> Barrett also said Washington must do its job to stay ahead of the semiconductor race. The U.S. has invested more in the semiconductor industry in the past year than in the last 28 years combined, but he says it should do even more, especially in academic research.
Oh. Okay. So why don't we split up Intel and then make this exact investment in the spun off divisions?
Why split Intel in the first place? Intel fabs die without the volume of Intel Design, and looking at the disaster that is the Arrow Lake launch, without the cost advantage Intel Design will also die.
Sounds like they're dying just fine on their own, but if we're supposed to be investing in this at a national level, why let Intel management ruin it first?
Can someone explain the logic behind the spin-off? In particular, 1) why an independent fab wouldn't end up exactly like GlobalFoundariess--i.e. no longer a contender in the leading-edge fab space; and 2) how Intel, a company that dominated in high-end performance largely because of the synergy between its design and fab units, would continue to be a contender in high-performance computing?
Intel as it currently exists may be destined to failure, but I don't see how splitting the company could do anything but seal its fate, if not hasten it. American MBAs have been hostile to vertical integration for over 50 years now, but TSMC is one of the few companies they can point to today as being dominate without embracing vertical integration. Most other dominant global industrial players, especially ones that aren't pure IP plays, are vertically integrated. Companies like SpaceX and Tesla that bucked the American trend and embraced vertical integration have done well precisely for that reason, whereas the story of those, like Boeing and countless others, that eschewed vertical integration continue to deteriorate.
Now, in the grand scheme of things vertical integration may not be economically ideal in terms of overall social wealth maximization, but to pretend Intel can remain competitive at the leading edge by eschewing vertical integration seems implausible. Far more plausible is that disassembling Intel is a well proven path for shareholders to profit on the decline with minimal risk. But given the political environment they can't sell it as promoting efficiency like they used to. They have to sell it as if they're attempting to do the very opposite of what they intend and has happened countless times; namely, managed offshoring of labor and capital intensive industry. But maybe I'm being too cynical? What's the steel man argument that is sincere about preserving Intel's market leadership rather than a managed decline?
> American MBAs have been hostile to vertical integration for over 50 years now
I think you are misusing the term "vertical integration". Intel is vertically integrated already, as it is an OSAT, Foundry, and Chip Designer.
TSMC is only a foundry.
Furthermore, the major difference between the Taiwanese ecosystem and American ecosystem is state support and opposition to antitrust.
The Taiwanese foundry ecosystem is split between TSMC and UMC, and both receive tens of billions of dollars a year in subsidized loans, land, and power from the Taiwanese government, as well as a blind eye to their overwork culture.
The US on the other hand never subsidized the foundry industry until the CHIPS Act was signed, and M&A has become much harder nowadays due to ideological differences, meaning that American companies cannot leverage the economies of scale the same way TSMC, UMC, Samsung, and other foreign chip fabrication players can.
Vultures are circling and with Intel trading at below the value of it's assets there's been an onslaught of articles suggesting that it would be best to cut up the company.
This would doom American leading edge fabrication and Intel as a whole.
Intel split apart will not survive. The fabs do not have customers with volume required to survive. They need Intel Design to be using them. With the Arrow Lake launch we're seeing that Intel Design, even with the latest TSMC node is behind AMD in performance. Raptor Lake, for all it's unfortunate voltage spike woes, was able to be very close to AMD when it came to performance in spite of being several nodes behind. This spoke to how well optimized Intel's designs were as an IDM for their own fabs. It's also why in spite of 18A being a very promising node has failed to pull many clients away from the TSMC ecosystem. Small advantages were built up over years which turned into larger ones.
Of all of those, only Amazon has a design planned for 18A. Just because chip designers exist doesn’t mean that their business is going to go Intel. If they had customers lined up they wouldn’t have been delaying their new fabs.
I expected at least a little bit more substance and insight, but the whole article basically boils down to, “if Intel’s fab splits off, they risk becoming another GlobalFoundaries.”
Which, yeah, I don’t think that’s much of a revelation to anyone discussing Intel’s future. There’s no argument put forward that it’s not still the best path for Intel.
Not specially Intel or Boeing but the entire USA-led system is derailed because that's how capitalism work: a big period of growth and a deep crisis "solved" with a reset, which is a war.
That's is. You can't save the corporate, you can only save intelligence of those working there, allow the corporation to fail and new small enterprises will born, some will became medium and some large again.
