There’s a lot in this article that doesn’t merit too much discussion, but here’s a couple of interesting points.
Here’s the closest the actual article comes to making a thesis statement that matches the headline:
> But there is next to no chance Apple will replace Intel chips for ARM-based ones any time in the next five years.
“Never” vs. “in the next five years” is a pretty big difference, but the actual claim in the article was correct. The first M1 Macs came out in 2020, more than five years after 2012. The headline was wrong, but headlines are often not written by the writer of the original article. That seems weird and bad to me but that’s not the author’s fault here.
> Intel is the Apple of the microchip world: everyone is at least a year behind Intel when it comes to competing with their cutting-edge technology and design in the x86 space. Because of Intel’s manufacturing strengths, they can make the entirety of their chips smaller (and therefore more power-efficient) than anyone else around, not just their CPU cores.
It turned out that TSMC caught up to and surpassed Intel. It’s tempting to prophesy that Intel will never catch up to TSMC based on the performance of the past few years, but things can change over time.
Saying that TSMC caught up, whilst "true" is probably not the right way to look at it. Intel just slowed down to almost a halt. TSMC continued to do what it always did.
So it is possible for Intel to catch up. They always have had a more aggressive strategy.
>But Intel’s now seen the light. They have almost infinite resources to throw at the problem of catching up. And they’re going to do so quickly.
>“By 2014, Intel will have gotten their power management ahead of everyone else and be using their manufacturing muscle as a major advantage in the mobile landscape,” predicts Kanter.
Hm, I'd be interested in a postmortem analysis on why these predictions seemed realistic in 2012 and why they so drastically did not pan out over the following decade.
I think we can infer some of that from what Intel is doing now.
There are two orthogonal competitions happening here. One is between x86 and ARM, and the other is between Intel and TSMC. TSMC can manufacture x86 chips (this is what they do for AMD) but they can also churn out ARM chips for Apple and Amazon, GPU’s for nVidia, and so on. So while Intel’s manufacturing was vertically integrated with their x86 chip designs, TSMC specialized in fabrication, and got ahead of Intel.
Intel’s response has been to undo some of that vertical integration. Part of this entailed plans to outsource some of their manufacturing to TSMC, but another part was to reinvest in their own fabs and compete directly with TSMC by providing fabrication for other companies’ chips. This latter effort has received lots of political support. Maybe Intel simply decided their vertical integration strategy was unworkable, or maybe the recent chip shortage and the geopolitically vulnerable position of Taiwan has created a lot of political demand for a domestic alternative to TSMC that Intel is trying to capitalize on.
TSMC didn't get ahead because they were specialized. Intel just had many years of delays / failures depending on how you see it. TSMC traditionally and still are a slow risk-adverse entity and hence were usually behind the competition e.g. Intel.
Intel had always wanted to open up their fabs - want being the term here. It was just never executed very well in the past.
Even ignoring the chip shortage and any recent political demand there has always been a demand for alternatives e.g. Apple / nVidia going with TSMC / Samsung. You can never rely on a single provider.
I see Intel fixing their problems rather than having a major shift in strategy.
Thanks, I did read that over, but I'm more wondering about a hindsight for why the 2014 prediction seemed realistic, and why Intel failed to execute as predicted.
Interesting that the CoM followup doesn't make any mention of architecture.
Everyone always overestimates Intel's ability to do things. I worked for IBM in the P4 days and we were worried about 9-10 GHz pentiums. But Intel has a tendency to coast when it's doing well, and then only wakes up when AMD (and ARM also now) starts taking its lunch money.
Here’s the closest the actual article comes to making a thesis statement that matches the headline:
> But there is next to no chance Apple will replace Intel chips for ARM-based ones any time in the next five years.
“Never” vs. “in the next five years” is a pretty big difference, but the actual claim in the article was correct. The first M1 Macs came out in 2020, more than five years after 2012. The headline was wrong, but headlines are often not written by the writer of the original article. That seems weird and bad to me but that’s not the author’s fault here.
> Intel is the Apple of the microchip world: everyone is at least a year behind Intel when it comes to competing with their cutting-edge technology and design in the x86 space. Because of Intel’s manufacturing strengths, they can make the entirety of their chips smaller (and therefore more power-efficient) than anyone else around, not just their CPU cores.
It turned out that TSMC caught up to and surpassed Intel. It’s tempting to prophesy that Intel will never catch up to TSMC based on the performance of the past few years, but things can change over time.