You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
Intel announced a €30B investment in Germany in June 2023 to build two semiconductor facilities as part of its manufacturing expansion plan in Europe. However, it's having significant trouble staffing the plants according to the Wall Street Journal. Intel says it needs 3,000 people to staff the plant in Magdeburg by the end of the decade. However, the apprentice program for chip making technicians this year is two, The Wall Street Journal reported.
The company is set to have an initial appreciate class of 20 that starts training next August, Stefanie Klemmt, a chamber-of-commerce official, told the Journal. Nearby universities, such as Otto von Guericke University, are also starting degree programs in advanced semiconductors, but these programs will take time to ramp up. In June, Intel and the German government signed a revised letter of intent for U.S. chip giant's planned leading-edge wafer fabrication site in Magdeburg, the capital of Saxony-Anhalt state. The Germany government is providing €10B in subsidies, including financial subsidies and price caps on energy. Intel's plant in Magdeburg is part of the chip giant's plan to invest $88B in Europe for manufacturing purposes over the next decade. In July, the European Union finalized approval from the European Council for its own Chips Act in an effort to compete with the U.S. and Asia for semiconductor production.
Magdeburg is not far from Dresden where TSMC will build and GF has one fab and land for more fabs if they choose to build more. Why they haven't I do not know. IT must be cheaper to build in Singapore:
The GF Dresden manufacturing site (formerly AMD) is recognized throughout as among the most successful leading-edge semiconductor production facilities in the world. Dresden fab represents one of the biggest international investments in Europe with a total investment to date of more than $12 billion, and about 3,200 world-class employees.
Fab 1, located in Dresden, Germany, is a 364,512 m2 plant which was transferred to GlobalFoundries on its inception: Fab 36 and Fab 38 were renamed Module 1 and Module 2, respectively. Each module can produce 25,000 300 mm diameter wafers per month.
Module 1 is a 300 mm wafer production facility. It is capable of manufacturing wafers at 40 nm, 28 nm BULK and 22 nm FDSOI. Module 2 was originally named "(AMD) Fab 30" and was a 200 mm fab producing 30,000 Wafer Outs Per Month, but has now been converted into a 300 mm wafer fab. Together with other clean room extensions like the Annex they have a maximum full capacity of 80,000 of 300 mm wafers/month (180,000 200 mm wafers/month equivalent), using technologies of 45 nm and below.
In September 2016, GlobalFoundries announced Fab 1 would be refit to produce 12 nm fully depleted silicon on insulator (FDSOI) products. The company expected customer's products would begin to tape out in the first half of 2019.
Magdeburg is not far from Dresden where TSMC will build and GF has one fab and land for more fabs if they choose to build more. Why they haven't I do not know. IT must be cheaper to build in Singapore:
For me there are multiple reasons why you would not want to build in Germany currently:
- German bureaucracy (building permits, environmental audits, etc.)
- Electricity cost (has more than doubled in the last 7 years)
- Labor Cost (incl. protective laws for workforce)
Locally I sense a feeling of joy (and sometimes even pride) so many chip producers are coming to Germany soon. I am still wondering who is going to absorb all of the new capacity coming online in a few years. From my perspective those new fabs will likely never become profitable because of underutilization. Let´s get some subsidies as long as they are available though!
For me there are multiple reasons why you would not want to build in Germany currently:
- German bureaucracy (building permits, environmental audits, etc.)
- Electricity cost (has more than doubled in the last 7 years)
- Labor Cost (incl. protective laws for workforce)
Locally I sense a feeling of joy (and sometimes even pride) so many chip producers are coming to Germany soon. I am still wondering who is going to absorb all of the new capacity coming online in a few years. From my perspective those new fabs will likely never become profitable because of underutilization. Let´s get some subsidies as long as they are available though!
Safety and environmental laws are almost never seen as negatives. The negatives are in the legal processes which favor continual challenges by groups looking to obstruct development, and approval processes which require lots of paperwork and time. Delays have direct business costs (taxes on acquired non-productive properties, interest on loans or bonds) and the legal processes involve attorneys who specialize in federal or state (in the US) litigation, and these attorneys are expensive pretty much everywhere. The US is very inefficient in these regards, and from reading the sagas of Tesla's plant in Germany, German laws and the local populations are no better.
IMO, it seems in every western world country we have groups of people who believe that humans are an environmental nuisance, and the goals of the objectors do not seem altruistic, but simply to use available laws to slow down or stop development in general.
Please don´t get me wrong: that is not my opinion ;-) To be honest I would not want a polluting factory next to my house either, so thank god there are regulations on that.
What I was trying to say is the way such rules are to be interpreted and handled in our local bureaucracy are very inefficient. At least from an economic perspective.
Companies, especially in our business, are expected to be profitable. The more the better. Now if those local laws result in higher potential profits elsewhere, I think that is a legit explanation for why GF might have decided to expand other locations.
I have never thought of GF Dresden as leading in any categories. Good yes, especially considering that it is a 200 mm retrofit.
