The thing that will surprise a lot of people is how low Intel pay is in the USA.
Just look up Hardware Engineer Intel on Payscale or other wage comparison site. Other companies are paying 50% more for the equivalent level. Eg. "Intel Bay Area Hardware Engineer salary" level 5 vs entry level "Google/Meta/Nvidia/AMD/Apple hardware engineer salary". On a direct, with no PPP or similar conversion comparison they do pay more for Bay Area talent than TSMC does for talent in Taiwan but Taiwan has significantly lower costs of living so you really wouldn't be better off even as talent in Taiwan migrating for an Intel job.
There's no way Intel is getting the best talent nor has it been for a long time. They do not compete in the slightest and they haven't in a long time. They do not even come close. Oh and they are doing layoffs of the cheapest hardware engineering workforce in the USA right now. Lets see how this pans out for them.
This 10000000%. They treat their employees like shit. Intel lost literally their entire red team to Oracle of all places over (at least) pay. Intel has a complete joke pay system that is like something out of the 1950s, underpays substantially and docks that already low pay based on certain qualifications like having a college degree (sorry Steve Jobs!).
I have zero interest as a taxpayer in bailing out their crappy company, let it fail or succeed on its own merits like the companies all of their employees moved to because they were better at retaining staff. Modernize your staffing system to hire competitively or die.
The "companies all of their employees moved to" don't make leading edge semiconductors with US based R&D, so I don't know why you're acting like they'd be able to do what Intel does.
This isn't about Intel, it's about keeping US companies competitive in leading edge semiconductor manufacturing.
If we let Intel fail, we'd have an even tougher battle to get some other US company to compete.
Ok, so let bankruptcy wipe out the shareholders and then let someone more responsible buy the good parts at auction. Or nationalize it. If shareholders aren't taking risk and responsibility, why the hell are we putting them in charge and letting them pay themselves massive dividends?
Excellent observation. Setting "preserving XYZ corporation" as a national priority is focusing on the wrong thing. Corporations are temporary constructs formed out of combinations of resources (human, capital, IP, etc) to organize and maximize the value of those resources.
If we really care about fostering and furthering these resources, we need to free them from being held captive, under-utilized and mis-allocated inside a repeatedly failing corporation - not propping up failing corporations. If the resources in question really are both viable and valuable, releasing them will allow the free market to do what it does best - efficiently allocate resources.
No central planning committee created U.S. economic leadership by picking which corporations to subsidize, so no central planning committee can "preserve" or restore that leadership by picking "winners" to subsidize. We need to let Schumpeter's creative disruption keep doing it's job of cleansing, educating and renewing the market through re-allocation of resources. That's what put us in the lead and it's what will keep us there.
1) No other US company is capable of using the parts for leading edge semiconductors.
Case in point: GlobalFoundry had EUV machines but returned them to ASML as they weren't able to compete technically and capital wise.
They would be the best to take over Intel's assets but they weren't competitive years ago, why would they be now?
2) idk where you're getting the idea that Intel is paying massive dividends when it has been one of the worst performing tech stocks for years.
The SW companies juiced up on low interest rates have done far better.
Shareholders already have huge risk and have lost big on Intel. Leading edge is too cutthroat, if it goes to bankruptcy it will be too late to ever catchup. It’s probably already too late and we just don’t know that we’re living in a Chinese century.
Earlier this month, Intel announced that they've suspended dividend payments entirely. Prior to that, their dividend payments were decent but far from "massive".
America succeeds when we let unprofitable companies die and it fails when we subsidize them to the point that they can't run an undistorted budget or be competitive anymore. That is the difference between here and the places that will never even get close regardless of how many unprofitable silicon wafer factories they allow to crawl around the ground like sick mutants.
These markets are already destroyed and have been by corporations.
America doesn't have an economy. Well, not really. We don't make anything. We used to, when we had a manufacturing industry, but they all left and now stuff is made somewhere else and shipped here.
Same thing happened with automobiles and will only get worse when the inevitably better Chinese cars come in.
But, at least, we had computer. Except now we lose that too.
