You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
Intel delays $28 billion Ohio chip factory in New Albany again, to 2030 or 2031
Feb 28 (Reuters) - Intel's promised $28 billion chip fabrication plants in Ohio are facing further delays, with the first factory in New Albany expected to not be completed until 2030, local media outlet The Columbus Dispatch reported on Friday.
The first factory will begin operations sometime shortly thereafter in either 2030 or 2031, the report said, citing the chipmaker.
Shares of the company, which originally scheduled to begin chipmaking in Ohio factories in 2025, were up more than 5%.
Maybe TSMC will buyout the Ohio fabs? Assuming it comes with some CHIPS Act money.... That I could understand. TSMC subsidizing Intel to make fabs in Ohio I would not understand.
Intel has built empty fab shells and filled them later, this is not new. I wonder what financial strings were attached to this project? Is Intel going to get a call from Trump/Musk on Government waste?
Intel has built empty fab shells and filled them later, this is not new. I wonder what financial strings were attached to this project? Is Intel going to get a call from Trump/Musk on Government waste?
I am still eobdering where is the new ceo ? Is there actually a regularatory deadline for this? Or Intel are allowed to operate in this manner indefinitely?. Maybe Frank Yeary wanted to be in control all along.
I am still eobdering where is the new ceo ? Is there actually a regularatory deadline for this? Or Intel are allowed to operate in this manner indefinitely?. Maybe Frank Yeary wanted to be in control all along.
Haha. He knows he can’t qualify to be the CEO. That’s why he wanted this situation to last for as long as possible. So he can be the de facto leader via being executive chairman
Intel has built empty fab shells and filled them later, this is not new. I wonder what financial strings were attached to this project? Is Intel going to get a call from Trump/Musk on Government waste?
no this is completely new. Intel announced a site that was not required by any planning data. They announced ...for some reason... 6 fabs. only one of which will be build. And this was done at a new site rather than existing site. last time this was done was Ft worth..... you can google what happened there LOL.
There is a reason why this was announced after 2 months of discussions on options with other companies.
I am still eobdering where is the new ceo ? Is there actually a regularatory deadline for this? Or Intel are allowed to operate in this manner indefinitely?. Maybe Frank Yeary wanted to be in control all along.
Don't you have to know if there is still a company before hiring a CEO? Interim with the board managing the deals is the correct org. Once the new structure is announced they can hire a CEO (just an opinion).
Intel has built empty fab shells and filled them later, this is not new. I wonder what financial strings were attached to this project? Is Intel going to get a call from Trump/Musk on Government waste?
"The changes were made so Intel can align its factory operation with market demand and "manage capital responsibly", Naga Chandrasekaran, general manager of Intel Foundry Manufacturing, said in a message to employees which was shared in a press release.
"We are taking a prudent approach to ensure we complete the project in a financially responsible manner.""
The subsidy is a small portion of Intel's whole investment. What Intel didn't say is that there is a lack of demand for its products and foundry services to justify the Ohio fab project. It's a financial problem and a fundamental competitiveness problem for Intel.
Maybe TSMC will buyout the Ohio fabs? Assuming it comes with some CHIPS Act money.... That I could understand. TSMC subsidizing Intel to make fabs in Ohio I would not understand.
The subsidy is a small portion of Intel's whole investment. What Intel didn't say is that there is a lack of demand for its products and foundry services to justify the Ohio fab project. It's a financial problem and a fundamental competitiveness problem for Intel.
Was there ever sufficient demand (internal and/or foundry commitments) to justify building a new fab site in Ohio?
Intel's recent woes have never been about not having enough fabs, and it seems unlikely foundry customers would fault Intel for lack of capacity as a deciding factor.
no this is completely new. Intel announced a site that was not required by any planning data. They announced ...for some reason... 6 fabs. only one of which will be build. And this was done at a new site rather than existing site. last time this was done was Ft worth..... you can google what happened there LOL.
There is a reason why this was announced after 2 months of discussions on options with other companies.
I believe they were trying to get a solution where someone (maybe TSMC) would buy it or JV it. It didnt happen. Even a shell and site is not worth much (TSMC already has a shell and a site). there are still stories of negotiations with GF, UMC, TSMC and partnerships with customers (Nvidia, Microsoft, Broadcom) but the hard part is making the numbers work and "what is in it for the partner"....
No one really wants or needs the Ohio Fabs. Certainly not Intel. I would list chance of Intel ramping Ohio ever as <10%. I hope I am wrong because I want more Fabs in the US and I want a TSMC Alternative.
