My bet is on their own fab(s) at some near-future point. They are one of the few who can afford the infrastructure (buildings/equipment) and talent, and they totally rely on a supply chain that is currently dubious, and has the real potential to get a lot worse. And it would fit into all current and likely all future projects Apple does and would likely do since they have proven they are more than capable of designing their own chips.
Hmm. As a chip designer, I think there is no chance this will happen. Running a modern foundry is a huge investment and while they move a lot of silicon volume, it is not enough to support a foundry. Not even close. Second, a foundry is a relatively low margin, high volume, high risk business. That is not the kind of business area that Apple enters. Third, running a modern foundry is an incredibly complex process with a huge level of specialized knowledge requiring years of experience to make it all reliably work with high yield. As someone who has been in the business for over 30 years, I won't hold my breath waiting for Apple to build their own semi foundry.
On the other hand, might they contribute some investment to help TSMC expand into the US or other locales to further improve their priority and pricing? That is likely to happen. They often make large investments to help their partners spool up and meet the required volumes for Apple business.
It makes logical sense from a supply chain standpoint, but I fail to see how the fact that Apple is good at designing chips has any relevance whatsoever to the idea that Apple may or may not be good at manufacturing those same chips.
"Everybody knows" that a big chunk of the M1/mac performance dominance right now is due to their buying out huge amounts of TSMC capacity at leading edge nodes. Thus, much of the recent M1/M1Pro/M1Max "great leap forward" (pun intended) are directly attributable to TSMC specifically.
If every ROI-seeking multibilliondollar hedge fund in the world can't build a leading edge node fab or two to compete with TSMC, what suggests that Apple's money would do any better?
Note that this isn't rhetoric - Apple may well be able to do this, and better than some/most; I just don't have any data at hand (myself, it may exist) to suggest that the Apple cash reserves are in any way smarter at this task than big institutional dumb money reserves (which dwarf Apple in market aggregate).
A decade and change prior we could have said the same thing about doing in-house chip design instead of just buying off-the-shelf Samsung parts. Or designing their own GPU technology instead of just licensing PowerVR cores. Apple did not have those skills internally and there was no guarantee that those projects would pan out.
If there is one thing Apple is actually "good at", it's hiring people who are good at things they don't specialize in, and using those people to build out their own capabilities. So if Tim Cook decided that they need in-house fabs, they would use their massive cash reserves to hire and buy whoever and whatever they need to set up in-house fabs.
Apple's best bet here would be to co-fund TSMC building fabs just dedicated to Apple, which they would operate as a joint venture or whatever arrangement makes the most sense. It wouldn't make any sense to go it alone on new fabs when they have the tremendous expertise of TSMC to utilize. It might also be reasonable to distribute the fabs as much as possible to different locations around the world, rather than concentrating them near China / in Taiwan (put them in the US, in Europe, in Asia).
It sometimes certainly works strategically to deprive other companies of chips by buying up TSMC's supply, however that can also obviously work against Apple in a big way. If you really believe in your products and the demand for them, you'd want the dedicated supply from fabs you co-own.
You’re close to describing the current arrangement, a mutually profitable agreement to develop and use new capacity. There wouldn’t be much to gain from TSMC not owning the fab because TSMC is better at selling the older process capacity to other chipmakers, which lets TMSC offer favorable terms to Apple. For Apple to own fab would essentially mean for Apple to build a slightly smaller clone of the entire TSMC organization.
If Apple were to start doing their own fab they would likely start with some of the simpler support chips and work their way up in complexity through networking before getting to the CPU. This fits their historical pattern of quiet experimentation with new technologies well ahead of large investments. Bringing PMIC and Bluetooth fully in-house, for example, could yield benefits in supply predictability and give them more design freedom for low power wearables. Later they might do the cell modem. Main CPU is a harder sell, letting an external partner like TSMC handle it allows them to move on quickly while the partner sells the older process capacity to AMD, Intel, etc.
