News Intel's next CEO could be Lip-Bu Tan, a former member of the board who reportedly clashed with Gelsinger about strategy

[phead128]

Are you sure that this was actually written in his contract and not just a mere marketing statement? The thing is, that a CEO absolutely cannot even remotely make such promises, considering Intel's more than half a decade long struggles with its 10 nm node.

The first phase of IDM 2.0 strategy where Intel regains technological leadership from TSMC requires five process nodes in 4 years (5N4Y). It's part of the official process node roadmap released by Pat Gelsinger right after he became CEO, and it's key to turning IFS Foundry into a competitive foundry business.

https://www.intel.com/content/www/u...mbraces-internal-foundry-model.html#gs.ieod2t

We remain on track to deliver five process nodes in four years, and we’ve invested in the capacity required to meet the industry’s demand for semiconductors, bringing much-needed balance to the global supply chain.
-Pat Gelsinger

Also on the official process node website: https://www.intel.com/content/www/us/en/foundry/process.html

Beyond Five Nodes in Four Years (5N4Y)​

With execution to our ambitious five nodes in four years (5N4Y) goal on track, Intel revealed at Intel Foundry Direct Connect '24 an extended process technology roadmap adding:

• Intel 14A to the company’s leading-edge node plan,
• Several specialized node evolutions for Intel 3, Intel 18A, and Intel 14A, including Intel 3-PT with through-silicon vias for 3D advanced packaging designs, and
• Mature process nodes, including new 12-nanometer nodes expected through joint development with UMC.

Customers ready to design can start their engagement with Intel Foundry today.

 

[JarredWaltonGPU]

Pat Gelsinger claimed a defect density of 0.4 (source), which would translate to >70% yield depending on chiplet size and yield curves towards HVM. The problem is Intel is stuck at 10% with HVM in 2-3 quarters, it's not viable for HVM, and no progress made.

I have serious doubts about the "10% yield" on Intel 18A being a reliable piece of information. That's very inside information, the most inside of inside information. That sort of stuff is highly protected. And those claims, by a Korean company (i.e. potential ties to Samsung) could very much result in a "hit piece."

Everyone is reporting on the 10% yields of Intel 18A as though it's a fact now. If it's actually true? Yes, that would be bad. If it's not true? It's equally bad, because the stock takes a hit regardless. It would recover faster if it's not true, but then that would be the purpose of a hit piece.

Did Pat lie when he said things were progressing so well with 18A that doing 20A was deemed unnecessary? That's tough to say without inside information again. Certainly it would cost a billion or so to bring up 20A to the HVM level, even with all the prior work that went into it. If it was underwhelming, or would be short-lived, cancelling it wasn't necessarily a bad move. If he lied, that could be why he's gone — or let's say he "exaggerated" rather than "lied."

Ultimately, time will tell, but Intel stock is taking a big hit right now regardless. Pushing Pat out the door without a replacement and with no information on the why of the matter isn't going to impress investors.

 

[phead128]

I have serious doubts about the "10% yield" on Intel 18A being a reliable piece of information. That's very inside information, the most inside of inside information. That sort of stuff is highly protected. And those claims, by a Korean company (i.e. potential ties to Samsung) could very much result in a "hit piece."

If you don't believe Chosun (Korean newspaper), then perhaps you would consider New York Times (American newspaper) that have insider anonymous source that say it's even worse (less than 10% yield for 18A).

New York Times:
Some customers were recently informed by Intel that its most advanced manufacturing processes, which it calls 18a and 16a, were far behind TSMC, a chip industry official briefed on Intel’s progress said. TSMC is producing 30 percent of its leading-edge chips, known as 2 nanometer chips, without any flaws, while Intel’s new process produces less than 10 percent of its 18a chips without flaws, the person said.

So New York Times has insider sources saying it's even worse than 10%. This is consistent with Reuters/Broadcom assessment of 18A unviability back in September.

Everyone is reporting on the 10% yields of Intel 18A as though it's a fact now. If it's actually true? Yes, that would be bad. If it's not true? It's equally bad, because the stock takes a hit regardless. It would recover faster if it's not true, but then that would be the purpose of a hit piece.

I highly doubt Reuters and New York Times would try a hit piece. They are American newspapers, not Korean. What incentives do they have to hurt Intel?

Ultimately, time will tell, but Intel stock is taking a big hit right now regardless. Pushing Pat out the door without a replacement and with no information on the why of the matter isn't going to impress investors.

The interim Co-CEO of Intel today has admitted in the investor call that 18A has production issues, the production ramp has been delayed from H1 2025 to H2 2025.

Furthermore, the anticipated progress on Intel's 18A node technology seems to have hit a snag, with the production ramp now postponed from the first half of 2025 to the second half.
Dec 5th: Intel Investor presentation by interim Co-CEO

So this is entirely consistent with reports of Broadcom's 18A unviability for HVM soon and low yield issues with 18A. The interim Co-CEO just admitted this few hours ago, and consistent with firing Pat (despite overly optimistic 18A tape-out in 6 months, which turns out to be a lie)

 

[JarredWaltonGPU]

If you don't believe Chosun (Korean newspaper), then perhaps you would consider New York Times (American newspaper) that have insider anonymous source that say it's even worse (less than 10% yield for 18A).


So New York Times has insider sources saying it's even worse than 10%. This is consistent with Reuters/Broadcom assessment of 18A unviability back in September.

I highly doubt Reuters and New York Times would try a hit piece. They are American newspapers, not Korean. What incentives do they have to hurt Intel?

The interim Co-CEO of Intel today has admitted in the investor call that 18A has production issues, the production ramp has been delayed from H1 2025 to H2 2025.



So this is entirely consistent with reports of Broadcom's 18A unviability for HVM soon and low yield issues with 18A. The interim Co-CEO just admitted this few hours ago, and consistent with firing Pat (despite overly optimistic 18A tape-out in 6 months, which turns out to be a lie)

Ah, I missed the interim CEO statement, as I'm buried in GPU testing for next week.

What would be the purpose of a hit piece? Tank the stock, buy low, sell high. It has happened many times, including with Bloomberg and Reuters — and I'm sure there have been "leaked" information to the NYT and various other publications over time.

That doesn't mean this is a fake leak, merely that it could be (based on my complete lack of insider knowledge). It could be that Pat was pushed out for his statement that 18A was ahead of schedule when 20A was cancelled, because if this is correct, it was false when he said 18A was doing well. (Proving such things is always difficult, naturally. And Pat is probably fine with $12 million and retirement.)