Look, the US is currently only propped up on invisible money and greed. What we do it make someone else do the work, and then turn around and sell it for 10x more, and pocket the difference. We justify it with having busy bodies in suits sitting around "creating value" or something. Tada! Highest GDP in the world.
It's only a matter of time before these countries, like Taiwan, China, Bangladesh, India, wisen up and realize what's going on. Soon, they're gonna put two and two together and realize THEY are the US economy. They make everything. They have all the capital. They have the means of production. We don't.
We're on thin ice. If we let more US industries die off to be replaced by manufacturing in other countries, we won't have anything left. In my opinion, the situation is unsustainable. I don't know how much longer we have, and I think the consequences will be more severe than anyone is imagining.
Correct, but one can make a point those were an impetus measures that allow those to kick off and operate at competitive level
In case of Boeing their business pivoted in order to extract the funds themselves, not to use funds to make products. Thats the problem, corruption and profit chasing over product creation.
Yes, because any subsidy is literally communism. /s
How's the US farming industry doing?
I wonder why you brought up a completely different industry in a different country ran by a dictator instead of a core industry right here in the US that has been heavily subsidized for decades and for very good reason.
Tons of farm production is now going not into food, but into massive ethanol subsidies that are energy negative, damage cars and pollute more (ethanol generates smog, plus gas for tractors and equipment, plus fertilizer runoff creating massive dead zones in lakes and rivers and the Gulf) than gasoline does, and will ultimately have to go away in order for EVs to replace fuel engines (which they will now lobby against because it ruins their subsidized industry). There's probably some examples that more prove your point to some extent, this is definitely not one of them. My argument is that when governments try to manipulate or force market demand based on political priorities they tend to create problems.
If I was going to fight my own argument I would probably try the heavy government investment into silicon valley for military purposes. It clearly worked out pretty well. At the same time, the government didn't bail out Shockley after all the talent left to Fairchild.
2) ethanol is less than 50% of corn production, which is in itself about 25% of the crops in the US.
3)
> If I was going to fight my own argument I would probably try the heavy government investment into silicon valley for military purposes. It clearly worked out pretty well. At the same time, the government didn't bail out Shockley after all the talent left to Fairchild.
This is a bad faith comparison.
Intel isn't asking for money because its top talent left to start an even more successful US based semiconductor manufacturer.
Instead, the whole US semiconductor industry has been slowly off-shored after decades of US capital and talent doing the math and realizing that SW pays a lot better.
Intel is simply the last holdout due to its vertical integration providing higher margins which allowed it to afford the higher US wages, until recently.
If the US had started the CHIPS act 10 years ago when there were other US companies anywhere close to the leading edge (i.e. GlobalFoundry) then it wouldn't effectively be an Intel bailout, but that's not Intel's fault.
Intel doesn't have a monopoly on leading edge semiconductor manufacturing. The company that kind of does is ASML, they are based in the Netherlands, but there's an office in Connecticut.
While I agree their comp is abysmal, if Intel fails America is completely screwed. TSMC will never place leading edge in America and Beijing has the most advanced fabs in the world, which they’re already largely blocked from using, in rocket artillery range. None of our software roles mean anything if we don’t have hardware to run it on.
The truth is that hardware just plain sucks compared to software. More punishing to produce and less profitable to do so. Nobody with money really wants to back hardware over software, at least right now.
It's more like it pays better to be on a small team that can create software that is easily scaled to millions of users, running on hardware that is very difficult to produce but done using 2nd world and 3rd world manufacturing and supply chains.
The only conclusion from this is I think taxing the ever living daylights out of huge software companies.
Google/MSFT/Amazon all need to be looked at. We shouldn't be subsidizing them with our free shopping information data, nor the text we write in our content now with LLMs.
What do you think it says about the economic system? Software inherently has better economies of scale than hardware. That will be true regardless of the economic system. I don't love advertising but it's a small fraction of the entire software industry.
Ya it’s simple economics unfortunately (I’m a hardware engineer). Software can pay more because the cost to produce is lower and so the margins are higher. Hardware has abysmal margins - design, r&d, failure, debugging, schedule push back, component cost, supply chain logistics, warehousing, manufacturing, etc… the list goes on there’s a million things eating into the bottom line.