Yes back when Intel products had a larger wafer demand than Apple pre 2022. 15 fabs of 10/14/22nm and Intel still couldn't keep up with demand even outsourcing chipsets to Samsung for 12th gen.
Partially it is. Outsourcing started because Intel's factory network was too small to handle all of Intel's needs. After all Samsung 14 is not better than Intel 14. And N6 isn't better than Intel 7. Losing the trailing edge business would have a large impact on Intel foundry during this period where leading edge runs on TSMC.
This couldn't be further from the truth. If there isn't garuenteed capacity you got no contact. Everything else is irrelevant if there is insufficient capacity for their needs.
Yes back when Intel products had a larger wafer demand than Apple pre 2022. 15 fabs of 10/14/22nm and Intel still couldn't keep up with demand even outsourcing chipsets to Samsung for 12th gen.
Thought Ohio was announced later, but was thinking about this in the context of the 2 Madgeburg fabs and new Arizona fabs fab expansions. A greenfield site is a major investment if it was just a matter of demand not keeping up.
Partially it is. Outsourcing started because Intel's factory network was too small to handle all of Intel's needs. After all Samsung 14 is not better than Intel 14. And N6 isn't better than Intel 7. Losing the trailing edge business would have a large impact on Intel foundry during this period where leading edge runs on TSMC.
Do you consider too much demand to be a "woe" though? I guess 2019 (for Samsung) is recent, but for even more recent outsourcing, you could argue process efficiency was a bigger factor than fab capacity.
This couldn't be further from the truth. If there isn't garuenteed capacity you got no contact. Everything else is irrelevant if there is insufficient capacity for their needs.
Thought Ohio was announced later, but was thinking about this in the context of the 2 Madgeburg fabs and new Arizona fabs fab expansions. A greenfield site is a major investment if it was just a matter of demand not keeping up.
My point was that over 15 fabs were needed just for Intel with simpler purposes technologies back in 2022. From 2022-2030 Intel announced plans to fully ramp 9 fabs. Granted these fabs were bigger than the average across Intel's pre EUV fabs. But they need to be bigger to support bigger tools and more complex process technologies. And all of that was spread across Intel 4/3, 20/18A, and 14A. Even with minimal external business Internal demand should have had no trouble if Intel chip demand was growing even at a snails pace. Of course Internal unit shipments went to like half of what they were before the downturn and there is little indication that pandemic era volumes are ever coming back even at the end of the decade. This is THE problem for both sides of the business.
And as far as I know Intel's AZ/IS/NM/IR sites are all at the maximum size or pretty close to what the governments will let them build. New greenfields are unavoidable if Intel wants more than 6 EUV capable fabs outside of Oregon (which even operating with TSMC productivity on a process with as few mask layers as N5 is only as big as ONE contemporary TSMC gigafab for ONE process node).
FWIW when Intel showed off a capacity roadmap before fab push outs. By 2030 they thought Intel 4-18A wafer starts from all those fabs would roughly equal their 10/14 capacity in 2022. That hardly looks like an aggressive build out to me.
It wasn't a problem when fabs were full. But now they aren't because the CPU sales can't fill Intel's fabs single handily like they could from 2018-2021 and they chose to move all of the supporting Si to external foundries specifically to make room for said underselling CPUs. Intel took huge write downs on Intel 7 equipment last year because there is no volume tail (chipsets/PCH/PCD/etc.) to ever fill those fabs with internal Si. Full fabs make money empty ones bleed it. Now given the poor Intel 7 margins, I doubt Intel foundry would be profitable with all that 14LPP business on i14 and the N6 stuff on i7. But the losses would be less and large asset write-offs wouldn't have been needed. But those billions in extra revenue and extra net income would make the catch up R&D and capex more palatable.
You say that as if Intel fabs run their tools at half the throughput of everyone else. Anyone who thinks that doesn't know what they are talking about. If that was the case Intel would need WAY more fabs to run the volume they did.
And what failed foundry deals would those be? There is no concrete information on who/what products Intel failed to win. And even if they won something unrealistic like 200 customers you wouldn't be seeing a dime of that revenue for YEARS.
But to answer your question if there were no wafers their would be no negotiations. If 18A was better than TSMC 5A at a lower cost not one company would sign a deal without volume commitments from Intel and a believe that Intel would deliver their wafers when they needed them. Having wafers is a prerequisite to even getting a seat at the table.
It still could be worse. EU is weighing putting 25% tariffs on all US-manufactured chips, to reciprocate the US gov robber-banker tactics and hit us where it really hurts. This would put Intel to its knees.
. Maybe Frank Yeary wanted to be in control all along. 
