Edit: AI claims HVAC and air filtering/handling can be 40-50% of energy usage. Plus a megawatt per EUV scanner (ok, the latter matches what I would have guessed, but compared to the 16 MW of a single large bucket wheel excavator or estimates I've seen for large shopping malls, that's not that insane either).
But even more importantly, they need 2 separate completely independent substation feeds, since a 15 second interruption can cause 0.1B Euro of losses.
Finally, and here is where Europe really really suffers, generally you need these things to be planned, approved and built within a year.
A great example of this is trying to compete with the New Space companies.
I think that is not even possible in China, let alone in the USA.
> Finally, and here is where Europe really really suffers, generally you need these things to be planned, approved and built within a year.
Like the offshore wind farms in the US? I think Europe and China are building these things much faster. ;)
EUV lithography and etching rooms require ISO class 1 clean rooms. Those can have a maximum of 10 particles of size 0.1 microns per cubic meter.
While a shopping mall can get away with 2 - 4 air changes per hour, an ISO class 1 clean room requires 400-600 an hour, or about 10 a minute. Airflow needs to also be controlled: vertical and laminar. Every human entering the cleanroom needs to undergo an air shower with speeds of 60-90 mph to shear off loose skin flakes, hair and lint. This area needs to maintain constant positive air pressure to repel outside dust as humans move in and out.
In the EUV lithography systems, temperature control is measured in milliKelvins (0.001 C). If a wafer stage has a temperature variation of 0.05 C, the silicon expands by a few nanometers and the multimillion dollar batch is toast. Meanwhile in the same room, the EUV machine is generating extreme ultraviolet light using a plasma light source that gets to above 200,000 C. That's 40x hotter than the surface of the sun.
> I always thought of these as high-precision, not-too-high-volume processes
EUV rooms are churning out about 150 wafers an hour, 24/7. That sounds like not too high of a volume, but each wafer is about 500 cpus. So 75,000 cpus an hour.
They have special security guards who make sure you do the procedure correctly.
Vacuum does not dissipate heat well (extra engineering in that department)
Vacuum is never 100% vacuum. If you have some dust in the air, it will mostly flow to the direction of the airflow, but in vacuum dust particles will do random walk thing and settle on surfaces.
You can't use normal plastic or any kind of lubricant anymore because it starts degasing inside the vacuum, pump down after opening to air is a multi-day process that involves heating everything up to 150°C to get rid of the water that will settle on basically every surface when you pull the vacuum, every electricity feedthrough is a potential point of failure, transfering objects between different stages is a huge hassle etc. pp.
I'm sure there's still lots of UHV and HV systems in a chip fab, but there's no way that's easier on a full factory scale than clean room operation.
Heat dissipation would be the obvious problem.
Just to address this separately: there is effectively energy rationing in several European countries, where the government intervenes to prioritize who should receive energy first (e.g., the Netherlands and Denmark).
And Japan, South Korea, and Taiwan are also energy poor states that have been able to maintain capacity.
The energy excuse is true for some manufacturing processes like automotive in CEE which were pants that were built in the 1990s to 2000s and leveraged Russian ONG, but is a red herring for a lot of other manufacturing processes in Europe.
European states have a coordination problem as Draghi pointed out 2 years ago.
https://spectrum.ieee.org/taiwan-semiconductor
https://www.datacenterdynamics.com/en/news/tsmc-could-accoun...
But no, this failure is due a mixture or privatisation and politicians looking no further than the next election, leading to structural underinvestment.
Oh, and the energy transition going faster than even the most forward looking scenarios anticipated.
Solvable but we need to be much smarter and redirect the fossil subsidies to reducing risk for grid-scale batteries and adding more grid capacity.
But yeah, we have one party that is most often in government that just loves privatisation, taxing workers more and the rich less, etc. and engages in a lot of short-term thinking. Apropos short-term thinking, one of them was a prime minister for 14 years and cut defense spending and did not even consider increasing spending when Russia invaded Crimea. And now he is the secretary general of NATO. Talking about failing upwards. Though he is probably primarily chosen for being able to communicate with Trump.
Water. Energy. Public transport (which is partially privatised). Healthcare has been a disaster.
Same thing for the UK but the results there were much much worse, especially the water (Brexit made it 10x worse). It was driven by the actions of the Unions in the 1970s (where they were taking the mickey) but of course went way way way too far. What was needed was to tweak the system such that the unions couldn't blackmail the Labour governments. Not to break it all up and sell it off for pennies on the dollar.
So that brand-new fab has to pay for itself within the first 3-5 years, or it'll simply never pay for itself. But indeed, once it has paid off its initial construction costs it'll happily keep running for another 10+ years, churning out pure profit.
An additional problem here is that the costs have been rapidly increasing as the nodes have gotten more advanced. The tech for something like, say, the 40nm node is now mature enough that building new fabs isn't too expensive, so there's a decent market for building new ancient fabs for cheap mass production. But with EUV everything has gotten so complicated that there's no reasonable way to make it cheaper as it matures, and scale-wise you're well into the diminishing returns for everything which isn't logic. It is very much doubtful whether building brand-new 7nm-and-smaller fabs for mass production will ever make economic sense.
More to the point, from 2019 they built €17 billion new fab development which doubled the manufacturing capacity available in Ireland and enabled the production of Intel 4, the company’s most advanced process technology.
It's because whenever Intel asks for anything, the Taoiseach, IDA, and Invest Ireland will move mountains to do it.
Ireland is extremely business friendly because their leadership actually listens to industry.
This is where the 'Inside Tennis' knowledge of the Irish Tech industry from someone working within it comes in handy. Enter the 21st Century Luddite, Thomas Reid, from stage left.
Thomas is an eccentric, independent bachelor farmer whose family has owned a 72-acre farm next to Intel's Leixlip manufacturing campus for generations.