Using those examples when comparing software and hardware engineering _in general_ gives a skewed perception of things. Three of those companies are only interested in hardware as a way to optimize their (incredibly profitable) software business, and Apple/NVIDIA are two of the three largest companies on earth. That's a teeny tiny slice of the job market.
I'd bet that if you averaged out compensation from the top 100 paying companies for both hardware and software engineering rather than the top 5, the software number would be nearly double the hardware number. Once you start look at "normal" companies, the difference in margins between software and hardware and its effect on pay becomes very clear.
It took me several years to figure this out. And then several more years to do enough software side projects to build a full software résumé and get hired as a software engineer. The work is definitely less fun, but I almost tripled my TC overnight.
100% agree on pay. Also Intel is ossified, there is less and less team where you are held accountable. Empire building is rampant. One of the fundamental thing missing is vision and how each team is relate to that vision.
A friend left his SWE position at Intel(He was part of the Altera acquisition) about 4 years ago. His weeks consisted of answering one email on Monday and working out the rest of the work week because his lead did pretty much nothing, and their manager not only did nothing, but made shit up regularly.
One problem with paying people significantly below market is that some of your people will be disgruntled and will optimize for working as little as possible without getting fired.
Soon enough there are entire groups whose culture aligns with cutting corners and pretending to work.
Since the mid 2010s at least. It was common then that Intel would expect far greater for less pay than others. They're also known for paying half than the market rate for contract work too.
I'd say the pay is on the low side even when comparing outside the FAANG. Globally tech wages went up but Intel didn't budge due to long standing poor financial performance (see stock price since 2000) leading to a slow rot of their core strength.
In a direct dollar to dollar conversion Intel in the Bay Area (the highest paying location) pays pretty much exactly 2x in raw salary for equivalent levels to TSMC roles based in Taiwan. That might sound like a pro-Intel comment but the simple fact is that it's a lot cheaper to live in Taiwan. 2x Taiwanese wages are not going to entice anyone to move to the highest cost of living region in the world. TSMC clearly has amazing talent and Intel are not only failing to compete against the other US tech giants, they can't even entice from lower wage regions of the world right now.
It's pretty obvious that there's talent Intel needs and if we want to onshore the production from Taiwan we need that talent but Intel isn't coming close.
The main contributor is high real estate prices. Much of the value generated by tech companies is eventually captured by property owners. Geography and open space preserves prevent building out anywhere within easy commuting distance. Politics prevents building higher density housing in most neighborhoods.
The secondary contributor is high taxes. The Bay Area, and California in general, isn't the worst for taxes but in terms of total tax burden (income tax, sales tax, excise taxes, mandatory fees) it's near the top. Especially because state income taxes are no longer deductible on federal income tax returns.
Are the real estate prices high because of high salaries? This is what I'm getting at- the infinite money glitch tech salaries skews a lot of things downstream.
> Atkinson says that while the CHIPS Act is a start, the United States is likely to need additional funding, support, or incentives if the goal is to see competitive leading-edge semiconductor fabs owned by U.S. corporations. “We’ve hit a double in the first inning, and that’s pretty good,” he says. “But the problem is everybody in Washington thinks we’re done.”
This is the problem.
Asia has seen semiconductor manufacturing as a strategic investment and so setup long term benefits, subsidies and education programs.
They also have leadership that stays in power long enough to see these plans through.
Meanwhile the US has been enjoying the easy gains of SW companies juiced up on low interest rates (which leads to inflation, which favors industries that require very little capex).
Finally people are realizing that we still need HW companies but all we get is a one off band-aid solution, just enough for the industry to survive until the next election.
It really is like the old joke that Asia is playing go while the US is playing chess (strategic vs tactical thinking).
One thing that Asian governments have insisted on is export discipline. Companies could only get grants, loans, etc if they exported their goods. The idea is to force domestic producers to be good enough to be competitive internationally. How does one make sure that the US chip manufacturing industry stay competitive without this sort of export discipline?
That policy made sense for Asia because they got their start through technology transfer agreements.
This policy ensured that these agreements weren't just thinly veiled attempts to bypass import tariffs and were actually commercially viable in the global market.
