Risk/reward? What if TSMC invests billions into the next process and never gets good yields? It's happened to other fab owners. (Remember when Intel had comp

Risk/reward? What if TSMC invests billions into the next process and never gets good yields? It's happened to other fab owners. (Remember when Intel had competitive products?) Then Apple is out billions of dollars for nothing. But if they are just TSMC's customer, then they don't take on that risk. Their risk is that TSMC raises prices, or other people can make CPUs as good as theirs. But that is not a huge risk for Apple; if they can't buy capacity, neither can their competitors. 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.