In a major push to give Europe pride of place in the global semiconductor design and fabrication ecosystem, 17 EU member states this week signed a joint declaration to commit to work together in developing next generation, trusted low-power embedded processors and advanced process technologies down to 2nm. It will allocate up to EUR145bn funding for this European initiative over the next 2-3 years.
Recognizing the foundational nature of embedded processors, security and leading-edge semiconductor technologies in everything from cars, medical equipment, mobile phones and networks to environmental monitoring, and smart devices and services, the European Commission said this is the reason it is crucial for key industries to be able to compete globally and have the capacity to design and produce the most powerful processors.
It’s kind of odd that Europe does not command a more prominent position in the semiconductor industry, since the one company that enables the constant progress in this sector isn’t American, Chinese, or Japanese – but Dutch. ASML is by far the world’s largest developer and producer of photolithography systems, which are the machines companies like Intel and TSMC use to fabricate integrated circuits.
Their machine are some of the most advanced machines in the world, and all the advanced, high-end chips from Intel, Apple, AMD, and so on, are built using machines from ASML. It seems odd, then, that Europe’s own semiconductor industry lags behind that of the rest of the world. This investment seems to aim to correct that, and that’s a good thing for all of us, no matter if you’re European, American, or from anywhere else – this can only increase competition.
Yes there is a bit of a problem with Europe lagging in tech. Historically pre-war Europe was a lot of fractured markets with a lot of built in protectionism. This was basically standard practice and had served Europe well until then. However, the United States with its large single language market operated differently. Europe knew the issues with fragmented markets which is why moves began to correct this with what later became the EEC then EU. The US backed this and it was one reason for the US funding the Marshall plan.
The UK was still benefitting from its pre war industrial and technological lead. As the UK was hugely in debt to the US over lend lease the US had the upper hand. Due in part to bad management by successive UK governments and the US outright lying and bullying to gain market advantage the UK has all but hollowed out as an industrial and technological base. Throw in language barriers and still to some degree a European fragmented market and this may explain why culturally, or psychologically and sociologically, conditions exist which hold Europe back from being more prominent in technology. Post Brexit the EU has lost the UK which looks like it’s going down the shitter again. The US is acting like an empire and its presence on the internet and economic policy and marketting in general tends to frighten competitors, buy them up or bankrupt them, or drown them in a firehouse of propoganda and marketting so Europe doesn’t get the mindshare it may deserve. The same is true to one degree or another with countries outside of Europe.
The EU plan is an acknowledgement of these issues and a welcome drive to correct things but I suspect unless there is some attention paid to fragmentation, language barriers, and social and business issues (not to mention problems with the US meddling and throwing its weight about) the money spent won’t be as effective as it could have been. Russia is also an untapped resource on Europes doorstep. The Russia question won’t go away and some kind of normalisation and harmonisation will need to take place with Russia. The recent treaty with China addressing market issues and human rights issues is another good step which leaves the US lagging behind and may help smooth the way for improving relations with Russia.
Good points. Also the transistor, IC and microprocessor were all invented in the US. Never underestimate first mover advantage. And there was all that military money and flowing directly and indirectly into R&D during the cold war. This is in a sense what the EU is trying to counter with the IPCEI programs.
Interesting that the C there stands for cluster. This is the key to success IMO. Prior EU approaches where they spread the research funding like butter and forced involvement of universities was too constrained, The latest approach looks more outcome vs input oriented and might work.
Moreover, NXP, Netherlands, owns unique CPU designs they can improve. They can even embrace the RISC-V route… So, what’s up, Doc ?
Kochise,
I am supportive of an open architecture like RISC-V and I think a considerable percentage of the industry probably is too. However I am extremely concerned that mainstream vendors will end up exploiting the openness for their own ends while selling consumers essentially a closed hardware implementation. Nvidia uses RISC-V cores in some of it’s products but it not user accessible. It does me no good whatsoever if the underlying CPU uses an open architecture if the overall platform is locked and controlled via 3rd party restrictions.
