It's silly to post a headline of a large operating loss when the unit was literally partitioned off a few months ago. Building fabs is expensive... literally the most cost intensive manufacturing enterprise in the world. Making those investments is the only way to stay competitive, and I'm sure we don't want to live in a world where there's a TSMC monopoly on the high end, and Chinese state run semiconductor fabs on the low end.
Yup. This is more of a signal on 1) experience with new EUV processes and 2) cost of labor. That's not taking into account the expensive EUV amortization.
A hard truth I've heard in CEO circles is that western middle classes are falling down in terms of purchasing-power; And they will continue to do so until they join Chinese middle classes levels of revenue.
Which country's business leadership did you hear that from?
I'd be fairly surprised that American disposable purchasing power would fall to around $3-4k USD a year, unless by "middle class" your peers meant the professional class in China (the kind that can afford to splurge on a new iPhone every year, buy a Xiaomi SU7, and send their kids to do a bachelors degree at UCB, UCLA, or UIUC), in which case I wouldn't be too surprised simply because of how much growth was captured by that subset, and salaries for that subset are at France/Germany levels, but those countries also severely underpay professionals tbh.
Even professional salaries at employers like Google India or Gojek Engineering in Bangalore, Jakarta, etc can pay German or French level dev salaries, let alone the subset of similar companies in China that can afford to pay around $20-100k USD salaries (depends on the city)
> signal on 1) experience with new EUV processes and 2) cost of labor. That's not taking into account the expensive EUV amortization
Yep. There's a reason Intel and GlobalFoundaries have the last American owned Fabs - most other players left the industry in the 2010s due to the massive costs, though it's a similar story in China as well which was only solved by the "Big Fund", but a similar Cambrian die off happened in the 2017-20 period which basically only left SMIC (comparable to TSMC or Intel in the early 2010s), Hong Hua (comparable to Tower), and YMTC (comparable PSMC or SK Hynix)
> like Google India or Gojek Engineering in Bangalore, Jakarta, etc can pay German or French level dev salaries
Not even close. Indian devs at top companies earn way more then European devs at local non-top-tech companies.
>unless by "middle class" your peers meant the professional class in China (the kind that can afford to splurge on a new iPhone every year, buy a Xiaomi SU7, and send their kids to do a bachelors degree at UCB, UCLA, or UIUC)
How many Chinese families (percentual speaking) can actually afford that?
Depends on the company you are working for in India.
At a product company like Ola or a top MNC like Google you are earning way above most Europeans excluding Netherlands, Switzerland, and maybe London, but if I had to estimate, Product+MNC companies only make up like 30-40% of the ecosystem in India today.
If you're working for a mass recruiter your mid-career salary will top off around the $8-20k range, which is Eastern European level, but lower than France and Germany (where it's around the $40-60k range).
Also, there are regional variations in salaries in India as well. You're more likely to earn much higher working in an established tech or MNC hub like BLR, HYD, NCR, or PN than in Indore, Kolkata, Chennai, or Ahmedabad, where the tech scene skews towards outsourced contracting.
Government dev salaries are lowish (depending on the Ministry and whether it's central or state), but the perks are solid.
I'd probably say the split is 35-45-20 Product-WITCH-Govt.
Edit: You said "Top". Fair enough. Yea, the salaries at top employers like Google India or Flipkart are extremely competitive by European standards and are comparable to those in Mainland China, Israel, and Canada.
Are you talking about software and tech in general? Because that's several times lower than the average. In every single East European country in the EU, the average ~~monthly~~ [edit: yearly] wage (I'm talking about the whole population) is at least 18-20k and you certainly can't find any actual developers for anything even close to that.
> single East European country in the EU, the average ~~monthly~~ [edit: yearly] wage (I'm talking about the whole population) is at least 18-20k
Median disposable household income in Poland is around $600/mo [0], only $200/mo more than in China
Median monthly household income in Romania is around $545/mo [1], so disposable is probably comparable to China.
Median monthly household income in Bulgaria is around $600-700/mo [2], so disposable is probably comparable to Poland.
The only Eastern European country that broke the $1k/mo barrier was Estonia with around $1k/mo [3].
These are all government statistics from the Polish, Romanian, Bulgarian, and Estonian government.
Eastern Europe has developed significantly, but still lags behind Western Europe massively.
Based on the hiring packages we've given to developers in Warsaw, the salaries were around $30-40k, but we could have paid much less if we hired someone from EPAM.
I'm taking umbrage to the fact that OP said the "average ~~monthly~~ [edit: yearly] wage (I'm talking about the whole population) is at least 18-20k" when it is obviously bull.
I gave direct sources from the OFFICAL Stats Agencies of Poland, Romania, Bulgaria, and Estonia in 2022-2023. You just gave Wikipedia.
The numbers which are on Wikipedia are AVERAGE. I'm using MEDIAN
The fact that the AVERAGE income skews significantly higher than MEDIAN income is proof that you have a subset earning much much more than the majority.
I have to question your reading comprehension. You're talking about median DISPOSABLE INCOME per capita (so it would include, disabled, retired, unemployed, part-time etc. workers). The Wikipedia link showing the average WAGE/SALARY. These are not the same thing.
Even in Poland the median household disposable income (the metric I am using for China) comes out to around $600/mo [0] and is comparable to Thailand.
Polish boosterism is almost as annoying as Indian boosterism tbh - both seem to come from a place of insecurity and heavily play up some positives while ignoring the negatives. If you listen to Polish or Indian nationalists, you'd think Poland or India are major superpowers already and have been for decades.
Is this a side effect of PiS's hypernationalist media in the 2010s? Or is this some weird alt-right Catholic White Nationalism fetish?
> both seem to come from a place of insecurity and heavily play up some positives while ignoring the negatives
> Or is this some weird alt-right Catholic White Nationalism fetish?
I see a lot of projection going on here. No, you were not talking about average disposable income you were talking about salaries for workers in tech. This is what you wrote:
> if you're working for a mass recruiter your mid-career salary will top off around the $8-20k range, which is Eastern European level, but lower than France and Germany (where it's around the $40-60k range).
The average salary for all employes in all sector in Poland is > $24k. Are you implying that you can hire tech workers or even software engineers for considerably less than the national average? Seems highly far fetched.
Was this an unintentional mistake or are you doing it on purpose. Considering the totally uncalled for diatribe about Polish(I'm neither Polish nor have anything to do with it for that matter) nationalism unfortunately it's probably the second?
You made it about average incomes when you said your average Eastern European - not Eastern European techie - earns around $18k-20k a year when every government from Poland to Romania to Bulgaria has provided statistics proving otherwise.
No, I was talking about average salaries because we were comparing salaries.
> government from Poland to Romania to Bulgaria has provided statistics proving otherwise
Can you please stop being obtuse? That's not particularly entertaining at this point..
> earns around $18k-20k a year
The average salary in Poland AFTER TAX is around $17k, GROSS is almost $24k. I'm not sure why are you fixated about this. That average income is tangential to the whole discussion fact is you certainly can't find any software developers/etc. for $8k primarily because that's actually below any minimum wage in any EU country besides Bulgaria. You would really struggle to hire anyone even remotely decent for $20k as well.
The phrasing of 'a hard truth' makes it sound as if it is inevitable, and those in 'CEO circles' are simply bystanders to the phenomenon. Why is it that the purchasing power of CEOs has drastically increased, while the purchasing power of the middle class has decreased?
I guess it's just one of those darned hard truths.
The average CEO's insight into the wider economy is zero. When a CEO (or anyone for that matter) says "a hard truth about x" where x is some complex system like the economy, roll your eyes.
In real terms I'd expect Chinese middle class to be converging up to European standards right now. The Chinese have quite a lot of infrastructure; by the numbers [0] they're capable of sustaining a really comfortable existence. It hasn't stabilised but China isn't a byword for poverty any more.
That being said, it does seem likely that the US living standards will drop. It isn't obvious what is supposed to be driving living standards given that the US has been undermining its own energy and manufacturing sectors. Their political situation also seems to be deteriorating.
The difference is population, there are A LOT of Chinese that are in the rural areas still making very little.
European wealth is generational. Land and homes that have been in the family from infrastructure that has stood the test of time.
Chinese wealth is rather New money and it’s all paper thin based on construction that isn’t built to last. Look at Guangzhou they are already relocating their downtown. Before it was Baiyun then it became tianhe, now they’re focusing on making New Baiyun already in just 20 years.
> Chinese middle class to be converging up to European standards
Depends where in Europe and where in China.
Beijing, Shanghai, and Greater Bay Area have definitely converged with Central Europe (eg. Praha, Warsaw, Budapest, maybe even Wien)
The rest of urban China has converged with Eastern Europe (Russia, Poland, Türkiye, Serbia) and Thailand+Malaysia.
Rural and some of the poorer provinces are still stuck around India levels.
There's a reason why household incomes per capita are still around $450/mo [0] - significantly lower than Thailand [1] or Serbia [2]. Per Capita Urban Household Income is around $7k/yr but for Rural Households it is around $2.5k/yr
I know some Chinese will find this a hot take, but this urban-rural divide is very similar to Brazil and Mexico, both countries that share similar developmental indicators to China.
Warsaw as most of the eastern part Poland is still heavily influenced by the post-Soviet mentality. Warsaw is even less diverse than ~380k-headcount Lublin (also eastern part) due to Lublin's universities welcoming international students with relatively cheap living costs for years now.
I live in Poland for 40 years, I'm biased, but I'd say that if you're looking for "western" cities - it's not Warsaw but Gdańsk or Poznań. Warsaw will keep its post-soviet mentality at least until the generation remembering WW2 is alive. Of course younger people may lack exposure to it, but I'm generalising on city/region-level.
I'm just solely going by developmental metrics and median household income.
Eastern Poland might have a fairly conservative society (as is reflected by Sejm and Presidential election results), but because Warsaw is the political, financial, and social capital of Poland, social and economic infrastructure is much stronger there than in other parts of Poland - west or east.
It's the same way that Beijing is luxurious but very socially conservative compared to Shanghai.
