-problems/
I work for a Fortune 500 company. About 70% of the employees has a MacBook Pro. Until recently, all of them had an intel chip inside. Going forward they will have an M1. We are on a 3 year refresh cycle. So within 3 years the majority of company computers will be running M1. About 90% of company phones are iOS. If Apple starts using M1 in those…
I also work for a Fortune 500 company. Unless the employees are doing anything related to Apple ecosystem, they will be carrying Thinkpads instead, or using classical PC desktops, with external contractors having virtual workstations via the various cloud offerings.
Overall Apple desktop market across the world, is still around 12%.
Yeah. At tech companies and tech conferences, you get a pretty distorted view of the sorts of computers most people are using. Especially taking into account the fact that you probably also see a disproportionate number of people running Linux on Thinkpads in a lot of places, one might assume that Windows is barely used by looking at laptops at the typical event I attend.
it's not really windows marketshare that matters - it's the instruction set. x86 compatible instruction set is dominant today.
If apple really want to disrupt intel (and i guess by collateral damage, AMD), they will release the m1 CPU as a commodity. but they will also have to figure out how to get microsoft to play ball as well (which i am not sure they will).
Definitely not the case that the full lifecycle of disruption is complete since that entails the full replacement of the incumbent. But with M1 Apple silicon surpassed Intel in performance after many years of inferior performance improving at a greater rate, which I think is one of the key milestones in a disruption story.
I’m not sure if there are any case studies of one behemoth fully disrupting another? So who knows what the end state will look like.
One confounder for predicting the endgame is that due to their vertical integration with their OS, Apple won’t replace Intel where Windows is a hard requirement, and so they probably won’t completely decimate Intel. I suppose in this aspect it’s not a full disruption.
I’m not really clear how much inertia there is behind Windows these days, now that MS is committed to Office on Mac. Definitely substantial, but if the price-per-performance was 2x or better for macs in a couple generations, what % of corporate buyers would switch over?
For that, they'd first have to sell some. Outside of Macbooks, I'm not aware of a single premium Linux/Windows Laptop that runs on ARM. Until HP, Dell and Lenovo offer those, there won't be any uptick in ARM laptops outside of Macbooks. And most companies won't buy the first model running an ARM processor, they'll first want to see how driver support and virtualization of x86 apps (of which there are a lot in larger companies) work in reality.
The vast bulk of Windows laptops that go out the door go to companies that want them to be as boring as possible. This is probably the primary reason Linux desktops never made mainstream headway. Pretty much anything that increases support costs is largely a non-starter.
Intel was "disrupted" by their own doing when they missed the bus on the mobile revolution and the GPU revolution. And the M1 is very much a niche thing. It's an SoC, not a CPU, which is what Intel mainly produces, so comparisons of M1 and Intel processors are bound to be somewhat flawed.
