Chipmaking depends
hugely on economies of scale. TSMC was relying on 2023 orders from both Apple
and Intel to pay for their planned level of 3nm manufacturing capacity. Car analogy: if you have one large and one smaller customer wanting limo rides, maybe you can afford to buy one new limo and cater for both of them. If the bigger one of those customers drops out, you can't buy 1/3 of a new Limo. Limos don't scale smoothly. Nor, I imagine, do 3nm chip production lines.
As
@MayaUser pointed out - with 20:20 hindsight
maybe Apple could have offered TSMC enough business to pay for their entire planned capacity - e.g. by moving more of the Apple Silicon range to 3nm sooner - but these things are planned long in advance and it sounds like the plan was to start with the relatively small volume M2 Pro
because of the contention for manufacturing capacity. Apple can't turn on a dime and have the entire 3nm Apple Silicon range ready to roll at such short notice.
I think that's one for
Hanlon's Razor. More likely, it's just Intel being short of cash:
'We must and will do better' CEO pledges amid big losses, Optane axed, expectations slashed
www.theregister.com