Ireland’s State Industrial Development Agency (IDA) attempted to use a Compulsory Purchase Order (CPO) to buy Reid's land so that Intel could proceed with a multi-billion euro expansion of its microchip production plant. The land was valued at €5 million. During the negotiations that followed, the farmer was reportedly offered in the region of €10 million. He turned it down. Reid had no interest in being a millionaire. Instead, he continued to live on-site without a telephone or modern amenities while taking on Intel and the state alone.
The battle culminated in a landmark legal fight that went all the way to the Supreme Court of Ireland, where Reid ultimately won a historic victory against corporate overreach and protected his property rights
https://en.wikipedia.org/wiki/The_Lonely_Battle_of_Thomas_Re...
https://www.reddit.com/r/ireland/comments/3s4bg1/how_a_farme...
https://www.irishtimes.com/business/commercial-property/inte...
https://www.rte.ie/culture/2020/1102/1003311-the-lonely-batt...
How does any of this make him a Luddite?
Luddites weren’t anti-technology. They were highly skilled workers that actually wanted to be involved with it.
They wanted technology to serve them not make their conditions worse as they worked with it.
This guy just doesn’t to sell his farm. That’s not the same at all.
The Double Irish (& Dutch Sandwich) where Companies routed profits through two Irish-registered subsidiaries, attributing intellectual property royalties to a management seat in a tax haven like Bermuda, is closed since 2015.
Subsequent sequential mismatch structures (the "Single Malt" tool) that replaced the Double Irish for certain firms using non-EU residency loops were also banned. Finally there's a 12.5% charge on unrealized capital gains when a company moves its assets or tax residency out of Ireland.
The headline tax rate and the effective tax rate are also very different things. France's effective corporate tax rate was actually lower than Irelands - but France is an absolutely brutal place to do business in for employers and they've comparatively no english-language tech FDI to speak of.
i.e. a 100,000 EUR annual salary, the employer pays ~40,000 EUR to the government, then the employee gets their wage slips with additional deductions of ~38,000 EUR.
= total of ~78,000 EUR tax on a 100,000 EUR salary.
I don't think anything like that level of taxation exists in the US?
At 100,000 USD (a bit lower than EUR) the combined tax load on income for both employee and employer is typically ~30-35% depending on locale. In the US, $100k is pretty close to the median household income.
If you have extremely high incomes (e.g. 1,000,000+ USD) the gap closes a bit but even in the highest tax locales the marginal tax loading maxes out at ~60% of top-line income.
US Federal tax rates are some of the most progressive in the world. Only ~20% of households are net taxpayers after transfer payments and subsidies. Ironically, this is why it is so difficult to significantly increase income tax revenue in the US. It would require the middle-class majority to vote to raise their own taxes.
If you are completely broke and go/are brought to a hospital in the US with an immediate life threatening medical issue, they will still try to save your life.
I am too young for M&M, public schools are the worst and I pay ridiculous amount of money for private education. welfare services? depending on where you live (I live in urban area) welfare services would make sure I am just as homeless as if I didn’t get any welfare service…
and yes, the ER will take me to save my life but that about as low of a bar as one could come up with.
I think what the OP wanted to highlight is right here: in China or Taiwan etc. this would not have taken 7 years.
2019: Kick-off
2020: Construction in Ireland basically halted due to Lockdown for 18 months
Q3'2023: Fab 34 begins running a ‘First Full Loop’ of silicon
Q4'2023: Fab 34 completed. Doubles the manufacturing capacity available in Ireland, enabling the production of Intel 4 and Intel 3, some of the company’s most advanced process technologies, and the first to incorporate EUV lithography
2024: Intel sold 49% stake to Apollo GM for $11.2bn
2026: Intel buys back stake for $14.2bn
Q2'2026: Intel announce a further $5.7bn investment to upgrade existing capacity for server chips, slated for completion by late 2027
The issue is with policymaking - coordination between the EU, national, and local governments is weak so high capex projects become risky.
Edit: can't reply
> Your first 3 countries are also incredibly poor comparisons for living standards as they have horrendous work/life balance.
Their hours worked are comparable to most Western European states as well now [0]. Germany is unique in Europe based on hours worked, and is why German industry is pushing to bring back the 40 hour work week.
[0] - https://www.oecd.org/en/data/indicators/hours-worked.html
So vague as to be utterly meaningless.
And slow moving regulation literally incentivises high capex long-horizon investment.
And slow moving execution and creation of relevant legislation is a hindrance for capex in high risk projects like those in the semiconductor segment.
The Draghi report was published 2 years ago, and remains largely unimplemented.
[0] - https://en.wikipedia.org/wiki/List_of_countries_by_Human_Dev...
A country with living standards at Bill Gates levels with perfect good health until 200 where everyone is illiterate has a lower HDI than South Sudan. Even if we fix it by setting the minimum education value at 0.1 billionaires with perfect health land would have a HDI lower than 180 countries. HDI has a 90%+ correlation with GDP. It’s not a worthwhile measure.
What?
Also, slow moving, stable regulation can be better than arbitrary flippancy, but that doesn't mean it directly incentivizes investment, it just means the investment already incentivized by demand can proceed with more confidence that it won't have the rug pulled out from under it, assuming the regulations didn't discourage a tentative proposal in the first place.
Your first 3 countries are also incredibly poor comparisons for living standards as they have horrendous work/life balance.
Those alive invest in long term planning, those with a dead drive get conservative in their action.
This does not counter that an idle, decadent mentality will have curbed growth everywhere.
This was strategically planned with EU as part of bringing back production locally, similar to CHIPS Act in the US.
And it's not like US has companies doing own manufacturing outside of Intel. CHIPS specifically wanted to address that.