I’ve long taken the position that owners are entitled to the keys to their own hardware, but in reality it’s the system integrators who are be buying and reselling CPUs to us, the consumers and for better or worse it is they who decide what level of access owners will have. I really hope RISC-V devices can turn out better than ARM devices in terms of vendor locking, but I’m not very optimistic. Ironically there’s a chance that those of us who strongly oppose the normalization of restrictive hardware are going to be left with x86 as the most solid choice. Even there traps are being set with secure boot.
Looking forward to the new bureaucracycore in 2084
Can’t be worse than that one time they decided to fund Symbian OS just before the parent company put in “maintenance mode” and gave it a termination date, can it? Fortunately the EU managed to pull funding before any money was spent, but it still shows how governments exist to regulate, not produce.
https://www.theregister.com/2011/05/25/ec_cancels_symbeose/
Back on topic, this is happening because the EU is butthurt that GlobalFoundries decided to not move past 14nm, because what will happen to the European grandeur? So they decided to throw some taxpayer money at the problem.
“The total initial investment in Intel was $2.5 million” (from https://en.wikipedia.org/wiki/Intel#Origins).
Why start small, when you can star very big! Like 100s of billions of Euros?
All the brand name US tech companies were founded by one or two really dedicated and talented people and first round of smaller investors that kept them rolling. Take Google, Facebook, Intel, AMD, or even HP, which literally started the Silicon Valley were very small operations.
Giving lots of donations to large companies would not easily jump start innovation. Even those large companies now buy smaller startups to keep innovating. Look at the list of (only the public) acquisitions by Google/Alphabet:
https://en.wikipedia.org/wiki/List_of_mergers_and_acquisitions_by_Alphabet
That money could be much better spent to grow some startup funds, as grants to universities, and research labs. It can also be useful for small-to-mid size operations. However flushing big companies with big money might not be the best way to spend tax dollars, … sorry tax euros.
sukru,
It’s all relative. Companies like google started relatively cash rich thanks to mad money in the dotcom era with money from billionaires including Bechtolsheim & Bezos. In early years facebook was desperate for money and needed microsoft to help fund it before becoming sustainably profitable. While it isn’t to write off the potential of talented people, the reality is those with rich & influential connections are significantly more likely to beat smaller startups with fewer financial resources. A more typical very small startup operation is one that is economically forced to be profitable from the get-go and grows organically. Being able to invest more up front without a requirement to be profitable for the early years it is a proven path to rapid success, but not everyone has the economic means to do it.
In a way the EU’s EUR145bn investment is kind of acknowledging that lots of financial power is needed to help get the gears in motion and for any hope to significantly change the cutting edge chip fab industry’s momentum in a quick timeframe. A very small organic growth startup isn’t going to cut it. It’ll be interesting to see what happens with this.
Well, many of us find these corporate acquisitions depressing due to the end game, which is ultimately hyper consolidated markets in the form of oligopolies, duopolies, and monopolies… It can be where innovation goes to die. Google for one is guilty or acquiring or building new products and then abandoning it. It’s sad, but all too often corporate acquisitions are followed by mass layoffs of the very employees who helped build the small company. Usually the big company is going after property (like patents) and market share. If you have the money, then buying your way into the market can be much faster and cheaper than having to compete.
@alfman @sukru
These programmes can be viewed alongside the EU firming up its regulatory environment with regard to monopolies and trade. There is also the issue of the EU being obligated to level up all member states. I don’t know about other EU states but France and Germany have fairly well established strategies at either a policy or cultural level to preserve strategic industries and ownership.
The Horizon Europe project has a nice chart highlighting the main points behind the investment programme.
https://ec.europa.eu/info/horizon-europe_en
On a slight tangent…
One of the reasons for creating the BBC was to create a framework for broadcasting and media production which could stand up against America. It has since been dumbed down and largely taken over at a managment and current affairs level by staunch Tories. There is a lot of talent in the movie production industry but it’s largely a screwdriver economy. The UK is appalling at management and is congenitally incapable of running large movie projects.
