The “it probably can’t get much worse” stance seems to assume there’s real upside, but I’m not so sure that’s achievable. Foundry is simply not competitive today, and it was supposed to be by now. Instead, it needs more time, more money, and more talent, while generating little to no profit in the meantime.
Products is kinda the same with their CPUs/GPUs not being competitive. Falcon Shores was scrapped and reduced to a test chip. Their CPU uarchs lag behind, and now they’re trying to merge the P-core and E-core teams (surely a smooth, frictionless process, right?)
They’ve also had to rely on external nodes to stay remotely competitive, which hits margins hard. Combine that with underwhelming uarch on both fronts, and again: it’s more time, money, and talent needed.
Meanwhile, they’re losing experienced engineers to Apple, AMD, and Nvidia. So how do they attract the talent needed to turn things around? And how do they fund and execute on a competitive architecture from here?
If you compare that with AMD’s situation when they bet everything on Zen (from-scratch, high-performance uarch) some rumors say Intel killed off its own high-perf design last year and the team behind it left the company.
1. the real economic value of the equipment in the kitchen is well below its book value - you got to apply a steep discount to book value of equipment that can't make chips (N3, 18A)
2. the return on Book Value, that is ROE, is well below the cost of capital - you won't buy "the book" at book value because it gives you negative returns
It can also mean there’s a big pile of acquired goodwill from foolish acquisitions that generate zero return. The biggest component of Intels book value are the empty shells in Ohio and Arizona waiting to be filled with equipment to stand idle because they have no customers.
In the same boat as you sold all TSMC on liberation day. Looking to buy back in. Noon value in seminal is not valid now way Intel fabs are worth book value. If Intel gets a viable node 2x if not 0.5 more valuable
You’re the best teacher at making fools understand; I learned a lot from this.
The “it probably can’t get much worse” stance seems to assume there’s real upside, but I’m not so sure that’s achievable. Foundry is simply not competitive today, and it was supposed to be by now. Instead, it needs more time, more money, and more talent, while generating little to no profit in the meantime.
Products is kinda the same with their CPUs/GPUs not being competitive. Falcon Shores was scrapped and reduced to a test chip. Their CPU uarchs lag behind, and now they’re trying to merge the P-core and E-core teams (surely a smooth, frictionless process, right?)
They’ve also had to rely on external nodes to stay remotely competitive, which hits margins hard. Combine that with underwhelming uarch on both fronts, and again: it’s more time, money, and talent needed.
Meanwhile, they’re losing experienced engineers to Apple, AMD, and Nvidia. So how do they attract the talent needed to turn things around? And how do they fund and execute on a competitive architecture from here?
If you compare that with AMD’s situation when they bet everything on Zen (from-scratch, high-performance uarch) some rumors say Intel killed off its own high-perf design last year and the team behind it left the company.
P/Book <1 can mean other things:
1. the real economic value of the equipment in the kitchen is well below its book value - you got to apply a steep discount to book value of equipment that can't make chips (N3, 18A)
2. the return on Book Value, that is ROE, is well below the cost of capital - you won't buy "the book" at book value because it gives you negative returns
It can also mean there’s a big pile of acquired goodwill from foolish acquisitions that generate zero return. The biggest component of Intels book value are the empty shells in Ohio and Arizona waiting to be filled with equipment to stand idle because they have no customers.
For TSM, is there a price level you’re waiting for?
For TSM, is there a price level you’re waiting for?
Hey let's make 21,000 layoffs to make Intel great again.
Can Lip fix Intel? The problems run deep, lits legacy and bad culture.
In the same boat as you sold all TSMC on liberation day. Looking to buy back in. Noon value in seminal is not valid now way Intel fabs are worth book value. If Intel gets a viable node 2x if not 0.5 more valuable