It's usually 5 years, Intel now do over 8 years, fabs can have a useful lifetime much longer than that. However Intel historically recycled old fabs to create new ones, so in practice they don't have many old fully depreciated fabs.
My understanding is that the various pieces of a fab are depreciated over different time horizons. Some of it as short as a few years, some of it as long as 30. I'd guess it evens out around high single digits.
This doesn't account for the fact that Gelsinger kicked off IDM 2.0 in 2021 and is this far in the hole already and the investments chosen have put them further behind TSMC before they even get started.
Gelsinger is trying to do too much in too short of a timeframe creating disarray at Intel, just like he left in his wake at VMware. I'm not sure that he has the landscape vision that's required in this market. I think he's likely better cut out in a more CTO type role at Intel rather than being at the helm.
And even if Intel is on a better long term path, Gelsinger didn't see the short term hurdles as he should have. These are the American taxpayer dollars being wasted in high executive pay and corporate greed under his leadership. Gelsinger should be held better accountable if Intel is looking for more handouts [0].
This and the fact that the point is obvious that Intel is already behind schedule based on the fact there are cost overruns. Then let's ask if, because Intel is behind, are the executives still being paid as if Intel is on a positive outlook in the short term? They are. Gelsinger, for example, took no hit even though is salary was reduced. [0]
And even though there were cost cuts they're still off by $1.8B.
Also $8.5B is in direct funding and access to up to $11B in federally backed (and discounted) loans. [1]
Executive comp and performance is interesting largely because it entirely depends on the counterfactual alternative, not actual outlook. There is bulletproof measure of impact, and it is subjective in the assessment of the BOD.
A CEO that oversees a loss of $1B could be doing a great job if you think the counterfactual was a loss of $2B.
A CEO that oversees a profit of $3B could be terrible if you think the counterfactual was a profit of $4B.
For the most part it is irrelevant to 3rd parties, because CEO compensation comes out of shareholder compensation, so it is their money to spend or throw away based on their assessment.
Intel's highest year of top-line revenue and gross profit was in 2018 @ at $22B. Their lowest, comparatively, was 2022 @ $11B. Most players of their size don't have that type of volatility. Your argument doesn't make much sense with imaginary numbers.
Gelsinger destroyed VMWare's focus and market prowess. His time at Intel looks unsurprisingly very similar.
Top line revenue and gross profit aren't my opinion.
You really think the BoD cares much about anything other than their profits? The board is the last place to look for an honest opinion. They're all friendlies of Gelsinger.
The problem is making those investments only guarantees cash outflow. Not that you will remain competitive. In the case of Intel it is rather obvious they are not going to be competitive again in the next 50 years for fab tech.
it's all internal revenue. IFS is charging Intel Products for making chips. 2023 was a down year bc of weaker cloud & IT demand. Outside of NVDA, semiconductor companies had a poor 2023.
"Revenue was $18.9 billion down $8.6 billion from 2022. Internal revenue was $18.0 billion, down $9.1 billion driven by lower intersegment volume. External revenue was $953 million, up $479 million from 2022, driven by higher packaging revenue"
Semiconductors are pretty cyclical in terms of demand, and 2023 should have been a bad year for them. This result may have been surprising in that the magnitude of the dip was worse than expected, but on some level this should have been priced in.
The other commenters have already more than addressed the internal revenue angle and the fact that Intel is just in a bottleneck any way you cut it. And that it’s a minimum of a year before we know if they make it.
But I will add that if they make it (basically that Gelsinger’s fundamentally sound-looking plan proves in fact to be fundamentally sound, and that they let a nerd run Intel long enough to find out), they’re going to be Kobe and Shaq in a bad mood.
A reborn Intel stepping into the room and seeing NVIDIA and AMD pillow fighting on top of the biggest pile of semiconductor bucks in the history of conductors or “semi-anything” will be like the “justice is real” scene from pick a Tarantino film.
Let’s just hope that this time around the they don’t intentionally hold back innovations in favour of slow-dripping a predictable fixed X% gain per year like they did 2014 to 2020 in the absence of competition.
Amazing how quickly Intel suddenly started leaping in performance/$ between releases once AMDs 5000 series came out.
I agree it's a weird world where Intel is our best hope for free, fair, and competitive markets. That's a really fucking weird world.
But Gelsinger looks at a cursory glance to be someone who knows how to build chips, has a pretty clear comeback story, and realizes that complacency in a monopoly is what made them uncompetitive in the end, and therefore desperate enough to put a qualified person in charge [1].
I think he might keep the pressure on even if they start winning.
> Let’s just hope that this time around the they don’t intentionally hold back innovations in favour of slow-dripping a predictable fixed X% gain per year like they did 2014 to 2020 in the absence of competition.
idk why you would cite this as a downside when this is specifically what the tech review community has been begging for with gpus. Having a gen do 40% and the next do 0–10% is grating to consumers and disruptive to the market. What consumers need is exactly this smoothing of generational progression and the removal of huge outliers in terms of generational step. Smooth the curve - which inherently will be much closer to that 15-20% progress generationally, given how slow the transitor tech is moving nowadays.
As Tim Hardwareunboxed says - if you don’t have a good uplift, don’t release the product. Intel made sure they had an uplift every year. You can’t change actual progress but it can be smooth.
> Why would I or anyone else care what the tech review community has to say about anything?
like, are you asking why thought leaders matter?
> What kind of idiot wants technological advances held back so they don’t get disappointed and the market doesn’t get disrupted?
Lots of people, have you never posted in an nvidia thread before?
tons of people would much rather RTX, DLSS, and AI never existed at all, because it means they have to upgrade their GTX 1080 Ti or RX 480 or whatever. A lot of people bought their first GPU since that launch and have never lived through a hard upgrade cycle.
recently it’s turned into this insane “if you can’t make cards like 3050 for $110 or get at least double the perf over 2 generations then don’t make a new generation at all” Neo-Luddism but it’s been brewing for a long time among the gaming public.
It's obviously dumb and regressive, literal modern luddism, but it's crazy to pretend that it's not an extremely visible and popular modern trend, either. People are mad the low end is stagnating and the high end is continuing to progress, and the natural human response is “if I don’t get to have a good time I don’t see why anyone else should either”. “Don’t launch a 5090 until there’s a good card for $110” is a very mainstream phenomenon nowadays.
Literally the drum the 2 biggest and most popular tech channels have been beating for a year now.
I don’t think a vocal minority under the sway of what it sounds like you’re describing as ill-intentioned propaganda (I mean, that’s a serious word but the shoe seems to fit) is a particularly good proxy for the overall preferences and welfare of everyone in the ecosystem.
To the extent that legitimate enthusiasts and early adopters and people trying to get the most out of their gear under resource constraints have a clear consensus, yeah, you want to be listening carefully. But that doesn’t really sound like what you’re describing. I hang out more on the “accelerated compute” parts of the “GPU Internet” than the “gaming” parts, but certainly over here in LocalLLaMA land, we want working, documented firmware, and cards that aren’t overkill on some things and weak AF on others (4090). That’s an effort to manipulate nosebleed margins by making the ostensible “prosumer/enthusiast” offering just kind of bad at everything (Quattro) except maybe CAD or Blender.
In gaming it seems like a cost-constrained public would be best served by credible offerings ranging from “the minimum to play modern stuff at a very attractive price” (I remember the 1060 being very well-received here), up to “Crysis at a million FPS” at Veblen, elaborate bespoke case, over clock with active cooling, and a sensible set of options along the way.
I don’t do enough gaming these days to have strong views, so please educate me if you’re an avid gamer and I’m full of shit, but both major gaming vendors seem to have good options for most budgets on mainstream platforms.
That’s just nothing to do with the 2T cap and 55.58% Net Profit Margins us llama.cpp folks are up in arms about.
> As Tim Hardwareunboxed says - if you don’t have a good uplift, don’t release the product. Intel made sure they had an uplift every year. You can’t change actual progress but it can be smooth.
I don't think that quote means what you think it does. If you don't have meaningful uplift every year… don't try to release a new generation every year.
I have a certain sympathy for all involved in your argument, but I don't think it's a good way to do business for either businesses and their executives, their shareholders, or ultimately the consumer and public.
The problem is short-term thinking around annual or even quarterly earnings reports and whatever drives that. I appreciate that a CEO or even a board may feel pressure up to and including losing their jobs if they leave literally anything on the table in the form of long-term investment, the "goodwill" (which is a real financial accounting line item that's sometimes a lot) of their customers, partners, etc. and the long-term sustainability of the business.
But it's that ability to deliver that magic combination of a product that people ideally love buying, but at least don't mind buying, a team that is healthy, well-compensated, and inclined to go the extra mile for the employer at rare need because the mutual priority of the wellbeing of the partnership goes both ways (this is massive, when you treat people well when they're in a bind, they'll go the extra mile when you're in a bind with little if any pressure, most of the time), and a consistent, low-beta (for the investor's risk appetite) return on investment for shareholders.
This is the kind of business that e.g. Warren Buffet at least talks about as the kind of business he learned to invest in. I'm not a Buffet scholar or watcher really, so I don't know how well he lives up to those ideals in general or especially recently, but it's certainly a pretty good way to describe the desired outcome of capitalism!
I've heard lots of arguments for why boards and CEOs and the like simply have to skin the sheep once rather than shear him many times (to borrow the old saw): but never a very convincing one. It's basically absurd in the kinds of dual-class share structures that are common in tech now: it's not only extant but common for the founder/CEO to be basically unencumbered by quarterly earnings reports, what's e.g. Mark Zuckerberg going to do? Fire himself?
And under this more "good for everyone over a long time horizon" model, you can just launch the next breakthrough card when it's done, and launch incremental improvements when you're out of inventory, or see a gap in your lineup, or whatever.
Fix corporate governance in what should be the relatively uncontroversial ways I and many others have described, and you fix this (among many, many other) set of perverse incentives.
