2. TSMC is much larger than Apple's manufacturing requirements. To keep it running, they would have to commit to doing a lot of contract work for other companies, which really isn't Apple's style. Moreover, many of TSMC's current clients (like AMD, Intel, Nvidia, and Qualcomm) might be uncomfortable with Apple controlling the company, as that would introduce conflicts of interest.
3. Owning TSMC would put Apple in charge of the company's technology development, and would expose them to significant financial risk if TSMC suffered a failure, much like Intel has been harmed by their failures to develop a 7nm process. Contracting out their semiconductor manufacturing allows Apple to switch foundries if needed.
I also have to imagine that if people are making a big antitrust stink to the government about how Apple takes 30% of Fortnite loot boxes, there will be an equally big problem when AMD, Nvidia, Broadcom, Qualcomm, Marvell, etc. can't have chips produced anymore.
On the other hand, Samsung owns a fab and makes phones, so it's not unheard of. Samsung has other big semiconductor businesses, though, and it's not clear that Apple really wants to enter into those. You don't need to manufacture your own flash memory to make good laptops; Samsung's SSD business does well on its own.
I really like vertical integration in general, but sometimes things work great even when they're completely decoupled. Two small projects is a lot easier to run than one big project, so you have to show some benefit for integrating things. In Apple's case, it was clear that Intel's CPUs weren't getting them where they wanted to go with thin laptops, so they made their own. They had to in order to sell competitive laptops. This whole thing is in its early days, though, and we don't even know if they're going to beat Intel or AMD. So the summary is, being TSMC's customer is working out great for Apple; they specify a CPU and they get a good one at a cost that the market seems OK with. They want to make a big bet and dramatically change the computing landscape, but that might not work out. What if Intel totally leapfrogs Apple Silicon? It is so uncertain that it is hard to say "they should just buy TSMC". Rather, they should sit and wait and see how they do, and then decide if more integration will let them make better phones, tablets, and laptops.
Fun fact - they account for 5% of Taiwan’s energy usage alone. Wild.
It is widely-licensed but also vulnerable to being superceded or abandoned by licensees if it tries to squeeze them for more money.
It exists because it provides a useful service, but not a critical one
Ahem. ARM.
Seriously, though. Take a look even at TikTok. I was pleasantly surprised that it's not for sale, either. Why does every tech company has to be HQ'ed in the US?!
After all, TSMC helps make Taiwan one of America's valued trading partners, even if America doesn't officially acknowledge Taiwanese sovereignty.
So long as a Chinese invasion will mean American companies lose access to TSMC and Chinese companies gain it, they hope this provides them the same protection Kuwait and Saudi Arabia enjoy. Or at least, if not certain protection then a high enough chance of protection to discourage an invasion.
As the UK does not have an expansionist neighbour disputing their sovereignty, ARM is not a strategic asset in the same way.
Apple also isn't the type to manufacture chips for other companies which is a lot of TSMC's revenue (this article is only for one size). So they would be taking a huge loss in value just to have control.
Much more likely for Apple is to try and create their own manufacturing lines.
The Taiwanese notably are framing the Huawei issue as a ban on a specific company rather than trade with China in general [1]. Whether that continues to be true might hinge on November...
1. https://www.reuters.com/article/us-taiwan-economy/taiwan-min...
Purchasing TSMC would lead to it having to deal with maintaining relationships to people it actively competes on several levels with. This would almost certainly lead to accusations of anti-competitive behaviour.
Buying ASML also wouldn't really gain Apple anything. There's tons of technology and know-how that goes into a fab beyond buying a lithography machines.
#249: A Sane Amount of Cash http://5by5.tv/criticalpath/249
I barely understand what Horace is talking about (square hole, round peg) and would only butcher his thesis. Sorry.
These podcasts are in the "think out loud" format, vs finally crafted presentations. So I listen at high speed and rewind to replay the key points.
Edit: Gonna risk it.
Large M&As are notoriously bad moves. Beats was Apple's largest acquisition, which Horace thinks was ok because it probably broke even. (I have a different take: it was an acqui-hire and they got some head phones as bonus.)
Reinvesting cash surplus is also bad, because it's shareholder's cash for them to invest themselves. Also, to do so would make Apple a hedge fund, which is a huge distraction.
Horace doesn't touch on Apple style partnerships, the deep investments into their suppliers. Not quite keiretsu. Not quite monopsony. Definitely not WalMart style partnerships, which is very abusive. Maybe more like defense or auto manufacturer partnerships, where it's more mutually beneficial. Like Toyota does. I would love to read case studies about Apple's partnerships.
Taiwan would do the same in the case of TSMC.
1. TSMC has very low margins. Buying the company means lowering Apple margins, that are the highest on the world.
2. Lots of companies depend on TSMC, if Apple did that they will be sued for monopolistic behavior.
3. Apple benefits enormously from the cheap devices that TSMC makes for other companies, that is volume, improvement of technology that Apple can use(ARM). It is over 5x the volume in sales of Apple products.
4. It is a strategic interest company of Taiwan,Europe, China and US.Buying it is not economics,it is geopolitics.
Also, tsmc has loads of clients. Apple doesn’t want to sell to them. So they shut them down?? It’s a mess.