> How does one make sure that the US chip manufacturing industry stay competitive without this sort of export discipline?
If we keep low tariffs on SK and Taiwan, I don't see how a domestic manufacturer has any choice but to compete in the global market, removing the need for this export requirement.
This is part of why US semiconductor manufacturers have had a hard time competing with Asia, especially in commodity semiconductors (i.e. memory and packaging the first to get off-shored).
> Not clear if US companies can catch up at this point.
I'm biased as I'm working on 18A at Intel but I'm very optimistic for this node. We could leapfrog TSMC (this is all public information).
And in the last few years Intel speedran 4 nodes to catch up, which was quite the feat.
> I think another efficient goal would be to invite politically friendly asian companies to build fabs in US.
This was part of the CHIPS act too (which I think shows how scattered the US semiconductor policy is), TSMC and Samsung got money to do exactly that.
It's a really large company so, has you pointed out, this is just my experience.
Pros:
It has provided compelling compensation in Oregon (significantly more than my previous position, but I'm junior so large jumps are normal) and work (leading edge tech in my field, plenty of external eyes looking at it).
All my managers all the way up to and including the CEO are very successful engineers and understand the technical detail of what they're managing extremely well. Pat is very capable of talking about details of upcoming processes during internal meetings, for example.
I also feel empowered by my manager and colleagues to point out how we could improve our methodology, although that often results in extra projects for me, requiring some sacrifice of personal time if I want to complete them all (which is what I'm looking for right now but not sustainable in the long run).
I quite like OKR system and I feel like my manager really understands what I'm doing and provides useful feedback.
Cons:
I feel there's a bit of a "fire-fighting mentality" in parts of the foundry, especially as you get closer to the fab. There's an unspoken pride in spending your day working late to fix some issue.
Many teams are constantly in meetings debugging issues caused by human error and no one seems to ask the question "how can we automate this so it doesn't happen again?" because they're HW not SW people.
A lot of my automation side projects require buy-in from other teams and there's been some friction in showing them the value of their cooperation ("this is how we've always done it" kind of attitude). Thankfully I have more senior engineers and managers who have a vision to fix these issues so it's going in the right direction.
I believe this is not unusual in HW manufacturing.
If you get placed in a team that has a bad methodology and has a manager that has no vision on how to completely revamp it, then it's going to be a real slog.
If you're thinking of accepting a position at Intel Foundry, feel free to drop contact and we can talk further about my experience here.
If you don’t mind me asking, what is your role in particular? Are you a module engineer, a metrology engineer or are you in design enablement? Also, is Intel Foundry currently hiring given all the cuts and layoffs? I hope your job is safe amid all the turmoil!
> If you don’t mind me asking, what is your role in particular? Are you a module engineer, a metrology engineer or are you in design enablement?
I'm a test engineer in Design Enablement, specifically in Advanced Design testing high speed IO.
I'm sure you can find my Linkedin through my callsign if you'd like to know more details or reach out.
> Also, is Intel Foundry currently hiring given all the cuts and layoffs?
Yes, there's case by case exemptions, just depends on that team's needs and staffing levels.
There's also voluntary internal movement as part of the pre-layoff process.
Companies that want bailouts, if they are of national interest, should get those in exchange for partial ownership. When private ownership fails, national funds should come with national controls.
Agreed, I'm willing to concede the point that any great power needs a chip industry but I hate my money getting put into the incinerator that is modern Intel with no strings attached
This is how the GM bailout was structured, though the government ended up selling their stake in GM after it was rebuilt.
Based on the performance of the companies that the US government does own, this seems like a really bad idea in general. Amtrak has been a money pit for special interests for decades and the USPS is only profitable because they have a legally protected monopoly they constantly abuse to shove spam into the same mailbox where I occasionally receive bills and jury summons (to the point where my USPS service is easily at least 90% spam).
It's also how the bank bailouts were structured. They weren't free money. They were loans plus government ownership (which the government typically sold off for a nice profit). The loans were typically 5% which would rise to 9% after the first five years.