I don't think anyone's under any illusions that people are going to upgrade more frequently, but the lack of predictability is pissing people off and miscalibrating expectations on what they "should" be getting every gen. Going 10%,40% is literally worse and more disruptive (in the bad way) to consumer psychology than if you'd just gone 20%,25%, because people will lock in on the most desirable possible combinations of perf gains and value offerings/prices.
the x70 tier launched at $349 in 2010 dollars and has bounced between $349 and $399 for every generation, but the one people remember is the singular time it was $329. And the x70 die has almost always fluttered around 250-300 mm2, except for the trailing-node era (when NVIDIA got cheap TSMC/samsung and chased price) where the 2060 was almost as big as a 1080 Ti (445mm2 vs 471mm2) etc. But people have never shut up about how the 4070 isn't a real x70 because it's only 300mm2 etc - they latch onto the one singular time it was bigger etc.
what you end up with in terms of node progress is similar - people tell themselves they deserve the gen when everything increases by 50%, at the best possible efficiency, with a bigger memory bus, but they also want the pricing from the value-focused generations. And they tell themselves that really things are actually advancing 50% per gen (because they see that sometimes/in some segments) and it's just evil corporations not letting them have it.
If, instead, you smooth the progression, so you get 20% then 25% over those same 2 gens, there is no bones made about the actual progress being made etc. Consumers have predictability and information, and they can thus make optimal, informed decisions about which generations they are going to upgrade (which I think everyone agrees at this point isn't "every single generation" for literally anyone anymore).
it's functionally no different in terms of progression, you are getting the same averaged gains year on year. It just means "don't launch the GTX 970 to be some super killer deal, because you have to top it again next year". And if next year you only get 10% over the absolute best sku in the lineup, you failed, your card is shit (or at least "disappointing").
you see this constantly with the 3060 Ti: the 4060 is deemed a failure because it doesn't leapfrog over the 3060 ti (the literal best-value card since RX 480 etc). And the 4060 Ti is a failure because it doesn't leave the 3060 Ti in the dust ("only" 3070 performance, with twice the VRAM). Etc etc. Stop doing that, just make normalized skus with consistent yearly improvement. That's what reviewers are asking for, and tbh they're not entirely wrong on that part (even though their overall performance/cost expectations are just completely bonkers). They're imagining it's going to be 50% a gen, not 20% a gen, but they're not wrong about how the public does need some consistency in their life (like any child) as far as generational gains.
The market simply doesn't reward you for making a 3060 Ti or whatever anymore, because that is taken for granted. You can make the same argument about the trailing-node strategy (for 20/30 series) as a whole, which legitimately bent the price curve downwards for consumers. But consumers will never reward you for a price increase you didn't make - in fact they will be extra mad when you finally bounce back to the "normal" cost curve, because now it feels like an extra big increase all at once.
Personally yeah, I'd rather have the god-tier value card and just understand that the successor is going to be feature/performance/efficiency focused and not value-focused. As high-information consumers we can make those decisions and claw back some consumer surplus. But that's not where the public is. And people have been hopping mad about it for years now, this is not an unusual sentiment.
Intel basically exemplified this mindset perfectly: every generation you got X%, and they aligned X% so that they'd get a similar number on every single generation. Going (eg) Sandy Bridge to Skylake is much "lumpier". And people dump on that era, but they also are (functionally) clamoring for that same consistency in the GPU market. "You get 15% a year whether you need it or not".
Right now they don't know what to expect, so they can imagine whatever they want, which only sets things up for disappointment.
I'm not sure that you and I actually disagree about all that much, and might be just arguing cases in different adjacent debates.
The realities of the modern hyper-sophisticated semiconductor fabrication business is that the chips aren't consistent with event a single logical "run" (or whatever the modern term of art is), and so you're binning and blowing fuses and all kinds of stuff to get even the same SKU in the ballpark as other cards with the same SKU (throw in MSI vs whoever, it just gets harder), and so clearly there is a mutually beneficial amount of "smoothing" that's more than zero (impossible) and less than anti-competitive (56% profit margins and a CEO saying "this is a gold mine" on camera).
And as someone who is clearly more current on the nuances of the gamer-oriented lineup, I'm inclined to take your word for it that gamers want a little more "smoothing" than ML people without a blank check (though I don't think I can endorse referring to the public as needing the same guidance as any other child, that's not me trying to chastise you, it's me noting that underestimating the public has burned a lot of smart people some rather, uh, recently).
It is complicated, and there are hard realities of the physics and tolerances of the manufacturing that complicate the situation, and there are hard questions in what constitutes a "major version" and what constitutes a "refresh" or a "rev". So to the extent you're saying that the vendors need some rope to give their customers something that isn't a casino by binning and labeling releases with an eye to some kind of non-ridiculous trendline in capability that isn't coming or whatever, yeah sure. I don't know enough about the pressures of a middle manager working on that stuff to offer hard policy prescriptions about it. I do know about the pressures of an upper-middle technology executive faces with this whole other bag of constraints over and above the already hard job about the "the Street is going to kill us if X isn't Y by Z for Q1": that's a problem I think you'd agree no one would actively solicit as part of their job description all else equal.
There's this different basket of stuff that's ethically and practically anti-competitive, anti-market, and at least used to be illegal most of the time (jurisprudence on how narrowly to interpret the anti-trust laws we do have has become an Olympic sport in narrowing their scope). That's the drivers I simply don't believe to be unfixable in a year or two, that's an unreasonably wide gap between a 2k card you can run PyTorch on and a 50-80k card (or whatever the real price is, I'm not sure anyone even knows in a public way) that's basically the next step up, that's the 4090 spending die area and TDP on vector units it doesn't have the GDDR6 to keep fed half the time. And finally, it's accepting margins that let you price cards at MSRPs you're willing to publish and meeting the demand for them at your stated price. This isn't an act of god chip shortage (in general, there are instances of that, particularly during COVID): I hate to pick on Meta here because the reason I'm using them as an example is that they're reasonably transparent about this (so keep it up guys, that's awesome), but IIRC their last publicly known block of H100 (Hopper v1 for lack of a better term) was 150k cards in one transaction, with a target of something like 700k "A100 equivalent compute units" by the end of 2024. I'm not surprised Lisa Su and Jenson Huang are inclined to do as much of that business as possible, but again, market failures left and right.
Maybe gaming and AI need to be different companies to get an efficient market, maybe export controls need to be fiddled with further, maybe a lot of things. Gamers should be able to get cards that play relevant titles at MSRP, startups and independent researchers should be able to get cards that work and don't require a blank check, and big AI labs should be able to get a reasonably level playing field in terms of access to the big iron. But a monopsony quasi-cartel of frenemies competing in some ways but with their fates intertwined in other, cornering the market on process nodes and still getting gouged by vendors having a "spirited debate" and not a UFC match. That's not getting us the benevolent global outcome of the locally selfish invisible hand. That's where sparing but well-refereed interventions in failed markets is what makes capitalism sustainable after the sugar rush wears off of having industrialized in the first place.
That's the stuff I think Intel is going to go on a rampage around if Gelsinger pulls off the turnaround.
I seriously wish Intel would pivot and stop being ultra worried about things like their atom or Celeron or whatever it's name is being competing with their own stuff and memory capping it. Lots of room for a really low power decently cored memory heavy CPU but they've always capped features and things like ecc memory away to the top skus which is really annoying.
Also not talked about but isn't Samsung and tsmcs fabs booked clear out the next 24 months?
Most cpu manufacturers do this. Binning, changing product numbers or lines around, and marketing tactics.
@GamersNexus on YouTube occasionally does stories on this. Recently they had a good intel slide deck trying to disparage AMD and muddying the water on their own product lines.
>Recently they had a good intel slide deck trying to disparage AMD and muddying the water on their own product lines.
To be fair AMD does this as well. A lot of modern CPUs series are older gen ZEN cores build on older TSMC nodes, rebadged as the newer major series with only the minor number identifying that it's older gen design[1], but AMD does not disclose this in their own comparison tables.[2]
They've been doing this since at least the Ryzen 5000 series(used to own one) so I'm not sure why everyone is so quick to point fingers at Intel but gives AMD a pass. It's not like one is more righteous than the other, they're both major for-profit corporations using marketing confusion to decisive buyers.
To be fair Nvidia Blackwell is basically the same thing, a doubled up Hopper chip that shares a few things between the two for slight efficiency gains.
Momentum of batch buys is running out with new buys coming at a cost? If your chips aren’t performance competitive in processing or energy, you have to discount until you hit cost parity.
I´m an outsider, but most demand in the last months have been GPUs and AI accelerators. And Intel foundries are not making them yet, only the ARC that are not very strong.
There's been demand for LLM accelerators since late 2022, but so far only the Groq cards have materialized and those aren't really viable for anyone but high end enterprise datacenter use.
Literally nobody's making these things yet, Nvidia gaming GPUs remain the best price to performance option lol. Intel Arc only does compute work through Vulkan which is not exactly comparatively fast yet, so despite being well priced for the VRAM they offer they're not seriously considered.
They aren’t making the chips that are in demand. Neither does AMD for that matter, but AMD does not make its own silicon. Whether that’s prudent, we shall see.
I think if we dissected chip demand by company we would see that TSMC/Nvidia are over represented. Chip demand as a whole can grow while demand for Intel or AMD chips shrink.
All Intel did was just report a new operating structure and re-allocate revenue and costs to that segment. So essentially expenses that went into manufacturing before are now just thrown under IFS. It's just a bunch of financial engineering.
It's not just financial engineering. It's a key step towards their goal of splitting the foundry and design businesses apart. They already manufacture chips they design on other people's foundries, and manufacture chips other people designed on their foundries. That's step one. Step two is to make both sides independently profitable.
At some point Intel might conceivably split into two fully separate companies.
Not a dumb question, but because they are so expensive to build, fabs have a high upfront cost before they can start to turn a profit. Oddly, the older a fab is, typically the more money it will make you.
It's also a big part of the reason as to why so many fab companies have died out. Building a new fab with the latest tech is extremely expensive. If you have a couple of big swings and misses, you can easily burn a tremendous amount of your capital.
Intel lost the high end. They don't seem to know how to make CPUs or GPUs that outperform their competition. But they do have fabs and could in principle make chips for other companies.
