Europe’s grand plan for industrial rearmament and tech sovereignty is at risk of disintegration on multiple fronts. Critical components are proving impossible to deliver.
The semiconductor renaissance is going nowhere. The lynchpin was supposed to be a €30bn (£25bn) project by US chip giant Intel to build two world-class “fabs” near Magdeburg in eastern Germany, a third paid for by German taxpayers in the most expensive undertaking in the history of Deutschland Inc.
These Intel fabs were to make chips down to the technological frontier of 1.8 nanometers (nm), soon to be de rigueur for AI, 5G, autonomous driving, and advanced weapons. They were to be the beating heart of Silicon Saxony, and Europe’s hope of playing itself back into the chip game after two decades of decline.
Intel should have started construction in early 2023 but was held back by wrangling over state aid and by surreal disputes over a neolithic burial site and what to do with the local “black” soil. The company said on Monday that the whole project – and another site in Poland – is on ice for another two years.
Intel is itself in trouble after missing the smartphone boom and the first stage of the AI boom. It is shedding 15,000 workers worldwide and retreating to Fortress America, focusing on four advanced fabs already in Arizona and Ohio under the spur of the $280bn (£212bn) Chips and Science Act – Washington’s national security plan to restore US self-reliance across the semiconductor spectrum.
Europe produced a quarter of the world’s chips in 1990. This has since fallen below 10pc. Almost none of it is at the cutting edge below 10nm.
The EU has passed its own Chips Act, pledging a non-existent €43bn to capture 20pc global share by 2030. World demand is expected to double by then, so the EU must quadruple its output. “It is totally unrealistic,” said Peter Wennink, ex-head of the Dutch lithography group ASML.
True funding at the EU level is just €3.3bn, and some of that comes from cannibalising Horizon Europe (science). It would take at least €500bn to reach scale and the 2nm threshold from a standing start, even if the EU had all the specialist skills, which it does not.
Wage costs are 40pc lower in Taiwan, and the island has a nexus of technical institutes geared to the industry. “The capabilities required to domestically innovate this technology are virtually non-existent in the EU,” said last week’s Draghi report on EU competitiveness.
Europe’s electricity prices are 158pc higher than in the US on average, and large fabs consume 100 megawatt per hour each. It is surprising that Intel was ever tempted.
In my view, Germany should consider itself lucky that Intel has halted this exorbitant prestige project, which was going to cost €3.3m in subsidies for each permanent job, only to replicate chips that can be bought on the open market from allies.
Furthermore, these 1.8nm chips may themselves be heading for obsolescence before long.
There is a limit to how far you can miniaturise silicon circuits. Advanced materials such as graphene and gallium nitride may soon leap-frog ahead. Cambridge start-up Paragraf is developing 2D graphene chips for sensors that are one-atom thick and a thousand times faster.
The UK is concentrating its £1bn semiconductor fund on niche areas where it has an edge. The EU should be doing the same, thinking like a mid-sized power rather than indulging in great power illusions. It should leverage its core strengths in sensors, lithography, optics, or quantum computing.
Europe’s other grand plan for battery gigafactories is scarcely in better shape. BMW has cancelled a €2bn order for lithium battery cells from Sweden’s Northvolt, Europe’s best-funded tech start-up and the great hope of the car industry. It could not deliver the cells in time. The contract will go to Samsung SDI in Korea.
Northvolt is an excellent company, and at least it makes sense to build a green gigafactory in northern Sweden where hydropower delivers Europe’s lowest electricity costs. But the company got ahead of its skis in a brutal world market.
It bet on standard NMC batteries made with nickel and cobalt just as China switches to cheaper and safer lithium iron phosphate (LFP) batteries for the mass market.
Northvolt is having to take drastic measures, halting cathode production at its core plant in Skellefteå. It will have to buy the cathode material from Asia. Other plants in Germany, Sweden, and Canada are under review. The Swedish state has refused a state rescue, leaving the company in talks with creditors. Such are the woes of the meteor once billed as Europe’s new Airbus.
Norway’s Freyr has given up trying to make EV batteries in Europe, switching to America to profit from the Inflation Reduction Act, although that is not plain sailing either.
Only one of Volkswagen’s six gigafactories has progressed beyond the drawing board. The company has axed a proposed plant in Saxony and is building just one of its two planned cell plants in Salzgitter.
“Western carmakers are just passengers travelling at the back of the bus. Others decide where it goes,” said VW board member Thomas Schmall.
“Batteries are a core technology of EVs but today the car industry is totally dependent on Asian battery suppliers. We must change that,” he told the Frankfurter Allgemeine.
Note that he did not join the political backlash against EVs and warned that any delay in the 2035 combustion ban would be fatal. “We all agree that the future belongs to e-mobility. There is no alternative”
Volkswagen is going through its own corporate hell as it pays the price for letting China steal a march on EV technology. But the immediate problem is that China ramped up battery capacity last year to 800 GWh, more than the entire global demand. It will have tripled again by late next year. This galactic excess is landing in Europe.
One European battery-maker said privately that the EU had not offered his company “a single inch of flexibility” on financing, which is extraordinary after all the talk of the European Battery Alliance. But that is the point. The EU’s lofty declarations have no serious funding.
Mario Draghi, economist and former Italian prime minister, is right to call for a double Marshall Plan of €800bn a year in extra investment to make Europe fit for the 21st century. But to do that the EU needs its own Hamiltonian treasury with the full borrowing powers of a unitary state. Such a Europe does not exist.
Either the EU grasps the nettle and goes the whole way with radical treaty change or, more likely, given the political currents in Germany, it devolves economic and legal power back to the nation states. The hybrid status quo is demonstrably failing.
Europe’s quest for ‘industrial sovereignty’ has gone horribly wrong
A semiconductor renaissance in the Continent is going nowhere