People talk about the bailouts like they were free money and say things like "the government should bail out ordinary people." The reality is that they weren't favorable terms - but reasonable enough that the banks wouldn't fight taking the money. The government wanted to stabilize the banking system (and automakers) and it wasn't totally clear who would need money. For example, how much money do you need over the next month? Well, that depends on factors you don't know about. Maybe you have a medical emergency or some other unexpected thing happen. So the government wanted the banks to accept the money rather than say "nah, we think we can weather the storm" and then fail creating more chaos.
If the government had said "we'll give you $10 and own half your company," no bank would have accepted that. Instead, it was typically structured as preferred shares (essentially a loan) offering 5% interest for the first 5 years and 9% after that plus warrants for ownership of the companies equivalent to 15% of the money invested. When the financial system stabilized, the banks typically repaid the loans and bought the warrants from the government.
For example, with Citibank, the government gave them $45B and made a $12B profit 2 years later - a 27% increase over 2 years. That's pretty good. With General Motors, the government gave them $13.4B in exchange for $2.1B in preferred shares and 61% ownership of the company.
People act like they got free money, but that simply wasn't the case. The government got a reasonable deal on that money, especially given that they were trying to force the banks to accept it. Some believed that they could go without the help and that the "bailout" terms weren't in their best interest. They didn't want to make a deal while their value was low since the government would be getting a lot for their buck. But the government's terms were reasonable enough that some cajoling got them to accept it.
I mean, Citigroup went from $900B in July 2007 to $400B in July 2008 to $50B in January 2009 (and around $85B when the government sold its ownership in 2010). Citi did need the bailout and the US government ended up with 36% ownership. Citi clearly wouldn't have sold 100% of the company to the government for $5B in 2008 or 2009. At that point, you might as well roll the dice and take your chances. The government gave Citi $45B and got $22.2B back as cash from Citi and 36% of the company (basically, 36% of Citi for $22.8B). Note: when you buy large chunks of companies, you don't get the market price. As you buy more shares, you get to investors unwilling to sell at that price and the price goes up. Maybe 1% of shareholders might be willing to sell at $50B, but then some will believe they should hold rather than sell. So the government didn't get a bad deal.
What matters the most is the product. Intel has had a serious of serious problems on that front over the last decade. From product delays, to Speculative Execution security concerns, to oxidation/voltage issues. Meanwhile their competition has been absolutely killing it, to the point where their X86 moat is collapsing. Their major software partner, Microsoft, is spending $$$ to open the option to work with other vendors. The fact that Apple successfully walked away is a major advantage for MacOS over Windows.
All the tax incentives, all of the trade deals, all of the weight of the federal government won’t matter as long as Intel continues to produce inferior products. Now they have 15% fewer minds to work with. In twenty years it won’t matter if Intel is still operating fabs or not, the chips will be so far behind all the serious workloads will be on other silicon.
I don’t have any mind for policy. Law just seems like very confusing hyper-spaghetti code.
But, I really hope the CHIPS act is well-targeted.
We don’t need Intel, the chip designers, anymore. I mean, their chips are fine basically, or maybe they aren’t so great, whatever, they can still sell them, I don’t care.
But we absolutely and desperately need a chip foundry company and if IFS can be that, they deserve a chance (and are probably the best bet we’ve got, unless somebody can convince TSMC to send us some folks, but that seems pretty unlikely).
Intel has 10 wafer fabs worldwide. Yes, not all of them are at maximally modern nodes...
But here's an idea. Let's spin off Intel's fabs not into one company, but two. Get a little anti-trust and market competition to go with the CHIPS investment.
Intel won't like it, but make the CHIPS funding dependent on it.
Yes I know, no way that gets out of congress intact.
It would be a fascinating experiment if we could try it safely somehow. Break up Amazon, Intel, Microsoft, Apple, and Amazon and watch their various parts try and figure out how to make money. A world in which I can buy a Graviton chip fabbed in the US by IFS, and then consider installing Windows on it (nah, but it would be nice to have the option) would be really neat.
Who will be their customers? Intel’s woes are all because their fabs fell behind, if they were fabless they would just go to TSMC. Global foundries could keep up and had to give up on leading edge. Without Intel as their own customers Intel fabs will die.
