It is still posted in their website. Are they still lying?
Exclusive: Intel's new CEO explores big shift in chip manufacturing business
- Thread starter XYang2023
- Start date
What a ridiculous title. Intel Foundry never had a shot at Apple business.
Is this the same person? Or did AI steal an identity?
https://www.linkedin.com/in/naumanhk/
Intel (INTC) May Kill 18A in Billion-Dollar Bid for Apple, Nvidia
Nauman khan
Wed, July 2, 2025 at 4:58 AM PDT
Yahoo is using AI to generate takeaways from this article. This means the info may not always match what's in the article. Reporting mistakes helps us improve the experience.
Key Takeaways
Is this the same person? Or did AI steal an identity?
https://www.linkedin.com/in/naumanhk/
Intel (INTC) May Kill 18A in Billion-Dollar Bid for Apple, Nvidia
Nauman khan
Wed, July 2, 2025 at 4:58 AM PDT
Yahoo is using AI to generate takeaways from this article. This means the info may not always match what's in the article. Reporting mistakes helps us improve the experience.
Key Takeaways
- Intel CEO Lip-Bu Tan is considering ending marketing of its 18A process to focus on 14A technology due to struggles in winning new external customers.
- The potential shift could lead to a one-time write-off of several hundred million to potentially over a billion dollars, with the board expected to decide on the matter at its fall meeting.
- Intel will continue to honor 18A commitments to Amazon and Microsoft, while investors will be monitoring whether a pivot to 14A can deliver the performance gains and customer wins Tan is aiming for.
siliconbruh999
Well-known member
Also how does 18A is getting killed for Nvidia and Apple this has to pass through AI multiple times
MKWVentures
Moderator
You are correct.... they did announce customers so i may have exaggerated with "no customers" .... my badwell Intel got 2 customers(AWS/Microsoft) ON 18ayou can say the volume for those customer is low but saying no customers is kind of misleading
but even then, lets talk IF the production stepping that is used in servers for those two customers tapes out on 18A. LBT has things changing more quickly than I expected.
@dkr1986
You are correct.... they did announce customers so i may have exaggerated with "no customers" .... my bad
but even then, lets talk IF the production stepping that is used in servers for those two customers tapes out on 18A. LBT has things changing more quickly than I expected.
@dkr1986
I am not aware of any 18A external tape-outs thus far and that includes Amazon and Microsoft. Maybe we will hear more real about customers on the next investor call. With the staffing cuts that is going to be a big call.
Last edited:
For any fabless to gain confidence in IFS requires four things
1) Accurate and Stable PDK
2) high yield
3) schedule confidence
4) track record
Not sure why anyone thinks after last three generations Intel would deliver on 14A
Being on Lip-Bu’s personal Rolodex won’t help Intel fix these
1) Accurate and Stable PDK
2) high yield
3) schedule confidence
4) track record
Not sure why anyone thinks after last three generations Intel would deliver on 14A
Being on Lip-Bu’s personal Rolodex won’t help Intel fix these
No need to waste time on those rumours.
www.intc.com
Q2 2025 Intel Corporation Earnings Conference Call
www.intc.com
Last edited:
That’s almost a month out from today.No need to waste time on those rumours.
Q2 2025 Intel Corporation Earnings Conference Call
20 days, 2/3
MKWVentures
Moderator
I am not sure Intel will give much of an update until any deals are finalized. My opinion/understanding is:
1) external foundry has minimal volume in the next 3 years based on the past 4 years of marketing and business development. Any 18A wins going forward wont ramp until 2029.
2) Intel most likely has locked in some DOD items. These are low volume and will take forever (like 2030+). but they are real.
3) It is not clear how intel foundry is financially viable if they have to wait and organically grow over the next 10 years. The maths aint mathing and there are some serious JV requirements
4) Intel internal volume cannot pay for development and fab build costs, volume is too low. This has been a problem for past 10 years.