Not the most likely scenario in my opinion, but it's how they would do it.
> "Everybody knows" that a big chunk of the M1/mac performance dominance right now is due to their buying out huge amounts of TSMC capacity at leading edge nodes. Thus, much of the recent M1/M1Pro/M1Max "great leap forward" (pun intended) are directly attributable to TSMC specifically.
That sounds like good common sense, but you don't get the huge leaps in performance that we're seeing with single process advantage (7nm vs 5nm). Qualcomm has 5nm process ARM chips and they barely compete against what apple was shipping on 7nm ~2yrs ago.
> but I fail to see how the fact that Apple is good at designing chips has any relevance whatsoever to the idea that Apple may or may not be good at manufacturing those same chips.
> If every ROI-seeking multibilliondollar hedge fund in the world can't build a leading edge node fab or two to compete with TSMC, what suggests that Apple's money would do any better?
Apple had 0 chip designers working for them not so long ago. They saw the writing on the wall that they were not going to be able to get where they wanted to go, as quickly as they wanted to, relying on a company that didn't see the writing on the wall, and falling further and further behind. Apple had hardware engineers to design their products using chips from other companies, and software engineers to make them do something useful, and design engineers to make them pretty. Then they decided to hire some chip engineers, some really good ones at the top, because they can afford to and they had a will to and concrete purpose. On their very first attempt at a new cpu, they were running with the best of them. How long has Apple been designing chips and how does it stack up against Intels cream of the crop and how long have each been at it? I do not see why Apple would not succeed if they put their minds, money and leadership style to it.
> "Everybody knows" that a big chunk of the M1/mac performance dominance right now is due to their buying out huge amounts of TSMC capacity at leading edge nodes. Thus, much of the recent M1/M1Pro/M1Max "great leap forward" (pun intended) are directly attributable to TSMC specifically.
Right, which is exactly why I think they will start to rely less on something that is only manufactured/available in a single political hotbed of a country, Taiwan, because Apple is now reliant on the best/most advanced process only offered by TSMC/Taiwan. Those leading edge nodes are not available to TSMC plants in any other country that I'm aware of. The crown jewels are kept in-country, under tight control. China is very vocally stating that Taiwan will be "reunited" with China and being fairly hostile about it. Apple might be peering into their crystal ball, wondering what that eventually means for their heavy reliance on that single country and top of the line products that make them their money. That's a very new worry that didn't exist when they were using Intel chips, which were manufactured in multiple countries. Apples future currently lies in whatever China ends up choosing to do with Taiwan. That's a major single point of failure and scary position to be in. If I were Apple, I'd want to ensure more control over my product line into the future and history has shown that is a very big priority for them, to their benefit. I'm quite sure the China/Taiwan situation is giving everyone at Apple quite a headache. It's hard to predict and Cook's steering a trillion dollar ship. If you have the money, hedge your bets.
It would also open up an additional major revenue stream, after they've perfected it for their own products.
> Apple had 0 chip designers working for them not so long ago.
Apple's acquisition of PA Semi was over a dozen years ago.
I agree that the second best time to plant a tree is today, but I also think that nothing today is indicative of Apple's likelihood of success at operating a fab 5-10 years down the line from today.
They tried cars and (so far) seem to have failed. I have nothing to suggest that their odds at succeeding in chip manufacture would be anything other than 50/50.
I'd hope that money could be put towards researching ways to make their machines less fragile, or how to engineer an M.2 slot in a professional device with more than enough room. Maybe a thinner camera?
Infinite money doesn’t translate into infinite R&D, unfortunately. As pointed out by books like The Mythical Man Month, you can’t necessarily speed up development by adding more engineers/resources.
Put another way, infinite R&D will produce diminishing returns very quickly in the same time frame with the same constraints. R&D is a very long game to play, and depends on anticipating and adapting to more favorable constraints.
Edit: that said! The benefit of investing in it anyway is potentially changing the course or severity if those constraints by having more control over them.