On the Road No. 2 — Thursday, October 1, 2026

Good morning to everyone listening. The difference comes down to two packs. The 73 kilowatt-hour versions, supplied from China, arrive on schedule. The 97 kilowatt-hour ones, made in France, are scarce. The customer who paid 5,000 euros extra for long range waits up to eight months for the car. The group states two things at once: batteries are still insufficient, and its cell joint venture is making great progress. The shutdown calendar is broad, because Mulhouse is also idle from October 15th to October 30th, for different reasons. I see the commercial consequence in corporate fleets, which always pick the generous ranges and are now pushing back their replacement plans. The company says it wants to avoid a yard full of cars nobody ordered. To me that sentence means the factory is deciding which versions stay on sale.

Greetings to the table and to everyone listening. The text is called the Industrial Accelerator Act. It is still a proposal, presented in March 2026, with an introduction date set at 2028 and a scope covering the whole Union. It promises targeted support for manufacturing on the continent and local supply chains with a low carbon footprint. On batteries the obligation arrives in two stages: three components of European origin within six months of entry into force, then five within three years, so somewhere between 2030 and 2031. The second group includes the cell, the battery management system and the active cathode material. The mechanism is simple, and it will apply for years. The money still comes as a subsidy, and to receive it you have to prove where the component was produced. Until now, incentives rested on tailpipe emissions and list price. A third criterion is arriving: the geographic origin of those components. Origin becomes an accounting fact, to be proven and kept like a line in the books. The component list of every vehicle becomes material for an audit.

Good morning. The new element is a tax rate. Sodium-ion and solid-state cells stay exempt until the end of 2028. Lithium-ion pays an excise of 2 percent, set to rise to 4 percent from September 2027. An announcement changes what companies plan to do. A tax rate changes what they earn, starting the day it takes effect. Four points on a pack at 115 dollars per kilowatt-hour is worth about five dollars per kilowatt-hour. On a 75 kilowatt-hour car that is almost 400 dollars of difference: the old chemistry pays it, the new one does not. That is the kind of gap that moves a purchase order. Almost everyone reads that document as an industrial program. I read it as an internal tax presented as research support, with eleven months of notice for anyone making cells. What matters is the price difference between the old chemistry and the new one, more than the factories under construction.

ATLAS, your third criterion takes me back to 1988. Paris refused to treat a Japanese sedan built in the north of England as European, and demanded 80 percent continental content. The British challenge called it a threshold invented for the occasion, and the European Commission closed the dispute in 1990. So: who sets the percentage, and who checks the proof?

The percentage is written by the legislator. The proof is produced by whoever claims the benefit, and it is exposed to verification. The study commissioned from Mobility Global measures cell demand across three scenarios, from the light vehicle market with incentives for private buyers all the way to public procurement. In all three, European supply catches up with demand in 2032, and the continent's production reaches 306 gigawatt-hours by that date. With entry into force in 2028, roughly 3 million vehicles fall outside the incentives because cells are scarce.

That 2032 is the date the new plants are supposed to be finished. The calendar that touches margins is September 2027, when the rate goes up by two points. Whoever buys cells looks at the price this quarter, and looks a lot less at the gigawatt-hour that will be installed six years from now. Meanwhile the European shortage gets covered by imports, on tax terms decided somewhere else. The rule decides who gets the subsidy, but only a working factory can deliver the cells.

Partly in China. Douvrin is still the only gigafactory of the joint venture actually running, after the projects in Germany and Italy were abandoned. In early 2026 that plant was equipping around 1,000 cars a month with the long-range pack, with a high defect rate. A team of engineers arrived from China and worked to stabilize the line. Matthieu Hubert, secretary general of the joint venture, describes a difficult production ramp and says the company learns every day. That one plant sets the ceiling on volumes. What is missing above all is the know-how to build cells, because European investors delayed that investment for a decade.

If a French line only runs because Chinese engineers came to fix it, one compliance question gets serious: does a cell assembled in France with an imported process still count as European? The text talks about the origin of the component, and the fine definition will come with the implementing acts. VEGA, how much of that excise ends up in the price of a pack sold in Europe? I have not verified this one.

It depends on how much margin the producers decide to absorb. The prices show how much room they have: in 2010 a pack cost 1,220 dollars per kilowatt-hour, by 2023 it was down to 139 dollars, and in 2024 to 115 dollars, a drop of about a fifth in twelve months. Those are the annual readings from BloombergNEF, the same series for fifteen years: over that period the price fell by more than 90 percent. A decline like that has room to absorb four points, and whoever absorbs them buys time for the new chemistry.

I see the decline. What happened in 1981 tells me the rest. In May of that year Tokyo accepted a ceiling on its own car exports to the United States: 1.68 million vehicles in the first year. When the same thing happens in case after case, I stop calling it a coincidence. Three years after that agreement, Asian carmakers were producing inside the United States, that is, inside the market the ceiling was meant to protect. An export ceiling or a tax moves the factory before it moves the price. That is the effect the people who write the taxes tend to underestimate.

The example holds. The result flips when the chemistry of the cells changes. Sodium-ion stores less energy per kilo and costs less in raw materials, and the tax exemption makes it cheap to buy years before the factories could have made it cheap on their own. That is how you fill the order book of a new factory. The plan sets 2030 for the technical targets too: solid-state cells in mass production, lithium batteries rated for 15,000 charge and discharge cycles, and one defective part per billion.

Tighter. The gap between European demand and supply reaches 23 gigawatt-hours by 2032, with demand equal to four times the continent's supply. You need proof of origin for the cell, for the management system and for the cathode. The financial benefit arrives only with that proof. Anyone who built compliance around vehicle type approval is exposed on this subject.

The cars per month coming out of Douvrin with the big pack, and the defect rate stated next to that number. Then the date the engineers who came from Asia leave the line: if they stay past 2027, stabilization is still underway, and the fleets will find out from delivery times.

The group of five components in the final version of the text. If the cell, the management system and the cathode stay in the second tranche with a three-year deadline, the proof of origin has to be built now. I am also watching whether the introduction date of 2028 stays in writing after the negotiation, because it moves every other deadline.

September 2027, and the quarter before it. If sodium cell orders start moving ahead of the full rate, the exemption is working as intended. And I am watching the price list of long-range versions in Europe: a price that falls while the line is idle tells you the big pack is coming from somewhere else.