Europe Lost the Cell Before It Lost the Car

Europe has spent three years arguing about Chinese electric cars. The more consequential contest was decided earlier and further upstream, in a business Europe entered late, funded generously, and largely lost: the manufacture of lithium-ion cells.

A car company that cannot buy competitive cells cannot build a competitive electric car, whatever its assembly costs are. Cells are the single largest cost in the vehicle, they set its range and charging behaviour, and they are the component with the longest and least forgiving industrial learning curve. Losing that layer is not one lost supplier. It is losing the ability to set the terms of everything built on top of it.

The European cell industry was supposed to be built by a handful of national and pan-European champions. The record since 2024 is bleak.

• •Northvolt, the Swedish company that carried most of the continent’s hopes, went bankrupt. Its factories were to have accounted for around 13 per cent of Europe’s planned 2030 battery production. Its Skellefteå gigafactory reached roughly 1 GWh of output against a planned 16 — a ratio that tells you the failure was in manufacturing execution, not in demand or in funding.
• •Automotive Cells Company, backed by Stellantis and Mercedes and the second-largest European effort, halted construction of gigafactories in Germany and Italy in June 2024, citing the need to move to lower-cost chemistry against slowing EV demand.
• •Cellforce, Porsche’s high-performance cell venture, also wound down.
• •Europe’s battery pipeline capacity out to 2030 fell by 176 GWh in 2024 alone, against an EU objective of covering 90 per cent of European demand from domestic production by 2030.

Meanwhile CATL began construction at Debrecen in Hungary of what may become Europe’s largest cell facility — a potential 100 GWh a year, with investment reported around €7.3 billion. Chinese manufacturers hold close to 70 per cent of global lithium battery installations, with CATL alone near 39 per cent.

It is tempting to explain Northvolt by pointing at management, and there is evidence for that. The more useful explanation is that cell manufacturing is a fundamentally different discipline from anything European automotive engineering was good at, and the difference is not obvious from the outside.

• •It is a process industry, not an assembly industry. Output is governed by coating uniformity, drying, calendering, moisture control and contamination — closer to semiconductor fabrication or pharmaceuticals than to a body shop. The skills do not transfer from carmaking.
• •Yield is the whole business. A line running at 70 per cent yield and one running at 95 per cent have similar capital costs and completely different economics. Nobody publishes their yield curve, so a new entrant cannot know how far behind it is until it is running.
• •The learning curve is empirical and slow. Cost falls because thousands of small process problems get solved in sequence, each one requiring a failure to learn from. That takes calendar time that money cannot compress.
• •The equipment and materials supply chain is itself concentrated. Much of the machinery, and the refined cathode, anode and electrolyte inputs, come from the same region as the competition — see our note on the refining chokepoint.
• •Scale-up is not a bigger pilot. Processes that behave at 1 GWh routinely fail at 16, and the failure modes only appear at volume.

Northvolt’s 1 GWh against a planned 16 is precisely this pattern. The company could make cells. It could not make them at rate, at yield, and at a cost anyone would pay — which is the actual product.

There is a second cause, and it is policy rather than engineering. A gigafactory is a multi-billion-euro bet on a demand curve five to ten years out, and European EV demand grew more slowly than the plans assumed while the regulatory target that underwrote those plans was itself being renegotiated.

The softening of the 2035 combustion engine deadline into a 90 per cent fleet CO₂ reduction with technology neutrality — covered in our piece on what that revision actually does — is a rational response to industrial distress. It is also, from a cell investor’s point of view, the demand floor moving after the capital was committed. One analysis put the capacity implication of target rollback at the equivalent of dozens of Northvolt-sized factories.

This is the trap in industrial policy that uses a regulatory deadline as the demand guarantee. The deadline is only credible while the industry it disciplines is healthy enough to meet it, and it gets relaxed exactly when the factories most need it to hold.

It would be wrong to write this as a collapse. Europe retains real assets: a large and wealthy vehicle market, deep engineering in power electronics, thermal systems, safety certification and manufacturing quality, strong recycling regulation that will matter more as fleets age, and a battery regulation framework that others are copying.

It also retains the ability to host production, which is what the CATL and BYD investments represent. Hosting is genuinely worth something — supplier ecosystems, trained process engineers and tacit knowledge do diffuse locally, which is roughly how Japan and Korea seeded parts of their own industries decades ago. It is simply a slower and more dependent path than the one that was planned.

India is attempting the same thing Europe attempted, later, with less capital and a similarly concentrated upstream. The European experience is the most expensive available case study, and it argues against a few comfortable assumptions.

• •Announced capacity is not capacity. Europe’s pipeline shrank by 176 GWh in a single year and its flagship reached 6 per cent of its target output. Judge a cell programme by shipped, qualified, warrantied cells, not by memoranda of understanding.
• •Subsidy does not substitute for process capability. Northvolt was not short of money or of political support. It was short of yield, and no incentive scheme produces yield.
• •Partnering with the incumbent is not defeat. Hungary will host Europe’s largest cell plant because it accepted Chinese investment. The knowledge transfer argument for doing the same is stronger than the sovereignty argument against it — provided the terms actually require local engineering depth rather than a screwdriver plant.
• •A stable demand signal is worth more than a large one. Investors need a floor they can underwrite for a decade. Repeatedly revised targets and subsidy schemes that lapse and restart do more damage to a factory decision than a modest incentive would fix.
• •Pick the fights that are winnable. Cell chemistry leadership is a hard fight from behind. Pack engineering, BMS, thermal design, safety qualification, refurbishment and recycling are adjacent, less capital-hungry, and closer to where Indian firms already have people.

Our longer treatment of the Indian position is in what catching up would actually require, and the upstream constraint that binds everyone is in the refining chokepoint.

Europe’s mistake was not trying. It was assuming that a large market, serious money and a regulatory deadline would substitute for fifteen years of accumulated process learning. They do not, and the price of finding that out was paid in full.