If 18A does not ramp externally, you cannot say "lets try again on 14A". the stock price will drop ever further below book. We are just talking about foundry.... Intel foundry will lose billions more in 2026. What does Fitch say that foundry revenue and expenses are in 2026?
I thought the lesson from TSMC and other foundries is that with the right amount of process adaptation, IP adaptation and new channels (like UMC and perhaps GlobalFoundries), fully depreciated fabs could continue to produce revenue at good margins.
They said break-even in 2027.
Guess I didn't make this clear. I didn't say foundry could skip 18A externally, I said get a few customers with small volume to kick the tires. They do have some pre-pays in the bag and the DOE purchases so it isn't like they will have nothing to show for it. I believe that has been the whole problem with the perception that Gelsinger created, despite the expectations he set, no one was ever going to just chuck TSMC and go all in on 18A. But compliments of Gelsinger everyone thinks Intel's foundry efforts have failed if they don't get 10's of billions of dollars out the gate with 18A.
I believe "if you build IT they will come" is true. The catch is in identifying what IT is. I believe IT is competitive technology with comparable ease of adoption and an established history of delivering what you promise when you promise. It is that last one that is going to ensure adoption of Intel's foundry is slow even if they deliver on the other two vectors. That said, Intel has publicly said they don't need large volumes in the near term to make foundry profitable. You're welcome to believe they are lying (or just overly optimistic) if you would like.
what is your estimate for external 18a revenue in 2026 and 2027?
They said it is a function of EUV usage. By 2027, I assume the majority of Intel products use EUV.Has Samsung Foundry broke even in the past 5 years?
14A really needs to win some designs and that means it needs to deliver PDK 1.0 well before TSMC A16. For Intel Foundry to break even 14A needs to be in HVM? 2027 seems a bit early. I say set the break even goal post at 2030 and beat it.
I would look at the recent earnings report to get an idea of Margins by process type. It would be good to ask what the foundry margins will be on 18A in 2026/7.
I think if they can get the loss within +-2b, that should be acceptable. In that case, the fab side would not drag down the valuation of Intel. The foundry can also have a positive impact on the valuation.
Couple items:
I also heard that (point 1). AMD gets high PE partly due to data center AI. In 2027, Intel should have its own data center AI solution. Also Intel's data center CPU is closing the gap with AMD.
David also mentioned that part of Intels problem was that the previous CEO was getting filtered information from managers and Engineers and that it was too optimistic (ie not CEOs fault).
That quote form David is regarding the Panther Lake\18A delay i.e. Pat saying Panther Lake will launch in 2H'25 to now only 1 SKU launching in end of this year.
I guess I might have some update about financial targets later today.
I think David is just saying that Intel 18A is a node process developed with a focus on being used by the likes of Intel Products in mind i.e. PC & HPC. Whereas Intel 14A is being developed with early inputs from potential foundry customers.
Previously, the Head of Investment Relations mentioned that they are targeting a level within the 15–25% range. I would think it's at the lower end of the range. This aligns with Lip-Bu’s view of utilizing more of their own fabs. Additionally, IF views internal and external allocation as equally important. Having a portion outsourced would give IF more flexibility.