What they could do is use the equipment they understand really well to make chips for the military. 45nm or 14nm or whatever. Or for automotive. Loads of stuff is useful without being on the bleeding edge.
Instead, they're talking about needing heavy investment in new machines and blaming lack of that for their problems. It's hard to read that as anything other than they're trying to retake the high end crown.
Perhaps the motivating defence angle is AI. The US panicking that TSMC makes better GPUs and tech built on that will end their empire. That's something Intel has decades of failing to execute on.
This just isn't how the semiconductor industry works. You invest in leading edge and then you try to find demand for the older stuff. There's tiny margins for the older stuff- because the people choosing to use it are literally choosing it for cost reasons and there's no competitive advantage - they can shop around to any of the fabs. The capital investment is all done and you're just trying to make the most of the sunk costs.
It would also be a choice to basically not be a company in 10 years. The low end stuff is the stuff that was bleeding edge 10-15 years ago. If you decide to focus on that older stuff and don't chase after the bleeding edge then in 10-15 years you simply won't have anything competitive at any level, your company will basically be done - your 45nm process will be competing with TSMC's 16nm (which will now be old tech) and you'll just slowly move up and up out of competitive markets.
Intel's ARC chips are actually pretty good, and would be pretty competitive with NVIDIA for the gaming market if not for the drivers...
NVIDIA has three decades of DirectX driver support for many, many old games. Intel's ARC GPUS are infamous for rendering big magenta materials for many games, particularly old ones.
ARC however has absolutely nothing to do with Intel Foundries however, as the Intel ARC GPUs are fabbed by TSMC!
Intel will certainly get to the point where they fab great GPUs, but its actually a question of whether the software will catch up, not the hardware!
No. Auto pays a 2-10x premium. They have higher temperature demands, high safety demands, and lock in a 10 year design cycle. Automotive is resistant to changing suppliers due to development environment stability needs which allows even more premiums.
The process misstep was an error. Sticking with monolithic dies well past their EoL was an error. Missing the low power chips for mobile was an error. Everything they've done in discrete GPUs was an error.
Intel make really great WiFi chips and C2D and Nehalem were spectacular. Since then they ran a fantastic profit margin.
It's not remotely obvious that they're still great at designing chips.
Didn't Intel exit the HBM market a few years back and now it's soaring?
Also Intel is being a dick about their wifi on AMD systems when that's one of their markets honestly - everyone else's wifi chipsets sucks honestly at least under windows so they should really consider partnering with their enemy instead of trying to fight them. One of the coolest Intel products was that NUC with Intel CPU, igpu, HBM and AMD vega.
Intel was never in the HBM market - you might be thinking of their 3D XPoint/Optane memory developed in collaboration with Micron and discontinued in 2022.
> Sticking with monolithic dies well past their EoL was an error.
Recalling how Intel tried to argue their multi-chip approach was better than AMDs single chip approach back in the early multi-core days, I find this one so hilarious. It's like the tables have turned.
Sticking with monolithic dies well past their EoL was an error.
I'm going to disagree with this one. Monolithic dies are more energy efficient and usually faster. I think Intel, Nvidia, and Broadcom were right to stay monolithic as long as possible. AMD has saved a lot of money by using chiplets and that strategy worked for them as the underdog but companies with better reputation and high margins can afford monolithic.
Note: I'm a left leaning liberal, who thinks that our system is fucked when it comes to taxes and how we distribute money...
Spend about a day catching up on where the Ukraine conflict is.
Up to speed: good! You will have noted that not having domestic chip production has fucked Russia, and the restrictions on china are to inhibit pass through and limit what they too can do.
You dont give a potential advisory (china) leverage (Taiwan as the sole place chips are made).
Even if Russia could make their own chips invading Ukraine would still be a very bad idea and has completely dragged down Russian future prospects. They have isolated themselves on the world stage.
Russia has resorted to selling oil to North Korea because they no one else will buy it. That is isolation.
Russia is also working to remove Taliban from list of terrorist states.
These are Russia's "friends".
But isolation here also applies in political sense. There is no political support for Russia's invasion from the rest of the world. Even China has not approved the invasion.
>But isolation here also applies in political sense. There is no political support for Russia's invasion from the rest of the world. Even China has not approved the invasion.
True. But there is support to replace the unipolar with multipolar world. So no one of the big and medium players outside the first world will be too sad to see that the massive amounts of face USA and EU have put on the line in Ukraine are lost.
While the first link in my comment is the 4th result in the related Google search, the first one is this below horsesh*t that literally falsifies what the Economist's Intelligence unit and NYT and various others now openly state:
I understand the argument for national security. Having said that...where's the accountability if what they build is terrible? Where's the return to the investor (the public) if they succeed?
I'm not saying the government should just take over businesses...but if the government is going to essentially be an investor in a business, it should be at least as rigorous as a private investor would be.
Sometime I don’t understand HN. Such an informative article gets no traffic, yet people jump to a sensational article to discuss the very topic covered in details over there.
Presumably because it's cheaper to pay intel to set up a foundry in the US and allow them to keep all the money that results from it, than it is to pay them to set up a foundry and demand a % of the revenue.
I would love to hear any explanation for why they should get a penny of taxpayer money after spending >$150B in stock buybacks. They spent their money on enriching the rich.
Because the money isn’t a bailout, it’s to incentivize intel to do a financially suboptimal thing that the US finds strategically desirable.
If building and running the fab in Arizona versus Ireland is going to cost intel $8b more, the options are do nothing and don’t get the fab, or pay the difference. Even if they are doing stock buybacks it doesn’t mean you can force intel’s shareholders to take a $8b loss without compensation - that’s a clear takings clause violation.
They probably also spend the money on pay rises. Does that also bother you?
When investors buy shares, they are in effect giving money to the company. It might be indirect if they purchase from some other holder of the share, but ultimately those shares were issued by the company and the initial purchasers gave the company that money.
Investors do this because they want a return. Either dividends or price growth or both. If there's no prospects for growth they will pull out their money and this will seriously hurt Intel, creating a downward spiral.
The US Government is subsidizing Intel because they don't want it to die. An Intel that has crushed its investors will end up on permanent subsidy life support because nobody will trust it with their money again. An Intel that uses some of those subsidies to keep investors on-side will retain their trust, and give it a pathway off government support and back onto raising funds from the markets.
So whilst this sort of thing makes the left very unhappy (often because they don't think of themselves as holding equities even when they do via pensions/insurance funds), it does make sense over the long run. Investors aren't an ATM that can just be raided at will. They're a coalition that can break.
I said I would like a better tax policy (that would address your issue) but that doesn't change the need.
You may not like it, I am not fond of it, but we are now Rome. THE US's military might is the same as Romes, the legion (the navy being the 2nd largest Air Force on the globe says it all). You know what happened when the legion got weak? The barbarians came in and ate Rome. It wasnt fast, and it wasnt pretty.
Chips made in Taiwan is a matter of national security, or the lack thereof.
Chips made literally anywhere else, including China, are fine; China might blow up Taiwan, but Taiwan can't reciprocate and the west will not fight China.[1]
I say that as someone who likes Intel, wants to see Made in USA come back, and abhors Chinese geopolitical ambitions.
The bottom line that shall not be stated is nobody wants to deal with Taiwan and everyone is happy dealing with China.
The US military can't do a single thing about a bunch of misguided cavemen lobbing missiles into and stealing freighters from the neighbouring shipping lanes. For that matter, the US military hasn't won a war since WW2.
The numbers might look impressive, but in practice it's vapid. Obama even went on record that the US is retiring from policing the world.
>>> The bottom line that shall not be stated is nobody wants to deal with Taiwan and everyone is happy dealing with China.
What? No one wants to deal with china. Everyone is moving their manufacturing OUT as fast as they can. China in its desperate throws to remain relevant in the face of a massive population decline is desperate (the same way Russia is in the face of their population decline)...
No one wanted to sell the tech to china to make modern chips before all the wars and saber rattling of belt and road (facing). And china lacks the will, talent and now access to catch up.
ASee the Ukraine war and Russia. To see what "no chips" looks like. modern optics on tanks are now outside what they can do, never mind glass cockpits of 4th and 5th gen fighters, never mind gps and smart mentions, never mind drones...
>> Chips made literally anywhere else, including China, are fine; China might blow up Taiwan, but Taiwan can't reciprocate and the west will not fight China.[1]
China isnt stupid, they keep pouring money into missles in the hopes of being able to do exactly what you're saying. The problem is that without sea lanes then china runs out of oil, steel and food inside 6 months. If you pay attention to recent conflicts you will note that the patriot missile system has been shooting down the best Russia has to offer. You will note that the anti missile ships off the coast of Yemen are using the same system and missles. That the brits and French are in on the game. You will note that all the intel on chinas missile force (farce) has leaked everywhere... China wont invade shit.
Agree domestic semi is crucial national security. For disagreements:
>No one wants to deal with china
Everyone wants to deal with PRC so much that US literally had to impose sanctions and export controls to make it illegal. And then extra territorialize bans/frictions on partners who didn't want to by leaning on American origin tech %s until they had no choice.
> will, talent
PRC is minting more STEM/highskilled than OECD combined, and only actor projected to build enough talent to actually have fully indigenous semi supply chain while everyone else is projected to have shortages.
> what "no chips" looks like
Except military hardware use older harden nodes that PRC has domestic supply for, they're not running out of chips for military applications.
> patriot missile system
If you actually pay attention, you'll note patriot in UKR got penetrated by 6 shit tier kinzhals @mach3 and now multiple launchers are getting wiped from 80s hardware right after RU unfucked their primitive killchain, tier1 flight IIA/III DDGs in redsea are letting slip through enough uncoordinated old gen AShMs attacks that US can't guarantee redsea shipping. Old ghetto hardware that travels at a few maches or even subsonic, launched not in coordinated fashion, with even shittier kill chain than RU. Yet state of the art US missile defense interception isn't performant enough against these ordnances not used at scale and likely wholely inadequate for modern AShMs that travel much faster.