I went in thinking the same thing. After reading it, it's not 100% sure which type of bias it is. In the most charitable case, it could be sincere opinions formed by a career as Intel's CEO. An aftershock of a career in innovation.
Maybe it's the "grizzly thinks latching trashcans are a sign of hedonistic excess and should be shed" kind of bias. Less clear to me as an ex-ceo. Equity ownership could bend the needle that way for sure.
Honestly, more than the click-baity title, I was more amused by the subtle demand for even more subsidies ( that after putting US through offshoring, nearshoring and now taxpayer sponsored reshoring ):
"Barrett also said Washington must do its job to stay ahead of the semiconductor race. The U.S. has invested more in the semiconductor industry in the past year than in the last 28 years combined, but he says it should do even more, especially in academic research."
Makes one question whether he still owns substantial stake in the outcome.
I never did that kind of analysis, but the point is largely moot. They are going to get the monies bar some kind of event that completely changes the rules of the game.
That said, at certain point, even US government won't be able to bail out everyone and everyone is holding out their hands lately.
Intel's new Fabs in AZ started construction in 2021, before the CHIPS act has passed. It has had a gross profit in the tens of billions yearly and right now has ~$30Bn cash on hand. Not a single penny has yet been given from the CHIPS act to Intel [1]. So no.
That is obviously what the headline is saying. Don't know if comment guidelines apply to when one responds to a headline rather than a person in the thread, but per "respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize" (https://news.ycombinator.com/newsguidelines.html) it seems like an easy critique to point out what the headline says rather than critiquing the person's actual argument or the article's contents
> Barrett also said Washington must do its job to stay ahead of the semiconductor race. The U.S. has invested more in the semiconductor industry in the past year than in the last 28 years combined, but he says it should do even more, especially in academic research.
Oh. Okay. So why don't we split up Intel and then make this exact investment in the spun off divisions?
Why split Intel in the first place? Intel fabs die without the volume of Intel Design, and looking at the disaster that is the Arrow Lake launch, without the cost advantage Intel Design will also die.
Sounds like they're dying just fine on their own, but if we're supposed to be investing in this at a national level, why let Intel management ruin it first?
Taking actions which kill it faster isn't a good plan.
Can someone explain the logic behind the spin-off? In particular, 1) why an independent fab wouldn't end up exactly like GlobalFoundariess--i.e. no longer a contender in the leading-edge fab space; and 2) how Intel, a company that dominated in high-end performance largely because of the synergy between its design and fab units, would continue to be a contender in high-performance computing?
Intel as it currently exists may be destined to failure, but I don't see how splitting the company could do anything but seal its fate, if not hasten it. American MBAs have been hostile to vertical integration for over 50 years now, but TSMC is one of the few companies they can point to today as being dominate without embracing vertical integration. Most other dominant global industrial players, especially ones that aren't pure IP plays, are vertically integrated. Companies like SpaceX and Tesla that bucked the American trend and embraced vertical integration have done well precisely for that reason, whereas the story of those, like Boeing and countless others, that eschewed vertical integration continue to deteriorate.
Now, in the grand scheme of things vertical integration may not be economically ideal in terms of overall social wealth maximization, but to pretend Intel can remain competitive at the leading edge by eschewing vertical integration seems implausible. Far more plausible is that disassembling Intel is a well proven path for shareholders to profit on the decline with minimal risk. But given the political environment they can't sell it as promoting efficiency like they used to. They have to sell it as if they're attempting to do the very opposite of what they intend and has happened countless times; namely, managed offshoring of labor and capital intensive industry. But maybe I'm being too cynical? What's the steel man argument that is sincere about preserving Intel's market leadership rather than a managed decline?
> American MBAs have been hostile to vertical integration for over 50 years now
I think you are misusing the term "vertical integration". Intel is vertically integrated already, as it is an OSAT, Foundry, and Chip Designer.
TSMC is only a foundry.
Furthermore, the major difference between the Taiwanese ecosystem and American ecosystem is state support and opposition to antitrust.
The Taiwanese foundry ecosystem is split between TSMC and UMC, and both receive tens of billions of dollars a year in subsidized loans, land, and power from the Taiwanese government, as well as a blind eye to their overwork culture.
The US on the other hand never subsidized the foundry industry until the CHIPS Act was signed, and M&A has become much harder nowadays due to ideological differences, meaning that American companies cannot leverage the economies of scale the same way TSMC, UMC, Samsung, and other foreign chip fabrication players can.