Saw a related post on Linked in. When industry accepts/needs public money, they forfeit some control. If the Germany polity decides that they want a fab in Berlin, there will be a fab in Berlin. Same with Micron in Clay NY or intel in OH. Economics of manufacturing has nothing to do with it.
None of the government sponsored efforts (Japan, US, Europe) will create a cost-effective manufacturing environment which would require concentration of capital, not dispersion. The likely result will be underutilized fabs in high labor cost areas.
One final point: the cost of compliance is much higher in the USA/Europe than it is in Asia. There are many reasons for that (ability to sue/delay (NIMBO), lack of coordination between overlapping authorities, norms) but the wide dispersion of new fabs does not help. Each town and city must learn anew what rules need to be enforced and how. Very wasteful. Best location for a fab: next to another one. see tsmc and Samsung.
I have never thought of GF Dresden as leading in any categories. Good yes, especially considering that it is a 200 mm retrofit.
Saw a related post on Linked in. When industry accepts/needs public money, they forfeit some control. If the Germany polity decides that they want a fab in Berlin, there will be a fab in Berlin. Same with Micron in Clay NY or intel in OH. Economics of manufacturing has nothing to do with it.
None of the government sponsored efforts (Japan, US, Europe) will create a cost-effective manufacturing environment which would require concentration of capital, not dispersion. The likely result will be underutilized fabs in high labor cost areas.
One final point: the cost of compliance is much higher in the USA/Europe than it is in Asia. There are many reasons for that (ability to sue/delay (NIMBO), lack of coordination between overlapping authorities, norms) but the wide dispersion of new fabs does not help. Each town and city must learn anew what rules need to be enforced and how. Very wasteful. Best location for a fab: next to another one. see tsmc and Samsung.
I remember a comment from a radio interview several years ago that NASA has a so called "50-state problem".
He said in order to get support from senators and representatives from all 50 states, NASA's supply chain spread widely into all 50 states. It creates enormous amount of problems and inefficiency for any major NASA projects. We can see while civilian developed rockets are progressing fast, NASA's own rockets get endless problems and delays.
He asked if Toyota or GM can make a nice and reasonable priced car under such 50-state practice.
If there are too much political considerations and interferences in the semiconductor subsidy decision process, it will eventually kill the industry and business the government wants to help. IMO, the government should keep itself towards a venture investor instead of acting like a know-it-all backseat driver.
He said in order to get support from senators and representatives from all 50 states, NASA's supply chain spread widely into all 50 states. It creates enormous amount of problems and inefficiency for any major NASA projects. We can see while civilian developed rockets are progressing fast, NASA's own rockets get endless problems and delays.
NASA has never developed their "own rockets". NASA has always contracted with private companies to engineer and construct everything. The Space Shuttle, for example, was designed and manufactured by North American Rockwell. The current Space Launch System (SLS) was contracted to Boeing, and Boeing subcontracts many assemblies to other companies. NASA does have integration sites (like the famous Vehicle Assembly Building) and is responsible for system-level testing, but NASA is more akin to an airline than a space vehicle manufacturer or designer. NASA's lack of schedule integrity on the Artemis program has more to do with Boeing's lack of rigorous engineering and manufacturing than with politics. Numerous Boeing programs are a mess, from the Air Force One replacements to the troubled 737MAX and 777x programs.
What SpaceX has that Boeing doesn't is Elon Musk. Competence starts from the top down.
NASA has never developed their "own rockets". NASA has always contracted with private companies to engineer and construct everything. The Space Shuttle, for example, was designed and manufactured by North American Rockwell. The current Space Launch System (SLS) was contracted to Boeing, and Boeing subcontracts many assemblies to other companies. NASA does have integration sites (like the famous Vehicle Assembly Building) and is responsible for system-level testing, but NASA is more akin to an airline than a space vehicle manufacturer or designer. NASA's lack of schedule integrity on the Artemis program has more to do with Boeing's lack of rigorous engineering and manufacturing than with politics. Numerous Boeing programs are a mess, from the Air Force One replacements to the troubled 737MAX and 777x programs.
What SpaceX has that Boeing doesn't is Elon Musk. Competence starts from the top down.
Under the political considerations, US government (including NASA) often demands prime contractors (such as SLS, Boeing, or Lockheed Martin) to spread out their subcontracts to multiple states or multiple backgrounds of companies. There are multiple reasons that cause NASA's delays and cost overrun and political interference (such as the 50-state supply chain) is definitely one of them.
Under the political considerations, US government (including NASA) often demand prime contractors (such as SLS, Boeing, or Lockheed Martin) to spread out their subcontracts to multiple states or multiple backgrounds of companies. There are multiple reasons that cause NASA's delays and cost overrun and political interference (such as the 50-state supply chain) is definitely one of them.
I'm amazed. I'm not convinced it translates into NASA program lateness as you are, but I'm still amazed. (Boeing can seem to find a way to make anything late on its own.)
www.wsj.com