You'll note last round of PRC rocket force leaks are retarded, rocket fuel for hotpot? Conveniently leaked around TW election to send msg that PRC won't "invade shit". It's called motivated propaganda. Meanwhile actual Teixeira leak - the stuff US prosecutes because it's actually credibly damaging - allege US clocked PRC DF27 tests at mach 8+, and many signs point to PRC pursing global prompt strike with rocketry = no one's going to blockade PRC without their own energy infra getting blown up. Including US who goes from US to Yemen without a few hundred oil refineries and LNG plants. Or domestic semis, or data centres, or F35 factories.
TLDR western ABM performed about expected relative to threats (engaging slow targets up mach3), if anything less than idea since they're letting enough slip through that their performance has essentially validates PRC's focus on high end rocket force. You think this is US showing sabers work, but it's exposing they don't work remotely well enough against threats they were designed for, AKA, what PRC rocketry was specifically designed to defeat. Now they're thinking instead of sending salvo of 100 mach 6+ AShMs against a carrier group maybe only need to send 50. Or 300 cheaper AShMs/TLAMs. What recent conflicts has taught us is scale>wunderwaffe, lesson learned from previous wars.
>> Everyone wants to deal with PRC so much that US literally had to impose sanctions
Everyone is happy to sell as much as they can to whomever they can. The number of western companies fleeing the PRC says that every one is excited to keep betting on china?
>> PRC is minting more STEM/highskilled than OECD combined, and only actor projected to build enough talent to actually have fully indigenous semi supply chain while everyone else is projected to have shortages.
What SEMI supply chain? Where are they getting hardware for litho? Lenses? They aren't building that, they have no idea how.
SMIC is state backed, they got one run of 7nm chips to huwei for the latest phone and fizzled....
As for skilled labor ... where? China stopped publishing youth employment, all those fresh grads dont have JOBS and most of their tech base is in shambles. Any one with any skill went outside the PRC for work and are kept in restricted roles for fear of them stealing everything in sight (tech wise).
>> Except military hardware use older harden nodes that PRC has domestic supply for, they're not running out of chips for military applications.
None of this particularly new tech. The problem is that just because you have the machines and undestand how it works doesn't mean you can operationalize it. The restrictions have been impactful across the board and "catching up" is not just "doing it". These are insanely complex processes and getting them to work, and then at scale is a multi year process....
>> If you actually pay attention, you'll note patriot in UKR got penetrated by 6 shit tier kinzhal
Spotted by a drone, at the front line (to take out Russian jets) and caught with its pants down. This is not the systems failure your portraying this is an intel win for the Russians.
As for the Red Sea, Live fire exercises are important, you can train all you want but doing it and proving it are another matter. The patriot tech for this sort of interceptions has had no live tests at sea till now...
Lastly Russia has tried on several occasions to take out patriot batteries. They have missles that supposedly have the ability to directly engage them then why have they not used them?
Every one tests these technologies... putting them into production is another matter...
>>> no one's going to blockade PRC without their own energy infra getting blown up. Including US who goes from US to Yemen without a few hundred oil refineries and LNG plants. Or domestic semis, or data centres, or F35 factories.
At the point that the PRC striking a US target in response to a blocked ends in a nuclear exchange... Your making the point that the PRC wont try for tawain pretty clear here with this argument as there is no outcome that doesn't end in a global catastrophe.
>> anything less than idea since they're letting enough slip through that their performance has essentially validates
Quick head count: Ukraine a massive country has enough patriots to basically defend a bit of critical infrastructure (5)... Taiwan a country of microscopic size in comparison has 14 at last count + their own home grown defense system, and the ability to fire retaliatory strikes at china (who has too massive of a cost line to deploy the same level of protection even if they had it).
>> PRC's focus on high end rocket force.
If it performs like everything else does in the PRC then they would have been better off buying bottle rockets.
>> You think this is US showing sabers work, but it's exposing they don't work remotely well enough against threats they were designed for, AKA, what PRC rocketry was specifically designed to defeat.
If Russia or china had a missile that could have taken out an active patriot battery it would have been used. Why? Because every country in the world would have bought them. The windfall for a cash strapped Russia or chines defense industry would have been huge.
TLDR: Active patriots have yet to meet a target that they cant take out. Chips, and chip manufacturing technology continues to hurt china and Russia. The decline of China has been sharp and painful for them.
Western FDI constituted negligible part of PRC domestic investment as % of GDP for years. Like low single digit percentage of total investment. It largely doesn't matter. Of course western companies they're not excited after US made it geopolitically difficult to invest in PRC, because without literally making it legally difficult to do business in PRC, companies wouldn't have China+1ed/fled. That's not PRC not being attractive, that's US forcing their companies to take the L for geopolitics, which is fair. But questionable if it hurts PRC more or US.
>SEMI supply chain
They're chaining together entirely indigenous semi industry. 12/24nm under validation from SMEE. Every component has domestic ventures to replicate. A few years out, expensive/wasteful, might run late, but will eventually get there as with all PRC industiral policy that gets the manhattan project fast track. Pretty much the only country with entire semi supply chain development on horizon and talent base to man it. Huawei increased orders for 7nm phones, eating away at Apple shares. Opposite of fizzle.
>skilled labor/youth unemployment
PRC hard tech sectors have exploded in employment and compensation. Broad unemployhment 5% which means new graduates get jobs eventually. There's new youth unemployment stats to discount students in tertiary like west, it's fine at ~20%. Even at 50% that includes youth in fulltime school, that's multiple times new talent than US. Meanwhile you have sea turtles / skilled talent from abroad going back to PRC and they're retaining more high end talent than ever due to domestic opportunities. Youth unemployment breakdown also biased towards low skilled. High skill are generally fine.
>operationalize it.
PRC's has operationalized domestic MIC for 10+ years now. This is old boring news talked endlessly by US thinktanks for years. You're imagining they're learning to crawl when they're near sprint. That's why they're recognized as the pacing peer threat. No one credible thinks PRC isn't operationalizing modern capabilities at scale.
> Spotted by a drone
This this was reference to attack last year that damaged active patriot batteries. RU haven't used them more because they don't have good ISR/fast kill chains for how mobile patriots are, can't build high end missiles fast enough unlike west/PRC. So not worth commitiment. They have hardware, but limited, theres like 100 Kinzhals total and they used 20?. TLDR is they lack capability to coordinate strikes. PRC tests more high end missiles a year than RU can produce. This looks extra bad for patriots because even RU, as incompetent as they are at ISR still managed to penetrate patriot when they dedicated some highend (really medium end) hardware to the task.
> Live tests
USS Mason shot down shit tier Yemen cruise missiles (like old PRC 80s tech) in 2016, shooting down similar missiles and drones doesn't bring new learning except from PLA sigint in region gathering data on US weapon systems since they weren't in Djibouti at the time. Really shotting down primitive single digit clusters of cruise missiles is not even a useful live stress test.
>NASA_X-43
Many reports on PLA rocketforce expansion, they're making them at scale, hitting moving targets at sea. Leaks verify speed/performance. Putting up 100s of ISR for kill chain. What hasn't been verfied is US building capabilites to defend mach 5+ threats, especially at scale.
>US target in response to a blocked ends in a nuclear exchange
At the point US tries to blockade PRC ends in a nuclear exchange. See how that works. Point is PRC already deters US from even attempting blockade by making CONUS as vunerable as PRC is to SLOC blockades. SLOC blockade math works on rationale that US can enforce unilateral/existential damage to PRC without damage to herself. Now US can't without basically being in same energy deprived state as PRC. Hence US likely won't. Only real deterrence catchup is nuclear parity.
>Quick head
Quick headcount, PRC has magnitude more missile inventory and missile defense relative to RU:UKR, focused on island 16x smaller with even less perimeter to preposition defenses. The geography and math here isn't difficult, PRC can trivially overwhelm launchers TW has, assuming they even work when sensors are gone in first strike - PRC can hit anywhere in TW in 7 minutes.
> like everything else does
I mean they can land on mars first go, so I think it'll land on a carrier fine.
>windfall
PRC doesn't export high end, and RU not in position to export in middle of war with their limited industrial base.
TLDR patriots and SM2 working as intended, with expected limitations. Which is to say overwhelmable by power that can coordinate more than a handful of high performance missiles at a time. Which is actually pretty high ask but certainly includes PRC. PRC semi doing fine, catching up faster than anticipated, remember how they were suppose to be capped at 28nm. Meanwhile, they're still climbing up on the supply chain ladder, displacing more and more high tech from west, growing in S&T/innovation indexes. They crippled their R/E with 3RL and soft tech on purpose and still chugging along. Eitherway this is departing from the Original topic, so I'm going to hang hat here.
>> Western FDI constituted negligible part of PRC domestic investment as % of GDP... US made it geopolitically difficult to invest in PRC
Foxconn leaving at apples behest (and the Foxconn owner getting vocal with the prc in a rare move). The banking exoudus from Hong Kong (now deep into its 2nd year). And the impact to the GDP... its a literal house of cards with the failure of ever grand with country garden hot on its heels...
Meanwhile the retail apocalypse has hit china. High end luxury brands. Out. Grocery chains, shutting down.
Finally the iron bowl, the local government workers are getting pay cut. (because land sales are a shitty source of funding)...
Post pandemic there was going to be a realignment of global supply chains to be multi source and closer to home. Chinas covid zero policy just accelerated a decisions already made.
>> Every component has domestic ventures to replicate.
These have been facing up for decades. Without EUV they have no path forward. There is no EUV without Zeiss and ASML. Meanwhile unless your TSMC with their knowledge your going to have dismal yeilds.
>>PRC hard tech sectors have exploded in employment and compensation. Broad unemployhment 5% which means new graduates get jobs eventually. There's new youth unemployment stats to discount students in tertiary like west, it's fine at ~20%.
The decline of Chinese tech from stem to stern has been in process since 2000. Again the pandemic was a blast of orders for them but it was the last one.