The UK under successive government (mostly the Tories who couldn’t manage this with domestic policy let alone Europe but other parties are guilty too) pretty much had none of the above imperitives. The UK lacks a modern constitution which enabled the Tory coup (and I do believe it is a coup in a very real sense) and has a lot of problems with fragmented governance, the entrenched class system, and a lot of other problems small and large. From the government and now largely right wing tax dodger non-dom or foreign Murdoch owned owned media you wouldn’t know these EU initiatives existed. Throw in a blatantly interfering and dirty City of London financial sector.
There are more than a few people in the UK who would welcome the arrival of European tanks at Dover. Can I hold your coat? Until the UK reforms I regret the EU is better off without the UK or at least until the Tory coup has been repelled and reform underway.
I’m not a nationalist. I just don’t like stupid.
THIS. ALL of it.
There is no thing as “starting small” in new process nodes. You see, Moore’s Law says that transistor density doubles every 18-24 months, but an unknown consequence of this is that the cost of a new fab also doubles every 18-24 months. Moore’s Law isn’t free.
Which brings the question: Why does this sort of thing need to be funded by taxpayer money?
kurkosdr,
I’ve always been bugged by the misnomer of “moore’s law”, it was never a natural law at all, only an observation over a limited dataset. Sustained exponential growth is physically impossible in the long term. Ultimately as the difficulty increases exponentially we should expect diminishing returns compared to past trends and there’s evidence this is already happening. There may not be a “hard limit” in a physical sense….well, technically there is in a theoretical sense, but we’re never going to reach it due to economic realities. The more data points we collect, the more we’ll see exponential growth giving way to linear growth even even logarithmic growth as diminishing returns becomes increasingly relevant.
That’s a fair question. As you point out it’s a very expensive industry. A lot of people do believe that private investment ala capitalism is best. Still, if we leave it to free market capitalism then we should acknowledge that capitalism gives a lot of preference to wealthiest entities who inevitably consolidate power and control at the expense of competition. The systemic shortages many manufactures are experiencing are partly a result of many mainstream vendors being at the mercy of ASML to fabricate their high performance chips. While I don’t have a problem with ASML per say, arguably having such extreme dependence on hyper consolidated supply chains is no good. Monoculture caused by market consolidation ends up negating one of the basic tenets of capitalism: consumer choice. One of the ironies of unfettered capitalism is that it leads to oligopolies, duopolies, and monopolies. When you lack meaningful choice, the entire foundation of capitalism based on consumers having choices falls apart. And this brings us back to your question. Outside of government, very few will have the means to compete with incumbents (except other incumbents).
The most advanced fabs in the world, TSMC in Taiwan and Samsung in S.Corea are totally backed by their respective goverments. And IBM when it was still relevant was also quite benefitting from US gov. contracts.
Competitors to ASML are Nikon (Japan) and Chinese OEMs. But EUV lithography is so extreme and expensive that development efforts were shared by all fabs.
Some techologices not as glamorous as cutting edge logic nodes are also important. For example European manufacturers are quite good at high power high voltage transistors (SiC, GaN…) now used in electric cars inverters.
“$2.5 million” well, a little more than that, directly from that Wikipedia article: “equivalent to $18.4 million in 2019”
But I’m very happy with the amount of money.
Because a the price of a new factory like Intel or
TSMC.
If someone went with millions today they could do nothing much. Because I think the goal is to create a manufacturing facility in Europe. Look at the price of a highly experienced chip building for a new factory:
“New production – D1X Oregon $5B expansion,Fab 42 Arizona $7B, Fab 34 Ireland $8B expansion, Israel, $10.9B ”
https://en.wikipedia.org/wiki/List_of_Intel_manufacturing_sites
Can you imagine how much it would cost to build something new for less experienced chip makers ?