It’s trailing edge. TSMC is doing it to soak up subsidies, deny it to others and avoid stick policy solutions. But they will never put leading edge outside of Taiwan, and should they become a true monopoly with Intel and Samsung throwing in the towels then it will become very unpleasant.
Sadly it has been from the beginning. Companies in the US only care about profits, not about what is good for the Country they are in. And in reality that is how capitalism works.
Pouring money into these companies will do nothing. What is needed is laws with real punishment for the Company and maybe the CEOs. IIRC, there was a time were all defense contractors had to be 100% sourced in the US, not now.
So I would say a law like this: "70% of all products you sell and all the components must be 100% built in the US. If not, you are fine of 30% of your pretax worldwide revenue, including international subsidiaries". If you leave the US, you will be nationalized.
Advanced economies simply don't work like this. You cannot have an entire generation raised with the mantra "You must go to college. Take loans. Go to college" and then expect to have a strong blue collar work force. I went to college to sit at a desk at home and get a salary, not commute to a job site and make an hourly.
And even without the mantra, who the hell wants to be the skilled technician working 60 hours a week in a clean lab to push out top quality silicon, when they can be the skilled SWE fiddling with a login page for twice the pay, from their home office, for 30 hours a week?
If you break it down, it becomes more and more of an irreconcilable problem.
because different people like to do different things? twice the pay is a benefit, sure, but then you have to sit in front of screen all day long staring at it. Some people wanna work with their hands, and really don't want to sit in front of a screen, no matter how much you pay them. some people even forgo a livable amount of pay, and choose the starving artist route.
not everyone optimizes for pay, nor do they have the ability or aptitude or even desire to become a software engineer.
You notice how the right (or really any politician or think tank blabber) doesn't trump the "free market" anymore? It used to be every other minute on CNN in the 1990s and 2000s.
Why is that? Well, pretty simple, the corporate powers have all largely settled into cartel, duopoly, or outright monopoly alignments within their markets. Market competition is pretty weak in the US, and it is hurting our competitiveness worldwide.
The reality is that the nth-iterative stability of "free markets" in the real world with barriers to entry and even more important, regulatory capture, is cartel in best case.
A lot of the centralization of wealth in the US is through the large cartel megacorporations. Using those, the "elite" can extract wealth and wage war on labor costs, which goddamn they LOVE to do, more than shirk taxes. Oh, and the megacorporations also enable massive tax avoidance (loans on existing unrealized stock value as a great example).
So what the government SHOULD be doing is antitrust right and left, and breaking up all these monopolies and cartels. It would inject a huge amount of labor demand (at good job wages), a huge amount of anti-inflationary competition, and produce massive amounts of innovation, since you'd have a dozen companies trying different products and production methods in markets, as opposed to ... two.
I just would like to point out the irony that you want to break up companies to improve our global competitiveness. I seem to recall hearing in the late 90s and early 2000s a theme that the US really needed to let these companies combine so we could have large national players who would have the scale to compete globally with other nations' mega-corps. (Not so much in terms of FAANG tech perhaps, but certainly in banking and many other sectors.)
I'm not sure if there is some definitive research on what corporate size truly helps global competitiveness (but I'd expect that, as you say, smaller competitors in larger numbers tend to be more competitive.) But even in China car industry as transitioning today, the natural evolution of such a system is towards a fewer larger competitors. Kind of related to the iron law of oligarchy perhaps, this phenomena? https://en.wikipedia.org/wiki/Iron_law_of_oligarchy
It might be a little bit "on the nose", kinda like "speaking ill of the dead" too soon. Intel is not yet interred...
Interesting to see how Intel's hubris has painted it into a corner. Will the US govt. consider it "too big to fail" and infuse it with more cash as needed? Time will tell.
P.S. If the govt. does provide cash, what will prevent some future CEO from short-term juicing the stock using that cash, collecting a big paycheck (and running away before the stuff hits the rotary object?)
The US government would be even worse than Intel at running their business.
The more likely scenario is that the US govt declares Intel "too big to fail" and they give them a bunch of money. They are already doing that through the CHIPS Act.
My favorite part of this is how not only do we kick the can down the road, but the executives also get even more money. I assume that's their favorite part, too.
