Intel Corp (INTC)

[Unkwn]

TSMC does have a high margin core business, it's called wafer manufacturing for an extremely broad range of clients across a large spectrum of manufacturing technologies.

Yeah, but definitely not as profitable as Intel's core business. Sure, if you take the margin mix, then TSMC may not be far off, but that actually means that Intel has products at fairly low volume (compared to TSMC) that already pay for what TSMC needs much more volume (in chips/wafers) for. That's the point. Intel's margins for its core products improve when they fab chips for others. They leverage their core business this way, just like Samsung already does it. This, for sure, is going to harm TSMC. We had this discussion a few times already.

Anyway, as far as your argument goes about fixed costs, I'm not sure that it's totally correct.

Apple is a customer that adopts leading edge technology nodes as quickly as it can. So this means that if Apple were to build on Intel 10nm, it would be running wafers -- and lots of them -- right alongside Intel's own 10nm products.

This means that to support Apple, Intel would actually need to factor in Apple's needs into its own capacity/demand planning. This means more 10nm capacity built up in advance, and it would also mean that by the time Apple and Intel move on from 10nm for theiar highest volume products, even more 10nm capacity would be left behind (or converted to a future technology).

Yes, that's true. In order to supply that additional capacity, you need more product lines with lots of expensive equipment, expensive staff and eventually, a new plant. The question is: How big is that part of the fix costs? Frankly, I don't know for sure, but I expect (!) it to be a lot less than the actual fixed costs associated to actually developing that process and make it work. If you look at the Capex figures and how they change when new processes are developed, I think this is a reasonable assumption. You are right, fixed costs increase when you have to add capacity, but I expect it to be not anywhere close to what you have to spend for the initial ramp up (which itself is part of the fixed costs).

As I stated before already, Intel and those other semicos, activate their huge R&D and ramp up costs and depreciate it over time, which simply means higher costs of sales which again reduces margins. More sales -better margins (for that part of the costs, not all).

Regarding Apple in particular: I can see one very interesting business case for Apple as well as for Intel: Intel gets the foundry deal, Apple gets Intel's modem IP to integrate into their own SoC. That's a business case that makes a lot of sense and would improve Apple's margins as well as help Intel claiming decent margins on those AP's, while getting some of their investments back from the modem development.

I see Apple also as a likely candidate to buy the modem business from Intel. I wouldn't like that to happen but it would allow Intel to exit this disaster gracefully. This wouldn't make me hold my Intel shares though. I would consider it capitulation. But that may just be me.