> Meanwhile you have sea turtles / skilled talent from abroad going back to PRC
This has always been the case, they go home with any trade secrets that aren't bolted to the floor. https://www.smics.com/en/site/news_read/4334 << there's SMIC stealing its "progress" from TSMC.
> PRC's has operationalized domestic MIC for 10+ years now.
This is tech that was hardly new and sexy when the first iPhone came out, china JUST joined this party. This isnt pzrticlulary hard silicon issue but it is process intensive... Again, based on how they did it likely stolen IP.
>> This this was reference to attack last year that damaged active patriot batteries. RU haven't used them more because they don't have good ISR/fast kill chains for how mobile patriots are
The damage was minor enough to be repaired locally, the battery was not rendered inactive by this attack (read they did not destroy it). Per later reports the damage was not from one of the missles fired at it rather it was from an Iranian drone shot down in a separate intercept (oops). The reason that Russia does not have good intel on the patriots is NOT their mobility, it's their reliance on human intel (never change Russia).
Do not that even AlJazer wont call the Russian attacks "effective" and hints at US shoot downs (that no one is gonna confirm publicly, because its bad politics).
This is the only credible source for their expansion into this territory... And again most of the leadership is gone, and did they or didn't they have to buy missile tech from the Russians. Meanwhile the bigger threat is in theater ICBM's with MREV's ... https://www.popularmechanics.com/military/weapons/a34702702/...
We have a lot of carriers, and have managed to keep pouring money into tech that will keep them safe (as they are large, slow, expensive targets).
> ends in a nuclear exchange
China invades tawian, china gets the blockade. People starve or china does something dumb that ends in nukes. It's a zero sum game for them and they know it. Same bluster from them since Nixon...
>> PRC has magnitude more missile inventory and missile defense
To be clear: the US navy (just the navy) has more Tomahawk's than all of the theater (medium and larger) missiles of the rocket force. Chinas combine anti missile systems (unproven) are enough to stop an attack from Taiwan (who has about half the missles of chinas rocket force). At a 50% shoot down rate, the rocket force would have to commit a significant amount of its inventory to overwhelm that defense. Thats a very precarious position to put yourself in if you expect the US to respond....
>> PRC semi doing fine, catching up faster than anticipated, remember how they were suppose to be capped at 28nm.
Nothing of the sort, see above, see the current state of china semi... They have had DUV for a long time (5 ish years from asml). They used it for 14nm, to get 7nm and can get down to 5nm.... Can isnt volume, can isnt yeild.
>> Meanwhile, they're still climbing up on the supply chain ladder, displacing more and more high tech from west, growing in S&T/innovation indexes.
The only people making this claim are china state propaganda. This is right out of the little pinks hand book. There is over a decade of court cases (both civil and criminal) for theft of intellectual property... Copying by theft isnt innovation. Meanwhile they can barely make the filters for 5g (see above, mems, 20 year old tech at this point, brand new in china). That isnt catching up, that's isnt even keeping pace... it's throwing money and espionage at a problem. China cant make camera lenses, they cant clone DU, how are they going to get to EUV (all for litho for chips).
RE impact is 2-3%, which is what you get when RE demand gets crushed from 15T to 12T in a 100T (RMB) economy. That's it, the sum of PRC RE house of card "collapse" inefficiency is 3%, aka how PRC growth goes from wreckless 8% to fine/good 5%. PRC luxury spending up 12% last year vs 7% for US. Grocery chains are fine, some Carrefour leaving, but as if PRC relies on western grocery chains. Disposable income up 6% nominal, while inflation negligible. Supply chain was reorienting to PRC+1 pre covid, meanwhile PRC during covid exports increased 1T in 3 years, more than previous 10 years combined. Held on to most of gains / wasn't hit as hard by export decline vs others in last few years. Q1 exports up, PMI positive. It's fine. Closer to home supplies chains, i.e. Mexico has large % of transtripped PRC exports to evade tarrifs.
>decades
No really, big fund semi push only in last 10 years, build uncoordinated base, but lots of players for entire chain. 7nm yields are fine according to forensics extrapolating from quality of production. Production only one component of price, PRC not paying IP fees for domestic chips to west/US = it's economical to produce domestically even with lower yield. EUV/SSMB 5-10 years away. Questionable if <5nm has any major benefits vs PRC brute force computing and making up operation costs on the power end with massive renewable rollouts.
>youth, tech
You need to follow space more closely if you don't know about new youth unemployment data. Ditto with tech sector, especially hard tech hiring and compensation.
>tech decline
Every S&T, R&D, innovation index/ranking and reality of PRC moving up supply chain in just about every sector in last 10 years contradicts this. PRC went from like 10m STEM in 2000s/ high skill talent to 40m matching US, most in the last 10 years, and that's translating to PRC moving up tech in every sector, not just software which they deliberately slapped so people moved to more strategic sectors. Next 20 years they're adding another 40m+ STEM/high skill to workforce. There's a reason ppl freaking out about PRC EVs, green energy, mature chips - because PRC is encroaching on all the encumbant high tech previously dominated by west, and that's just PRC fucking around for a few years after spamming STEM in from academic reforms ~15 years ago.
>always been the case
PRC talent from west have never returned back to PRC on this scale before. Amount of PRC talent retention and inability for west to brain drain due to domestic opportunities have never been this high.
>stolen IP
Who cares? It's MIC, and huawei leads in 5g / telco which is what it focuses on. Working on filters was contingency post sanctions and they did it in a few years. It's a good example of how fast PRC catches in critical sectors it wants to - timeline for sanctions against PRC is more limited than most hoped - they indigenize fast. This is like people wanking about PRC finally making ballpoint pen tip (which US also can't make), when that was cover for LKQ bemoaning state didn't focus enough on precision manufacturing for advanced munitions. A year later state developed that capability.
>AJ/patriots
Quoting US officials, as if that's AJ position. It's he said she said, of course Kirby wants to play down damage, because in your words, admitting is patriots can get overwhelmed by mach3 duct taped RU missiles is "bad politics". Mobility and ISR goes hand in hand, the point of mobility is to be faster than opponent ISR, to pack up before they can plan mission. Hence RU intel failure + patriot mobility is linked. When RU ISR picked up, patriots started getting hit, of course in transit, pants down because ISR tells you thats best time to strike.
>locked up
You should look up all the CSR reports and think tank analysts who track PRC missile developments. Or former Pac fleet deputy chief Fanell who confirmed tandem PRC hypersonics struck moving ships at sea - they work fine. Like this isn't hard technology for country that has space program.
>we did we can
We already kenw that. Carney used boring old SM2, same as Mason in 2016. We already know SM2 can hit shit tier cruise missiles. Intercepting TLAMs and slower ballistics is not new. What we learned is it's not enough to deter Houthis whose still hitting ships despite USN protection. This isn't patriot test - Lockheed pitched to navy, but not integrated - no one's seen them because they don't exist. Deploying more THAAD/Patriot in MENA is not because they're peak, it's just what US have. BTW same system got overwhelmed by Houthis when they hit Saudi refineries, but we'll just blame Saudi operators because that's convenient.
>rocketforce leadership gone
For corruption, which in PRC =/= RU. Their corruption is graft of contracts not replacing rocket fuel with water or selling tires. The former requires a signiture and gets you a few years in jail, the latter bullet to the head, and for 1/100 the money. If anything PRC corruption is associated with overcapacity, grafting building more hardware. PRC has grown way past RU tech in everything except maybe sub quieting, but anything electronic/higher tech they're better, see PRC modernized flankers. Look up latest airforce university report on rate of PLA rocket force modernization.
>ICMB/MIRV
IMO Glide vehicles that defeats interceptors is more dangerous to carriers, and direction US missile defense/advanced interceptor programs are targetting. MIRVs is fine, but mostly for nukes. Conventional strikes against navy targets going to be just saturating limited local defense / magazine depth. US pours a lot into ship defense, but no indication they work against high end missile threats right now. We're up to FTM48 which hit a couple short range ballistics, but still controlled conditions, no countermeasures etc. Ultimately it doesn't matter how much US spends, amount of VLS cells in a fleet is known/finte and can be weaponeered around to saturate.
>China gets the blockade
China runs the blockade, dares US to do something, with next run on escalation ladder hitting US critical infra. US loses all oil refining capabilities and downstream inputs like fertilizer. US starves by doing somethign stupid like trying to blockade PRC. It's a zero sum game for everyone. Diffrence is PRC actually fought with every NPT nuclear state over issues less important than TW. Including US in Korea when PRC wasn't nuclear power herself. I mean if that's bluster... PRC knows US won't eat nukes for TW... and US won't risk losing her own energy supplies and not eat for TW. Hence blockade won't likely happen because US interest caluclus less suicidal vs adversaries who can strike CONUS at scale. This isn't 2010s when that argument had merit.
>navy has more Tomahawk
No USN has more coutnable VLS, people think US Navy has more Tomahawks, which btw is close to nothing even if 11k VLS cells were loaded for that mission. PRC can produce that in 11 days. 20 days and they have enough to overwhelm entire US security architecture in 1st island chain, most of which will be in range (nearly all of SKR/JP/most of PH from mainland). Ultimately 11k doesn't buy you very far when estimates for just taking out SCS island bases (missile sponges) run into 500+ for just temperoary degrading runways, let alone mainland targets. The fundmental inbalance in theatre is PRC makes more stuff, can hit more stuff, has more stuff to hit and therefore more difficult to proportionally degrade, even if you throw entire USN + get agile basing in region. Ultimately we don't know what we don't know, but we know relative to industrial output, USN can't throw enough at the problem. Like it took 5 carriers + good regional basing 3 weeks to degrade Iraq, scale that up to PRC and that's a 5 year campaign assuming nothing PRC makes work. But if they work advertised (not hard), then 11 carrier groups and 11k VLS cells is defeatable.