Sort-of. You can't just take without just compensation so it'd probably cost the government $200B (you can't buy a whole company just at its market cap).
But when what? OK, you've nationalized Intel. All the same people are working at Intel. You've bought Intel, but you haven't really changed anything. GovernmentIntel faces all the same problems that Intel faced like high US wages and having been behind on EUV lithography.
The biggest problem with Intel is that it ignored EUV for a decade hoping to save money. If the US nationalized Intel in 2012 and the government said "you're going to invest in EUV," then Intel would be in a different place today. Nationalizing Intel in 2024 wouldn't really do much. They've already changed course and should have High NA EUV chips out years before TSMC.
Of course, the US could have also just said "we'll offer billions if you invest in EUV fabs in the US" back in 2012.
I guess the question is: what does nationalizing Intel give you? It gives you control, but you also basically already have control if you're willing to give them money - and nationalizing companies typically costs more than incentives.
Some people think countries should keep a certain amount of certain industries around, regardless of whether they can be run profitably, just in case a big war breaks out and you need those things and can't source them from abroad.
A government might keep things like shipyards in business, even if they can't compete in the global market, just in case we ever need a rush order of new warships. And keep steel mills in business to supply the shipyards, and so on.
Because modern weapon systems need modern integrated circuits, if there was the political desire the state could keep Intel on life support forever.
- Xi's cult of personality and curtailing of capitalism
- A Taiwan invasion attempt that brings massive blockades and boycotts
- Demographic collapse from decades of poor fecundity
- Financial house of cards / property crisis
Zeihan used to be a cuckoo clock of saying this stuff every year, but recently most of the poo-pooers have gone quiet and even started saying things that quietly agree with him, so I think China doomism is entering mainstream policy planning.
China is undeniably being far more militaristic and aggressive in recent years, and this combined with the COVID shock has led to a large rethink in policy and a questioning of outsource manufacturing reaching the point of sharp security concerns.
Globalization being "over" is debatable, but the notion that the world is in a post-war "new history" is dead with the Ukraine war and China's increasing aggression.
The underlying big effect is that countries that are industrializing can grow economically at about 10% per year by copying the advanced countries. When they catch up, growth drops to about where their new peers are, because the fast catch-up phase is done. Happened to all the "Asian tigers" - Japan, Taiwan, Korea, Singapore. Now it's happening to China. Normal part of progress. Good summary in "How Asia Works", by Studwell.
Everybody has a property overbuilding crisis. "Build it and they will come" seems to be a universal for developers. What matters is how they come back from it. Japan took about three decades. The US took about three years.
>>Xi's cult of personality and curtailing of capitalism
The president of China is just a glorified code reviewer to some policies made by policy agency bodies. Sure he commands lots of political control over the broad contours of those policies, but nevertheless the control is limited.
What matters is making the right decisions, regardless of whichever 'ism' camp they belong to. Just do what is right, and what it takes to win.
This is seems to be more like a guiding principle behind whatever China does. During Deng era they redefined what communism means, they keep redefining communism as whatever they have to do to win.
Which in itself is a strange thing. They don't seem to care all that much ideologies, or anything for that matter. Only winning and what it takes to win seems like their ideology.
>>A Taiwan invasion attempt that brings massive blockades and boycotts
Another great part about China seems to be a stubborn refusal to fight wars. Which is a very scary thing if you think about it. The USA has perfected one weapon in all its existence, and its biggest nemesis is inert to it.
By and large they are just minimising going down decision trees that end in failure.
>>Demographic collapse from decades of poor fecundity
Tall order. Every one's fertility rate is declining. More importantly due to proliferation of AI and automation, this is actually a welcome thing. Lesser the mouths to feed, lesser instability when the machines inevitably take over.
China fought Vietnam after America. It didn’t go very well. It’s a defacto island with difficult terrain to the south, west, and north, while the first island chain keeps the lid in it. It hasn’t been until recently that it had a navy which could seriously reach beyond. It has been building everything it can and needs to break out of the first island chain by taking Taiwan and it knows that to do that it would first need a force capable of beating the Americans before doing anything.
Just to add to that:
AMD did used to make their own chips, but spun it off as GlobalFoundry many years ago to avoid bankruptcy when times got tough.