Vultures are circling and with Intel trading at below the value of it's assets there's been an onslaught of articles suggesting that it would be best to cut up the company.
This would doom American leading edge fabrication and Intel as a whole.
Intel split apart will not survive. The fabs do not have customers with volume required to survive. They need Intel Design to be using them. With the Arrow Lake launch we're seeing that Intel Design, even with the latest TSMC node is behind AMD in performance. Raptor Lake, for all it's unfortunate voltage spike woes, was able to be very close to AMD when it came to performance in spite of being several nodes behind. This spoke to how well optimized Intel's designs were as an IDM for their own fabs. It's also why in spite of 18A being a very promising node has failed to pull many clients away from the TSMC ecosystem. Small advantages were built up over years which turned into larger ones.
Related:
Crises at Boeing and Intel Are a National Emergency
https://news.ycombinator.com/item?id=41920987
US still has AMD and Nvidia.
Which is a good reason to split Intel into fab and fabless design. We have the latter outside Intel. We need the former.
Yup, we don't need Intel, we need fabs.
Who uses the fabs then? Without Intel Design giving the volume, the fabs die.
AMD, Qualcomm, Apple, Nvidia, Amazon... Did you miss the part where we are awash in chip designers?
Of all of those, only Amazon has a design planned for 18A. Just because chip designers exist doesn’t mean that their business is going to go Intel. If they had customers lined up they wouldn’t have been delaying their new fabs.
Those are fabless companies.
I expected at least a little bit more substance and insight, but the whole article basically boils down to, “if Intel’s fab splits off, they risk becoming another GlobalFoundaries.”
Which, yeah, I don’t think that’s much of a revelation to anyone discussing Intel’s future. There’s no argument put forward that it’s not still the best path for Intel.
Not specially Intel or Boeing but the entire USA-led system is derailed because that's how capitalism work: a big period of growth and a deep crisis "solved" with a reset, which is a war.
That's is. You can't save the corporate, you can only save intelligence of those working there, allow the corporation to fail and new small enterprises will born, some will became medium and some large again.
Given his position, is it possible the Intel CEO is a little biased? /s
I went in thinking the same thing. After reading it, it's not 100% sure which type of bias it is. In the most charitable case, it could be sincere opinions formed by a career as Intel's CEO. An aftershock of a career in innovation.
Maybe it's the "grizzly thinks latching trashcans are a sign of hedonistic excess and should be shed" kind of bias. Less clear to me as an ex-ceo. Equity ownership could bend the needle that way for sure.
Ah well, I doubt either of us know the man.
Honestly, more than the click-baity title, I was more amused by the subtle demand for even more subsidies ( that after putting US through offshoring, nearshoring and now taxpayer sponsored reshoring ):
"Barrett also said Washington must do its job to stay ahead of the semiconductor race. The U.S. has invested more in the semiconductor industry in the past year than in the last 28 years combined, but he says it should do even more, especially in academic research."
Makes one question whether he still owns substantial stake in the outcome.
Intel would barely survive, possibly even declaring bankruptcy, if it weren't for all the subsidies they receive from the US Government.
What are you talking about Intel had yet to receive a single dollar!
https://www.tomshardware.com/tech-industry/intel-ceo-is-frus...
I never did that kind of analysis, but the point is largely moot. They are going to get the monies bar some kind of event that completely changes the rules of the game.
That said, at certain point, even US government won't be able to bail out everyone and everyone is holding out their hands lately.
Compared to other companies that are developing their own chips? Hardly moot at all. The CHIPS act is the only thing keeping Intel afloat lately.
No it isn’t, because no CHIPS act funding had been given to Intel yet.
How do you think Intel started construction on fabs in Arizona? Couldn't have done it without the CHIPS act
Intel's new Fabs in AZ started construction in 2021, before the CHIPS act has passed. It has had a gross profit in the tens of billions yearly and right now has ~$30Bn cash on hand. Not a single penny has yet been given from the CHIPS act to Intel [1]. So no.
[1]: https://www.tomshardware.com/tech-industry/intel-ceo-is-frus...
That is obviously what the headline is saying. Don't know if comment guidelines apply to when one responds to a headline rather than a person in the thread, but per "respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize" (https://news.ycombinator.com/newsguidelines.html) it seems like an easy critique to point out what the headline says rather than critiquing the person's actual argument or the article's contents