Also PRC does their own FTMs, they intercept cruise missiles (easy, known party trick) and balistics themselves. Reality is no one knows PRC missile stocks, because unlike ship/planes they can't be easily counted via imagery. All we know is they just showed an automated factory that can make 1k cruise/TLAM components A DAY. It's trivially easy for PRC to stockpile enough that interception is meaningless due to saturation. And for all we know, they have, because munitions factories aren't sitting idle. Having a 100k cruise missiles lying around is trivial cost. They don't even have to be advanced at that point since quantity > quality of interception. The reality is we don't know how much actual missiles PRC has, only they can make alot, and even most competent western analysts down to estimating launchers/TELs, but that's like counting guns without knowing how man bullets. Rocketforce also barely has to invest any stockpile to TW, airforce and army MLRS has most of the TW distance fires covered. And really it's matter of satuating patriots, again magazine size known quantity, and mop up uncontested.
>can isn't volume
See huawei increasing handset production to the point of eating at Apple share. Techinsights breakdown of SMIC N+2 that suggests decent yield. Now HW also planning to hammer their own AI cards. That's sufficient volume short medium term on ASML. Once they have their own twinscan DUV equivalents they'll hammer 7nm at scale even. Litho 1/3 of production cost when doesn't accoutn for other cost (licensing) etc that PRC will save on by just not using western products. When Nvidia is selling AI chips for 400% markup, there's plenty of room to be profitable at lower yield.
>china state propaganda.
I mean these are western reports on PRC export composition. Trade data/flows has to balance, they know what's coming out of PRC since... it's going into their borders. Steal IP and innovate where they lead isn't mutually exclusive. And per your filters / MEMs example, PRC fine buying from west, but when it sees strategic need to indigenize, they do it in record time. If you don't think PRC can make camera lenses then you haven't kept up. Lots of buzz in photography circles last few years of PRC lenses catching up to Japanese. That's just commercial sector end spamming for commodity. As for EUV, see SSMB. They don't need to match EUV when they can just jump it. EUV is great and all but it's also failure in sense it was suppose to be cheaper tech that ended up as 200m machines due to development difficulties. You're pretending it's blackmagic when it's an engineering problem like every other. One that's IMO much more manage than something like turbojets, commercializng EUV took ASML had ~10k employees, Zeiss SMT had ~3k, Cymer had ~1k, who worked without strategic urgency for 20 years. It's really a modest effort vs Manhattan or even Boeing/Lockheed/Airbus at 100-150k employees for aviation. It's a problem that PRC can throw people at money at and do in ~10 years, as hinted by ASML CEO himself. So trend so far is western experts overestimate.
E: Last post to not bother Dan. You can have last word.
And yet everyone keeps dealing with China, because money talks and China has lots of it.
Whether it's Intel and Nvidia exporting their latest tech to China, Apple manufacturing iPhones in China, BYD poised to crowd out the car industry (sans America), or having our cheapass blue jeans produced in the finest sweatshops, everyone wants to deal with China.
No one wants to deal with the dynamite basket that is Taiwan. Initiatives like CHIPS are so we can eventually leave Taiwan out to dry, not because we want to overcome China.
>No one wanted to sell the tech to china to make modern chips before all the wars and saber rattling of belt and road (facing).
Except everyone wanted to sell to China because the monies involved was ludicrous and that is still the case today.
>And china lacks the will, talent and now access to catch up.
China has practical implementations of 7nm process and are shipping product today from what I understand. Pedantically, China is ahead of Intel and thus the US right now. This is something the west should be terrified of, now is not the time for coping or denialism.
Let alone the implied sentiments of "Ha ha, Intel's losing big bucks! Losers!" in threads like this which are the epitome of counterproductivity if the goal really is overcoming China instead of just ridding ourselves of Taiwan.
>See the Ukraine war and Russia. To see what "no chips" looks like. modern optics on tanks are now outside what they can do, never mind glass cockpits of 4th and 5th gen fighters, never mind gps and smart mentions, never mind drones...
And yet Ukraine still hasn't won. With western high tech.
Mind you, I want Russia to lose and lose hard because I hate the idea of warmongering validated as a means of diplomacy in the 21st century, but it is what it is.
It also goes without saying you don't need high tech to kill people and destroy entire landscapes.
>China isnt stupid, they keep pouring money into missles in the hopes of being able to do exactly what you're saying.
Yes? I said China might blow up Taiwan and they most likely eventually will. China couldn't care less about TSMC or really anyone and anything that exists on that island, they just want the island and they will invade if or when the geopolitical winds blow appropriately for them.
>If you pay attention to recent conflicts you will note that the patriot missile system has been shooting down the best Russia has to offer. You will note that the anti missile ships off the coast of Yemen are using the same system and missles. That the brits and French are in on the game.
Congratulations, we've managed to do the same thing as the misguided cavemen but did so behind schedule and over budget. Truly an achievement for the Nobel Peace Prize.
>>> China has practical implementations of 7nm process and are shipping product today from what I understand.
Yes china made a 7nm chip. There has been quite an effort to figure out how they got an ASML machine to do that because they dont have the tech to build that locally. They also dont have the tech to build the lenses for it either (Zeiss holds that). So our German(Zeiss) and Dutch(asml) partners have the tech to build the chips on lock (and aren't cutting that loose). It's not like that chip got stellar reviews either... Furthermore china then went and banned most government workers (a large chunk of the population) from using iPhones at work. Even with all this the sales numbers and reviews out of china were ... mixed. It's hard to filter whats propaganda, but there are some legit complaints about the device. ... this doesn't help with cameras (chips) or band filters (5g, where the mate 20 really sucks) but these are less critical techs and easier to catch up on than something like EUV... A good primer on ASML and chip production is asianomitry on YouTube.
>> And yet Ukraine still hasn't won. With western high tech.
60 tanks and air defense isnt going to win them the war. The modern jets are coming but far too late. The long range weapons they have needed were also very late and in low quantity. If you haven't been paying attention their home grown drone program has made a mockery of the Russian navy and their long range air drones are now strking deep into Russia.
Everyone is afraid of "poking the bear" the French seem to be getting over that now. However you should be alarmed that the French are suggesting putting troops on the ground in Russia. Their nuclear policy is one that allows for them to use those weapons preemptively....
There has also be a fair bit of learning in the Ukraine war.... Drones have made 155 artillery into precision weapons with incredible impact. We also cant give the Ukrainians enough 155 to keep them in the fight, and there is some debate about who should foot the bill for building more factories to make shells. Our lack of "basics" has done as much to harm the war as limiting ranged and high tech weapons.
No one would win the fight with what the Ukrainians have been given... and we would not take the fight if that was what we had to work with.
>> Yes? I said China might blow up Taiwan and they most likely eventually will.
Their moment to do this was after Korea, or Vietnam. The problem is that they dont have the navy, missies or landing power to take the dam island from the Taiwanese.
>> China couldn't care less about TSMC or really anyone and anything that exists on that island,
China thinks that Taiwan is filled with Chinese. Some of the tawinese see themselves as Chinese. Blowing up the island would be counter to decades of their narrative that Taiwan is part of china.... China blowing up the island is about as likely as the US blowing up purto rico.
The idea that this is the goal of china is deeply misinformed.
>> Congratulations, we've managed to do the same thing as the misguided cavemen but did so behind schedule and over budget. Truly an achievement for the Nobel Peace Prize.
Right. You really need to catch up on news and geopolitics.
1. China/russia/india all spent a lot of money on hypersonic missles with the idea that they would be a game changer in taking out carrier groups and bypassing missile defense.
2. The few patriot batteries that we gave to Ukraine have shot down Russias modern hypersonic missiles. This makes 2 decades of posturing by china and Russia moot.
3. Chinas rocket force have been all over the news. Some intel agency has leaked all sort of internal info about them. Chinas rocket force are their "defense" against a us carrier group, they are the tip of the spear to invade Taiwan. You can read the news yourself but the highlights are, corrupt, and unprepared.
4. Those patriots are now on ships. Our navy has the capacity to protect a carrier group from the biggest non nuclear threat hypersonic. And these folks are getting practical training off the coast of Yemen right now. Our allies are even in on the action, the people who would be blockading china with us in the event of a war.
China is saber rattling. We did it right back to them... and showed that our sabers work and showed the world theirs dont. It's a bit of detant but its working for now.
Part of the USAs superpower is letting the market decide where to invest. China micromanages its economy and can’t anticipate everything the market can.
> You will have noted that not having domestic chip production has fucked Russia, and the restrictions on china are to inhibit pass through and limit what they too can do.
You must have caught up on the Ukraine conflict from the places that have been claiming that the next Wunderwaffe (TM) will surely change the war for 2 years now. Russia is perfectly able to feed its own domestic daily demand on chips and everything else, not only Chinese but even European products are entering Russia through intermediaries still, and their economy doubled its growth. The smoke screen that the corporate media creates on this conflict is much bigger than the one they created on the Iraq war.
Yes, it seems they are following the large defense dinosaurs of the past into this perpetual too-big-to-fail mode of milking the public for free money. However, we've found ways before to push some things back into the private sector, to find people with energy and new ideas to run circles around stagnating boeings and lockheeds, and show them how it's done in space and other areas. It's been pretty fascinating to watch. Maybe the USG will find some way to light fire under Intel's ass too, like maybe with a little help from Taiwan or something. How these companies go from a super-innovative disruptor to a welfare program and then back has been fascinating and scary to watch. Doesn't really fit into any of the "left" or "right" cliches.
Given the geopolitical (and seismic) risk of having many of the world's most advanced fabs in Taiwan, it seems like it makes sense to try to locate more of them in the US from a national security standpoint (I guess that would be a 'right' argument - or at least traditionally the right was concerned about national security). Given that Intel (for better or worse) is the US' most advanced semiconductor company and that they know how to build fabs it seems like it makes sense to send some $$ their way to build more fabs in the US (I guess that would be a 'left' argument?). The market doesn't seem like it's going to optimize for the national security issue so the gov stepped in with the CHIPs act.
Can you imagine any market-based mechanism that would address this problem on it's own? Keep in mind that it takes at least 5 years from the time a fab is planned until it's operational. And it takes about $10B to build a competitive fab. The market prioritized social media from about 2005 to 2020 or so, and seemed to de-prioritize investment in things like semiconductor production. That doesn't seem to have worked out so well for us.
it feels terrible when incompetent people take over and start pissing away our money, essentially holding whatever defense concern they managed to monopolize as a hostage. So that part to me is very depressing.
