siliconbruh999
Well-known member
Ireland will be safe lol buy yes in the end TSMC will pass tariffs to customers
Intel delays $28 billion Ohio chip factory
- Thread starter MKWVentures
- Start date
I believe there was something like $600m from Ohio to start by 2028 and like $1-2B of the CHIPS grant contingent on starting this. But if you don't need the fab, it's probably not worth spending another $20 billion before you need it even if it means giving up these subsidies. And of course it could be a negotiating tactic with the current admin.
Arizona Fab 52 is just being tooled up now and 62 is still empty. Intel has plenty of capacity for the next few years and the 2028 Ohio project was probably over-ambitious/politically-motivated target.
hist78
Well-known member
"Arizona Fab 52 is just being tooled up now and 62 is still empty. Intel has plenty of capacity for the next few years and the 2028 Ohio project was probably over-ambitious/politically-motivated target."
Former Intel CEO Pat Gelsinger might have had another intention in mind. He believed he could use the empty shells, Ohio Fab and Arizona Fab 62, to secure more federal and state subsidies to help resolve Intel's financial problems without actually buying and installing the expensive fab equipment.
Fortunately or unfortunately (depending on your perspective), the Biden administration insisted on performance and milestone based funding delivery. An empty shell was not sufficient to meet the purpose of the CHIPS Act legislation.
Maybe, but subsidies would only likely partially pay for fabs, not the whole thing, and likely wouldn't be able to "solve Intel's financial problems." It just mitigates the cost of future growth. I think Pat did want to get as many subsidies as he could: remember, these things were announced before the CHIPS act was even passed so he was probably campaigning for the legislation, but I also think Pat thought he would land more external customers quicker than he did. Now Intel has to recalibrate to fewer external customers. That's just my impression, with all the language around ROIC/ focusing on internal products more (but still keeping the same strategy of trying to land external customers) around the time of Pat's firing.
If Ohio was initially intended for 14A HVM, delaying the schedule from 2025 to 2030 suggests that Intel may struggle to secure major clients for 14A. While Intel might produce enough 14A for internal needs, it would be challenging to cater to large external customers and secure substantial contracts.
There is no news about external customers for Intel 3. Therefore, Pat's best strategy might be to establish a strong reputation with 18A and secure significant orders for 14A through IFS.
However, postponing Ohio's 14A fab to 2030 (likely 2031 for HVM) almost guarantees that Intel will miss the opportunity. TSMC is set to begin N2 mass production in the second half of this year. If they maintain their three-year node cadence, TSMC will have 14A HVM by the second half of 2028. Who will need Intel's 14A in 2031?
While Intel can still claim to be in the foundry business, obtaining orders from non-government entities and making reasonable profits seems nearly impossible.
There is no news about external customers for Intel 3. Therefore, Pat's best strategy might be to establish a strong reputation with 18A and secure significant orders for 14A through IFS.
However, postponing Ohio's 14A fab to 2030 (likely 2031 for HVM) almost guarantees that Intel will miss the opportunity. TSMC is set to begin N2 mass production in the second half of this year. If they maintain their three-year node cadence, TSMC will have 14A HVM by the second half of 2028. Who will need Intel's 14A in 2031?
While Intel can still claim to be in the foundry business, obtaining orders from non-government entities and making reasonable profits seems nearly impossible.
I don’t know what you’re talking about no news on Intel3. Amazon announced last year. A recent article says volumes could be fairly substantial:
“Ronak Singhal, senior fellow and chief architect of Xeon products, said the Xeon 6900E is designed to meet the custom chip needs of cloud customers, including a large one that has it deployed now... “When they’re ready to talk about what they’re doing there, I think it’ll be pretty interesting,” he said. https://www.crn.com/news/components...s-it-won-t-broadly-deploy-288-core-xeon-6900e
The Xeon 6900E is an Intel product, meaning Intel isn't actually performing a "foundry" role with it.
Volume is volume. I think that is splitting hairs. I also wouldn’t rule out MSFT cobalt but MSFT has been very secretive about its custom silicon plans. I’m also not sure Intel was going hard after customers for Intel 3, I think 18A was supposed to be the one that opened up to external customers in a significant way.
MKWVentures
Moderator
it is based on fab or stamped country of origin? where does Intel package parts. Fab is not country of origin.
siliconbruh999
Well-known member
well intel can fab at TSMC and put the packing stamp at China/Malasyia/New Mexico
MKWVentures
Moderator
correct. Every tariff comment is useless unless we define Country of origin in the tariff. It took cars year and years to work out details. % of value, % of parts, locations of initial transformation, Location of final transformation. once we know the details, people will know how to adjust to (and circumvent) the tariffs.
"A large one" means a large cloud player, but it is not implying that this player deployed a large volume of Xeon cpus.
You reckon Trumps tariff team have got all this sorted out for todays activities?
Or the tariff is based on the last place the product was before it enter US?
MKWVentures
Moderator
they do not have details.
Tariffs based on where it is shipped from is easy to implement. PCs, Phones, servers are made in other countries. TSMC wafers are not shipped to US. Tariff problem solved.
Tariffs are always bad.
they do not have details.
Tariffs based on where it is shipped from is easy to implement. PCs, Phones, servers are made in other countries. TSMC wafers are not shipped to US. Tariff problem solved.
Tariffs are always bad.
Historically proven to only provoke negativity.
Maybe Donny and his crew know something the greatest economists in history do not