EU should put the brakes on its chips strategy
Driving high-end production back home may do little to ease shortages that have hit growth
Hello from a rainy and increasingly grey Frankfurt, where we’re covering the latest round of results from Germany’s big industrial players.
Among a set of rosy figures from premium carmakers, BMW’s have stood out. While rival Daimler saw quarterly sales drop 32 per cent because of the supply chain crisis, the Munich-based brand only suffered a 12 per cent drop in the same period, and registered a whopping 42 per cent rise in net profits.
Boss Oliver Zipse told all who would listen that this was down to BMW’s superior supply chain management — a view echoed by analysts. It somewhat puts paid to the idea — in vogue in Brussels and parroted by the auto lobby, the European Automobile Manufacturers’ Association (ACEA) which Zipse chairs — that proximity to semiconductor foundries has anything to do with securing supply of the scarce chips.
This vexing topic is the focus of today’s main piece. We’ve been out to meet the chipmakers with existing bases in Europe — such as Infineon, GlobalFoundries and Bosch — and found few who are convinced the EU’s plans to reshore production will achieve its most vocal proponents’ aims. Given the subsidies on offer, they are unlikely to voice their scepticism out loud, at least not before it is too late.
Brussels’ chips plan won’t solve carmakers’ woes
Earlier this year, roughly around the time that the chief executive of a German premium manufacturer assured the Financial Times that his company had “full visibility” of its supply chain, a sales executive at Europe’s largest semiconductor site received a call.
On the line to the Dresden branch of GlobalFoundries was a purchasing manager from a large carmaker — a rare occurrence, given that auto groups generally leave their tier-one suppliers to look after the mundane business of semiconductor procurement.
The question from the car executive, voiced with a nervous tone, foretold the crisis that was about to envelop the entire sector. “Would you consider running your fabs overnight and on weekends?”
Those familiar with the semiconductor industry know that its profitability hinges on “utilisation” rates, or how often a company can run its factories around the clock. In Europe, where personnel and energy costs are high (GlobalFoundries’ Dresden facility uses the equivalent electricity to roughly 200,000 households), this is especially true.
No fabrication plant survives for very long with idle capacity. In fact, large amounts of research-and-development spend is geared towards reducing the amount of downtime needed for equipment repairs. After a short, pandemic-induced lull, GlobalFoundries has been running its fabs at full pelt from August 2020, its boss recently told CNBC.
The car executive’s ignorance characterises, insiders say, the ill-informed response to the chips crisis from corners of both European industry and politics.
A shortage of semiconductors has hit Europe’s manufacturers hard. Nowhere more so than in the region’s economic engine, Germany, where automakers contribute sizeable chunks to gross domestic product growth. The latest official figures show German manufacturing output fell 4 per cent in August compared with July, with production of motor vehicles and trailers slumping almost a fifth because of supply chain woes.
The shortage has also underscored the degree to which the rest of the world relies on Taiwanese foundries for crucial chips. One hardly has to be an expert on geopolitics to appreciate that the territory is not the most stable place on earth.
In September, the European Commission acted. President Ursula von der Leyen proposed a “European Chips Act”, arguing that reshoring production of semiconductors, the vast majority of which are manufactured in Asia, was a matter of “tech sovereignty”.
While EU internal market commissioner Thierry Breton helpfully clarified that “the idea is not to produce everything on our own here in Europe”, (good luck making substrate manufacturing economical in the developed world) it remains the EU’s ambition to double its semiconductor market share by the end of the decade, by supporting the building and expansion of cutting-edge fabs on the continent.
Trade Secrets has long had its doubts about the strategy.
The problem with this plan, as we see it and as a cogent piece from Niclas Poitiers at Bruegel points out, is threefold.
The first, as underscored by the boss of Europe’s largest semiconductor company, Infineon, is that the biggest buyers of chips are elsewhere. The automotive sector accounts for roughly 8 per cent of the total semiconductor market. No amount of proximity to the car industry, for example, will make up for the fact that Apple and Samsung and the like are at least eight places ahead in the pecking order (as a purchasing executive at VW put it to us).
Second, the amount of investment needed is eye-watering. Taiwan’s chips giant TSMC will spend $100bn to expand its footprint over the next three years. The EU expects to invest up to €30bn by 2030. As TSMC’s American founder put it to our colleagues at Nikkei, “even after you spend hundreds of billions of dollars, you will still find the supply chain to be incomplete, and you will find that it will be very high cost, much higher costs than what you currently have”. That EU funding pool will have to get a whole lot bigger if it is to make much difference.
Third, and most importantly, there is a mismatch between the type of investment that interests the EU — the most advanced chip production facilities — and the semiconductors that Europe’s automakers are desperate to procure.
As part of its “2030 Digital Compass” plan, the bloc says it wants to build “manufacturing capacities below 5nm nodes, aiming at 2nm and 10 times more energy efficient than today”.
But two-nanometre chips are hardly used by carmakers. Modern cars are filled with hundreds of much larger chips, mostly low-margin power, radar and memory semiconductors, which make the windows go up and down and help the GPS work. In a two-tonne, 4.5-metre long SUV, chip size and weight are hardly a factor. Cost is.
“We have nobody who makes phones, we have nobody who makes PCs,” Jalal Bagherli, chief executive of Dialog Semiconductor, told the FT, referring to Europe. Investing in bleeding-edge processing chips, he added, “doesn’t make a lot of economic sense to me”.
That hasn’t stopped those who see subsidies coming down the line laying out their stalls. In September, car registrations in Europe dropped almost a quarter to the lowest level in more than 25 years. “In order to avoid this from happening again in the future, Europe also needs to come up with a strategic plan to increase the production of semiconductors in the EU,” the industry’s Brussels lobby, the ACEA, said last month.
Yet, as we mentioned earlier, their current president seems to have managed the crisis well despite relying on Asian suppliers — suggesting that a better understanding of the semiconductor market, and one’s place within it, is a far more effective solution than splashing the cash from Brussels.
“We know exactly who our suppliers are, not only on the tier-one level but right down to the raw materials,” Zipse told journalists on Wednesday. While he declined to comment on whether BMW was paying more for its chips than others, he did say that the company “always had a fair agreement in terms of how we deal with suppliers . . . that’s paying off now”.
As well as being somewhat misdirected, there is every chance that Breton and the EU’s ambitions will backfire.
Automotive companies are already considerably over-ordering chips. Infineon’s chief told the FT last month that “current orders look more like 110m to 120m cars [per year]”, when 80m-odd are expected to be sold in 2022, according to data firm IHS Markit.
Stacy Rasgon, a forensic semiconductor analyst at Bernstein in the US, said chip orders by automotive customers were 42 per cent “above trend” last quarter, meaning above the amount of cars actually being produced.
Investing in semiconductor capacity for political “bragging rights” (as one semis executive termed it) is risky enough. Investing in it based on current demand might be detrimental to existing Europe-based fabs, leaving them with the surplus capacity they so desperately try to avoid.
In its prospectus before going public this week on the Nasdaq in New York, GlobalFoundries had a telling warning to investors about its business model. “If we overestimate customer demand . . . we could experience underutilisation of capacity at these facilities without a corresponding reduction in fixed costs.”
We doubt that’s a message well understood by their customers in the car industry, or lawmakers in Brussels.