My bigger point was that this process is not linear or predictable. Saturn rocket was a miracle that seemingly no market forces could afford to build. Organizations that built that miracle have exhausted and are nowadays essentially welfare. Now, SpaceX and starlink are a private miracle that was hard to imagine just a couple decades ago. Not sure what the moral is here, but somehow we seem to find a way, whether private or public.
The question mark makes me think there is some uncertainty here and I am irrationally upset about the ambiguity. Of course it is taxpayers’ money! What is possibly unclear about the situation?
By that logic it's actually Intel's money, because it used to belong to the taxpayer, but the law says that Intel gets it now. The trillion dollars spent on the F-35 was quite an inefficient use of Lockheed's money.
While that's how things work at the district or even state level, that's not quite how money works at the federal level. To a first-order approximation, federal government spending is a tax on dollar-holders, with federal tax corresponding roughly to fealty (such service being delivered to dollar-holders). They're decoupled because the US federal government, via the Treasury, is able to create and destroy dollars.
The IRS is responsible for collecting taxes, but delivers them to the Treasury for reallocation or destruction. Since money is fungible, only the net change in extant money matters, and they don't actually have to destroy and reprint notes… though in unrelated news, they do also do that.
Though, at organisational levels below the central bank… once money has been set aside for the interests of the commons, does it truly still belong to those it was appropriated from? If I give money to a malaria charity, I don't then get to say "hey, spend that money on vaccines instead of nets": it's theirs to allocate as they see fit. If I don't like how a charity organises money, I'll choose a different one: just like how, if you don't like how your local government organises communal resources, you should vote accordingly (which, in theory, works) and otherwise take political action.
> While that’s how things work at the district or even state level,
Aside from the fact the fiscal argument you are making isn’t the most important reason why the “taxpayers’ money” thing is wrong at all levels, its also not accurate even within the narrow view of “where does the money come from” at the district or state level, either, since those entities get funded by sources that aren’t “the taxpayers” as well, including other levels of government, like the feds, for whom you acknowledge (correctly) that it is not true because of money creating power.
But "other levels of government" are also getting the money from taxpayers, and money printing is constrained by its propensity to cause inflation, so in practice the amount of it that happens will correspond to the amount necessary to hit the Fed's inflation targets rather than scaling with the amount of government spending. Consequently any incremental increase or decrease in government spending would have to correspond to the same increase or decrease in tax collection, whether as the full amount in the current year or as future interest payments via deficit spending.
Inflation is when the number of dollars increases faster (or decreases slower) than the amount of value in the (USD) economy does. If the money is invested in a way that creates over a dollar's worth of value per dollar, printing money can actually reduce inflation.
In typical circumstances, such investment possibilities do not exist. However, in disasters (e.g. a 4-year-old virtuoso needs surgery to repair a ventricular septal defect; an average person is homeless and needs a place to live; an asteroid's on a collision course with the planet), creating additional money is often the correct decision. Experts disagree whether the recent pandemic was such an opportunity and – if so – what the correct response was: we saw lots of different strategies playing out in lots of different countries, but I don't think that's helped resolve the controversy.
You're combining two different things. If you print a dollar and use it to improve productivity by more than a dollar, this might not cause inflation, but the same is true of raising general taxes by a dollar and using it for the same purpose. And those purposes are scarce, because by definition they have to have above-average productivity.
If you have one then it makes sense to fund it regardless of whether the money comes from printing or taxes, and the inflationary/deflationary consequences of that spending being priced in. And then we're back to the amount of printing being set by the desired amount of inflation, and any incremental spending on top of that coming from taxpayers.
> but the same is true of raising general taxes by a dollar
No, because taxes pull resources out from specific, easily-measurable areas of the economy, and that has effects on people's motivation. Increasing the value of the dollar applies deflationary pressure. High taxes on economic activity, combined with deflation, is not a good combination: people will stop spending money, and while that does relieve more opportunities for the government to invest in the economy… the competence of governments tends to be concentrated in a few specific areas, and things tend to go wrong when an unprepared government tries to run an entire economy directly.
> And those purposes are scarce,
Assuming the Efficient Market Hypothesis applies, yes. In practice… no, they really aren't. The average person contributes more to the economy than they are paid (a corollary of the theory of mutually-beneficial trade), housing costs less than ¾ of the average person's income, and people are a lot less effective at producing economic value when they have poor living conditions. The only way it would make economic sense to leave people living on the street, or constantly in and out of hotels, is if we assume the currently-homeless are inherently inferior at economic activity. Every UBI study ever has yielded evidence contrary to this assumption… and yet, we still have homeless people and we aren't doing large-scale UBI, because… reasons.
(Note that I'm ignoring the moral, human rights argument. I actually care more about human rights than numbers-go-up: it's possible that my bleeding heart is interfering with my cold, hard rationality, letting me overlook some logical flaw in this explanation.)
> If you have one then it makes sense to fund it regardless of whether the money comes from printing or taxes,
Yes! But humans are not ideal rational economic actors, and groups of humans are even worse. There is so much low-hanging fruit, you don't even need to get into the controversial things (like how exactly to fix / avert a recession) to take advantage of them. We've got years-long waiting lists for chronic health conditions that can be solved by a single drug, or a three-hour surgery. Public transport is eschewed in favour of just increasing the width of roads, leaving people shackled to slow, expensive-to-maintain, time-consuming-to-drive private cars for their travel needs. I have never seen an economic argument that supports the status quo: only political ones.
For some reason, there's usually political will to reduce taxes. (There's currently a European Citizens' Initiative to increase taxes,[0] which has reached the required threshold in France and half the required threshold in Belgium,[1] making it the most popular Citizens' Initiative currently running,[2] but politicians tend to get elected by promising tax cuts.) There's also political will to keep inflation in check. There is rarely, if ever, political will to reduce the numerical quantity of money in circulation (except as it impacts inflation), so I predict that's where many of the resources will come from, if one of the currency-controlling governments (UK, US, EU, Sweden, etc) figures out how to fix some of these problems.
That is the government that decides to spend the money is a fact. Whether the "taxpayers" (and what "taxpayers" means) are the ones that will pay for it, is an interpretation up to perspectives on political economy.
There would be an ironic twist if nvidia (who first competed with AMD nee ATI) was forced to license their IP to intel for national security reasons, but it would rhyme with history (x86 licenses to AMD) and benefit the consumer.
Why couldn't Intel fab GPUs for Nvidia? Taiwan just had 2 7+ earthquakes this morning, so we can add that to the geopolitical risk of having a large number of the world's most advanced chip fabs on the island of Taiwan.
Other than process specifics, they could, and that's exactly what happened when the 1980s US Government got worried about a single CPU supplier. The result was (to some extent) modern cheap computing.
This comment is an absolutely perfect example of Brandolini's Law [1]:
The amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it.
One sentence of the same tired myopic cynical trolling bullshit causes 20 responses and paragraphs of explanation to try and refute. Absolutely perfect.
Partly yes, I mean my effort was just to search for a link about the subject to something that seems reputable enough to share.
On the other hand, I just stated a fact. If by "explanation to try and refute" you mean just the story around china and geopolitics, it is not any high effort one. It is just repeating the same official story over and over. It does not change the fact of who is paying for it, as any sort of geopolitical justification for a war (or supporting a side in a war or supporting the continuation of a war) does not change the fact of who dies for it. Or who profits.
> a blow to the chipmaker as it tries to regain a technology lead it lost in recent years to Taiwan Semiconductor Manufacturing
Is it a blow, though? I'd expect exactly a big loss if they're investing to regain market leadership. If they're investing well, the higher the loss they can take, the better long term prospects. Am I missing something?
apparently they attribute the loss to deciding not to purchase EUV earlier from ASML. did they really think they knew better than company that's defining SOTA?
Intel & US research labs were where the state of the art was created. We gave it to ASML.
They've advanced the state of the art enormously, but it's i.possible to ignore how much is because we helped them. Because we wanted a factory for chip making machines, in a liberal free and democratic part of the world. Because just having vertically-integrated Intel's and AMD's (back when they made their own chips) wasn't enough .
I'd love for much more detailed write-ups on these technology transfer, this history. Alas trying to search for Intel + ASML is massively overwhelmed by search engine's rececy bias, filled with what machines Intel is buying. But here's this ASML history write-ups brief description:
> The research ASML received from US laboratories was even more valuable than the revenue provided by American firms. In the 1990s, Intel invested heavily in EUV research, subsidizing national labs. Insights developed at Lawrence Livermore and National Sandia laboratories were eventually passed to ASML after Intel realized no American firm could effectively commercialize these findings.
However, it is true that a company can report a loss on it's P&L statement (due to non-cash expenses like impairments / write-downs, reserves) while they in fact did generative positive cash flow in that same period.
But the company is still profitable on the whole (even if their margins are not great these days) they basically just shifted some of the costs from one department to another
"Expenses" and "revenue" are rather fuzzy concepts in this case: Intel's foundry business (as in, Intel manufacturing chips for somebody else) is still mostly aspirational; almost all of the chips they make are Intel chips, and there's a lot of wiggle room for how much of the revenue from selling Intel-branded chips gets assigned to their foundry business vs eg. the Client Computing Group or the Data Center and AI division. If they were reporting Intel Foundry revenue only from outside foundry customers (they're not), then Intel Foundry would never be able to come close to profitability.
not your point at all, but you say almost all of the chips are Intel chips. I was under the impression that literally all of them were Intel chips. Have they started manufacturing chips for others?
In a world where inter-company expenses [1] charging out for IP use, accounting & HR services, strategic direction etc. etc. exist, something so simple as expenses being larger than revenue being an accurate economic measurement is wholly unreliable at best. Intel market analysts will be starting with the statement of cash flows and reconstructing how they think it should all look from there using the financial statement footnotes and all their other research & knowledge.
[1] Frequently cross-border for tax purposes but there are plenty of non-tax reasons for filling and draining hollow logs.
