Fab Heaven; Fab Hell

It’s not as bad as you think. Europe actually has 320 fabs on 265 sites, says Heinz Kundert, President of SEMI, and, this year, the installed fab capacity in Europe will grow by 6% and, next year, SEMI expects it to grow 5%. 300mm capacity in Europe will grow 23% this year and 14% next.

The top three countries have half the 320 fabs. Germany has 30% of the total; the UK and France have 20%; the fourth largest number of fabs is in Russia, and the fifth is in the Netherlands

In the last ten years, Europe has spent $55 billion on semiconductor manufacturing equipment and materials, says Kundert.

So we’re not yet reduced to the debt-ridden, wastrel, Theme-Park-Europe status some people say is our fate.

In 2011 alone, $7.5 billion has been spent in Europe on equipment and materials.

Of the 320 fabs in Europe, 79 make ICs, 84 make power devices, 48 make MEMs, 25 make RF devices, 17 make printed electronics and the rest make an assortment of products.

However, in one way danger lies. “State of the art fabs are too big to be financed by one country only,” says Kundert.

The coming era of the multi-country-owned fab leads one to adapt an old joke:

Fab heaven is where the machinists are Japanese, the R&D is done by the British, the IC design is done by the Indians, the canteen is run by the French, HR is run by the Swedes, the management is Taiwanese, customer service is done by Italians and the marketing is done by Americans.

Fab hell is where the customer service is done by Germans, marketing is done by Belgians, HR is run by the Chinese, the canteen is run by the British, the machinists are French, the management is Indian, the R&D is done by the Irish and the IC design is done by Australians.



  1. Ouch! Alien94086, that’s a low blow for a Continent which took the lead on 300mm.

  2. The Landshut plant was upgraded at least once since it opened and when it closed a few months ago it was a 200mm line capable of >10K Wafer Starts Per Month with 150nm CMOS and other processes. In the European scheme of things it would have been a significant plant.

  3. Strewth, Robert, stone the crows don’t be a frog in a sock with this Pom and come back and spit a yarn soon

  4. David, Face it you’re a Whinging Bloody Pommy, so don’t “put on airs” or try to “don’t come the raw prawn” with me, I’ll be banned from this site before I put up with that sort of behavior, from a POM! 🙂

  5. Bonza, Robert, fair dinkum cobber, no spitting the dummy for an Aussie – that’s for Aliens

  6. On behalf of all Australian Analog IC designers, (both of us) I’d like to formally protest this clear case of discrimination.
    We’ll be holding our biannual general meeting behind the bike shed, this afternoon, don’t be surprised if we decide to occasionally boycott this site. But don’t worry too much, the boycotts are usually forgotten after a few beers.

  7. No apologies required, Alien94086, because Indian managers, Aussie IC designers and Irish researchers have something called a sense of humour which, obviously, Aliens don’t. As for Landshut, I seem to remember going there about 20 years ago when Hitachi opened it. Is a 20 year old 6 inch fab ‘significant’ in the Alien scheme of things? If so, I’m surprised.

  8. Don’t worry – I accept your apology on behalf of all offended Indian managers, Irish R&D staff and Australian IC designers. TO return to the more serious point – Europe lost a significant 200mm fab recently with the closure of the LFoundry ( formerly Renesas ) 200mm plant in Landshut near Munich. This would have been one of the bigger European facilities of the 320 on the SEMI list.

  9. I know, I know, Alien94086, it’s all very reprehensible, indefensible and quite nonsensical. It’s appropriate to have it pointed out by someone so sensible.

  10. Why did the Indians get the management job, the Irish get to do the R&D and the Australians get to work on IC design in Fab Hell ? And Fab Heaven has the British doing the R&D ? The joke does not add much value to the article and detracts from an otherwise well made point.

  11. Yes this is really tricky now. There are so many metrics for ‘state of the art’ that the phrase is becoming meaningless. We need a new definition.

  12. I think this depends on how you define ‘state of the art’. Advanced nodes and advanced technology are not always the same thing, even though it is easy to say one and mean the other, as I did myself only yesterday. Infineon’s new thinned 300mm wafer technology is very much ‘state of the art’ and shows Europe at its best.

  13. With design costs spiralling ever upwards, the European customers may exist, but are they going to trust their latest $100m baby to a European fab, whatever its credentials or are they going to stick with TSMC where they know the people?
    I think that Torben is probably right, now is the time for Europe to seriously start getting looking into the post silicon world.

  14. Judging by the number of European customers TSMC has, Frédéric, quite a few European companies could use a European fab

  15. GlobalFoundries could argue that a good fab shouldn’t be judged on its location, canteen quality or customers list, but on its ability to actually produce working chips.
    More seriously, which european company would really need a state-of-the-art fab, apart from ST for its STB chips maybe ?

  16. Spot on, Ian, the issue isn’t: Can you afford to build a fab? It is: Can you fill a fab? If you can fill a fab you can make a shed-load of money and pay back the building costs.

  17. The issue isn’t just whether you can afford to build the fab — which is a big one, but isn’t really a problem if the returns are good enough, which they obviously are for the likes of TSMC.
    Everyone in the world buys the same fab equipment from the same suppliers, and labour costs are a small part of the cost of running a mega-fab. The big reasons that the mega-fabs are where they are is low-cost (subsidised?) finance to build them, low-cost (subsidised?) power and water, and low-cost (subsidised?) buildings with laxer regulations built in no time flat.
    Given all this a fab built in Europe simply wouldn’t be able to compete on wafer price, unless the governments put similar support into not just building the fab but running it. Which would of course be seen as an unfair subsidy by other EU businesses — which it is in a way, but that’s what’s needed to compete in this market.

  18. Nice one Robert

  19. what a crock of ****
    Of course many European countries can afford wafer fabs BUT they choose not to. Thats it, nothing else
    Taiwan is a country of about 23M people but they seem to be able to find the money to have more than one megaFab and even maintain some local competition. Korea has about 60M (from memory) yet also has multiple state of the art fabs. So don’t tell me that individual European countries cannot afford fabs.
    The real problem is that money managers are being fooled by other money managers into making “LOW” risk loans in soon to fail ventures. Of course that make great returns for a few years (all ponzi schemes do) but then it all blows up. and the tax payer gets asked to bail it out…
    Whenever a money manager rejects a factory as too risky they are making a differential judgment that they can find other better returns at lower risk… The problem for all manufacturing is that we don’t know how to lie anywhere near as well as the professional lairs AKA wall-st so all fabs will always be too risky the joke of course is that loans to a fab would rate at best as CCC but loans for CSD/CDO’s rated as AAA with no underlying asset securing the loan. Explain that one to me and I’ll think about accepting your original premise

  20. If the EU can finance the LHC, they would surely also be able to afford a front-line fab.
    But maybe that isn’t what needs to be done. Maybe Europe should invest in post-silicon technologies rather than trying to squeeze the last bit of performance out of silicon at ever-increasing prices. Maybe it is like trying to breed better horses when someone in the next town is inventing the automobile.

  21. Who would be the paymaster?
    On a serious matter, the article captures a key point about the cost of a serious state of the art fab and that no single European country or company is capable of affording one.
    For all the successes, not being able to compete at the leading edge must say something about Europe.

Leave a Reply

Your email address will not be published. Required fields are marked *