5) LBT is looking at solutions.... not "just wait it will be great next time" or "it will take 10 years to build a foundry business"
- Get JV, partnership, with a few major foundry customers... give them a reason to use Intel and not be direct competition with TSMC
- Get JV, Partnership with other foundries. UMC has always made sense, GF makes sense, even Samsung makes sense.
- Go back to the pre-Pat plan. Outsource, stop process development, get out of IFS.
the Challenge with all of these is what @BruceA said. Intel has to deliver on 18A, and fixing ifs finances, or no one will want to partner with them or buy from them.
I think we will see some major changes announced across Intel over the next 9 months to try to fix the balance sheet.
I think people will definitely ask the question during the earning call. For Intel, they have to find one or two large customers to make 14A work.
3 weeks. Looking forward to it. Personally I think this call will be pivotal for Intel.
18A is not a flop. The 18A PDK flopped for external use, different thing. From what I am told 18A yield is good.
If 14A does not yield Intel will be fabless. Hopefully Lip-BU put a hold on HNA-EUV to reduce risk.
I still say Intel could do deals with Samsung, GlobalFoundries, etc... like they did with UMC. Intel could be the foundry's foundry to fill up their fabs.
MKWVentures
Moderator
Intel planned on about 3x the number of wafers for 18A that they will actually run (assuming 18A continues). so that is a flop. Volume is what pays for TD and new fabs .... that was the Intel problem statement before Pat . the financial numbers for 18A look really problematic now
18A was supposed to be ahead of N2 so how did TSMC get 10x more tapeouts lined up than Intel 6 months before either process is qualified?
"I still say Intel could do deals with Samsung, GlobalFoundries, etc... like they did with UMC. Intel could be the foundry's foundry to fill up their fabs." - 1000% agree
I don't know how IFS can survive waiting to get to 14A. I would love to see the spreadsheet that shows this is possible. I think LBT has other plans that have not been announced yet.
Becoming fabless and supplying legacy nodes are both alright. It is good that Lip-Bu is intercepting further capital spending. If it was under PG, I don't know where he would lead Intel to.
Interesting tidbit from the story (source is "industry analysts"):
"However, according to some industry analysts, the 18A process is roughly equivalent to TSMC's so-called N3 manufacturing technology, which went into high-volume production in late 2022."
18A used Intel restrictive design rules to achieve parity with TSMC (and boost yields). Once you relax the design rules to a more industry standard, the performance advantage disappears. This means there is a new performance gap to overcome: The de-restriction gap between how Intel designs their x86 custom parts and how fabless houses design for a standard PDK.
My fear is that a slimmed down Intel won't have the resources to choose between custom and standard PDK. If they go all-in on standard PDK, they lose yield and performance on x86, with no assurance they will gain external customers. It must be agonizing to leave yield and performance on the table.
"However, according to some industry analysts, the 18A process is roughly equivalent to TSMC's so-called N3 manufacturing technology, which went into high-volume production in late 2022."
18A used Intel restrictive design rules to achieve parity with TSMC (and boost yields). Once you relax the design rules to a more industry standard, the performance advantage disappears. This means there is a new performance gap to overcome: The de-restriction gap between how Intel designs their x86 custom parts and how fabless houses design for a standard PDK.
My fear is that a slimmed down Intel won't have the resources to choose between custom and standard PDK. If they go all-in on standard PDK, they lose yield and performance on x86, with no assurance they will gain external customers. It must be agonizing to leave yield and performance on the table.
I wonder about this. AMD is producing very competitive (many would say winning) x86 products using TSMC PDKs. So there certainly is proof it can be done. And Intel has products using TSMC so they know what it takes and what can be achieved.
It is very likely that Intel is presently facing the scary decision to go all in on industry standard PDK. They likely don't have enough staff to fully develop the industry standard PDKs (with associated tools support/vetting, etc) while still fully supporting their internal kits and flows. And the staffing levels are going down with the cuts. (Though maybe they will reassign some smart folks from other groups.) If LBT can get Intel through this massive sea-change decision, history may consider him successful at Intel even if it comes too late to completely save the firm.
Exciting times indeed, in every sense of the word.