GF was on the leading edge for a long time and even had EUV machines but actually decided to send them back to ASML and stop chasing the leading edge due to lack of capital.
Since then, Intel has been the only US based company on the leading edge.
The US government blocks all kinds of company sales. Sometimes for anti monopoly reasons, sometimes for national security reasons.
10 years ago I worked for a company with less than 2,000 people. We made a chip that was used by a few gigantic companies over 150K employees. There were only 3 companies that made this kind of product. My company tried to sell a business unit with about 200 people to our competitor and get out of this market. Our two gigantic customers went to the US FTC Federal Trade Commission and got them to block the deal over anti competitive concerns. With only 2 competitors these gigantic companies expected that they would have to pay more per chip. They used their influence to get the government to block the deal.
At least some of the equipment (EUV machines being the big example but even things like high bandwidth oscilloscopes and many other tools we never hear about) are already export restricted (export is still possible but requires approval).
Whether the executive could place a blanket "no export" on all of it without a legal battle is a very different question.
Edit: this is only for whether or not the equipment can leave US soil. Whether it can be sold to a foreign company but stay in the US is a different question (the answer is yes if I had to guess)
What’s interesting is that AMD used to have a fab… then it was spun off… then it gave up on keeping up with TSMC's latest nodes. Possibly for the same reasons Intel is struggling right now.
There’s a big difference between designing chips and manufacturing chips. The concern here is Intel is the last US HQ CPU fab that matters, and its relevance is waning.
> The concern here is Intel is the last US HQ CPU fab that matters, and its relevance is waning.
I gave up on US fabs a long time ago; are they really that relevant to begin with outside of paying lip-service to those concerned about offshoring manufacturing jobs? We just went through (and are continuing to experience) a chip crunch and it's not like Intel came through despite having a domestic fab.
The cost of a current top end fab is astronomical for experienced semiconductor companies. There is no "skunkworks" to get a top end fab going.
The real prayer is that Moore's FIRST law / node shrinks run fundamentally out of steam in producing efficiency or speed results, and that gives the entire industry time to optimize and drop fab equipment costs for near-top-end fab performance. This may actually start to happen in the next 5-10 years.
I think Jensen Huang stated that Moore's law is dead a few times. I've seen projections of cost per transistor for the leading node being more expensive than the trailing node. The only reason to pursue it would be that the market can bear the additional cost because power or some other attribute is better.
Our legacy "computer industry" was a lend/lease styled business model industry. Our legacy "computer companies" were financing companies. Not "computing" companies. That lends/lease financing business model justified the sales forces operating on commissions. All that is over now.
Everything has changed now. Computer automation (CPU, RAM, Storage, I/O) is now priced in the mere 100's of dollars. That signifies "stuff" you can buy at Best Buy or even 7/11. Stuff you can "buy" signifies "stuff" that cannot be sold on commission. There is no more role for financing (lend/lease). There are no roles for sales forces operating on commissions.
What? We are talking about a major IC designer/manufacturer not gateway desktop computers from 1999.
Theres still tons of money to be made in chips - the whole world runs on them. but the cost is in the multi-multi-billion dollar r&d to build a fab that actually makes functional chips. And then having the work force that can produce successful yield after successful yield 24/7 that keeps the economics of the whole thing on par with Taiwan and other major supply chain countries in the indo-pacific where wages are lower.
Just look up Hardware Engineer Intel on Payscale or other wage comparison site. Other companies are paying 50% more for the equivalent level. Eg. "Intel Bay Area Hardware Engineer salary" level 5 vs entry level "Google/Meta/Nvidia/AMD/Apple hardware engineer salary". On a direct, with no PPP or similar conversion comparison they do pay more for Bay Area talent than TSMC does for talent in Taiwan but Taiwan has significantly lower costs of living so you really wouldn't be better off even as talent in Taiwan migrating for an Intel job.
There's no way Intel is getting the best talent nor has it been for a long time. They do not compete in the slightest and they haven't in a long time. They do not even come close. Oh and they are doing layoffs of the cheapest hardware engineering workforce in the USA right now. Lets see how this pans out for them.