In parts of southeast Asia, white elephants were long considered a symbol of prosperity because they were expensive to maintain and served little practical purpose. It’s an apt metaphor for Europe’s latest plan to catch up with the US and Asia in technology.
The European Union is considering the construction of its own advanced semiconductor manufacturing facility, Bloomberg News and others reported on Thursday. The ambition to make next-generation silicon chips is admirable, but it’s also wildly misplaced. Yes, these underpin most of today’s cutting edge technologies — smartphones, virtual reality goggles, data centers and plenty more besides. But there’s simply not enough demand for these particular chips in Europe, and there isn’t likely to be any time soon.
In chipmaking, three countries dominate the market: the US, Taiwan and South Korea. Because their domination extends back decades, an ecosystem of suppliers, customers and skilled workers has developed around them. It means the big players, particularly Taiwan Semiconductor Manufacturing Co. and South Korea’s Samsung Electronics Co., can invest in new production facilities with the confidence that they will not only be able to source the components and employees they need, but also that customers will be waiting on their doorstep.
European chipmakers such as Infineon Technologies AG and NXP Semiconductors NV also rely on them. But events of the past two years have heightened fears about Europe’s technological sovereignty and the idea that it is too dependent on imports of critical technology. First the US-China trade war highlighted the precariousness of the global supply chain. European companies like Nokia Oyj and Ericsson AB, for instance, had to develop contingency plans to shift production away from their Asian manufacturing facilities.
Then came the virus. Many of Europe’s carmakers rely on Infineon and NXP — chipmakers that do have their own factories (called “fabs” in the lingo) but also outsource some fabrication to TSMC. Due to the pandemic’s capacity constraints, the Taiwanese firm, which also supplies Apple Inc., got so busy churning out widgets for iPhones that it couldn’t produce chips for the European auto companies trying to ramp up production.
Those sort of perturbations seem to have sent a chill down lawmakers’ spines, not least that of Thierry Breton, the European industry commissioner. As we head toward a world where everything is connected by 5G networks, reliable access to chips seems even more crucial. Breton’s response is to seek the construction of a so-called foundry, a semiconductor facility capable of making different chips for different customers. The bloc wants a foundry capable of supplying the newest tech: semiconductors with circuit patterns less than 10 nanometers wide.
The problem with this plan is that European factories don’t make the devices that need such chips. There is some demand, of course, but it can usually be supplied from outside the region. Europe just doesn’t have much of an electronics industry anymore. Many of the products it does make — industrial machinery, cars, telecoms equipment — are pretty big, so in a rudimentary sense, it matters less whether chips rely on 10-nanometer or 25-nanometer technology, because space is less precious than on a smartphone.
What’s more, foundries are expensive to build — between $20 billion and $30 billion a pop. And they need additional investments of billions of dollars every two to three years to keep pace with technological innovation. To get a return on that investment, the production lines need to run 24 hours a day, seven days a week, according to Jean-Christophe Eloy, a semiconductor consultant and the chief executive officer of Yole Developpement. Europe is a more expensive location to operate too.
Without sizable demand, it makes little business sense to invest in a European foundry. “It’s like building a cathedral in the desert,” Eloy told me.
The EU is in discussions with both Samsung and TSMC about the foundry project, Bloomberg News reported. But it’s hard to imagine either company proceeding without massive EU subsidies. And such funds would be better spent elsewhere. Fostering an electronics industry that might create subsequent demand for a foundry would be a good place to start. Or investing in alternative chip tech — an area where the region is already forging a lead.
Technological sovereignty is an admirable goal. But Europe doesn’t need a $30 billion white elephant.
Bloomberg
