What happened at Northvolt?

Set the epistemics first, because they govern everything else. Kann opens by saying that neither he nor Jaffe has particularly inside information and that they will be working from the public domain. Almost every causal claim below is therefore inference from public reporting, from Kann’s conversations with industry contacts, and from one employee account. The founding facts are firmer. Peter Carlsson, a Tesla battery executive who helped build the original gigafactory, left and founded Northvolt around 2016 to build a European-headquartered cell producer on equal footing with the Asian incumbents, which in 2016 meant LG, Samsung SDI, to a lesser degree SK, and Panasonic, with CATL and BYD already present but not yet dominant. Jaffe thinks the founding thesis correctly anticipated China’s rise, which he calls the most poorly hidden story of the last decade, running the same playbook that took solar manufacturing east. The money followed: public reports of something like 15 billion euros, which Kann immediately qualifies as not all awarded and partly debt, and which Jaffe puts at roughly $14 billion covering capital actually raised, future commitments and tax credits, with tax credits a large share. The headline figure is a bundle of different things, and neither speaker treats it as a single raised number.

Jaffe’s reading is that the Western venture model plainly can raise that much money; the question is what comes attached to it. In his account the investors wanted Northvolt to be huge, and huge meant not one factory but six gigafactories simultaneously, which he calls an impossible task to handle. Kann draws a distinction worth preserving here: building one very large factory was not the error, because the economies of scale in cell manufacturing make that unavoidable. The error was everything around it. A second factory in Germany to serve German automakers, a third in Eastern Europe, a fourth in North America once the Inflation Reduction Act made US subsidies available, the acquisition of Cuberg with its lithium metal aviation batteries and its own planned US plant, a large California research facility, and a sodium-ion program with yet another planned factory. Three different battery types, different technology generations, three continents. Meanwhile the demand side was genuinely solid: realistic purchase orders from European carmakers including Volkswagen and BMW, which is exactly what a manufacturer wants, and worth nothing if the cells never get made.

They did not get made at anything like the intended rate, and this is where the episode does its real work. Jaffe’s framing is that a cell plant is not one factory but roughly eighteen mini-factories in sequence, each doing something different, all of which have to be choreographed and timed against each other, and the underlying process is chemistry rather than assembly. Every battery factory he knows of has struggled to ramp, including Tesla’s first gigafactory, where he says close to half the first year’s cells had to be scrapped, hedging the figure with “apparently.” Northvolt had it worse than most, and both speakers converge on the same explanation: the choice of Wuxi LEAD as main equipment supplier, and the relationship with them. The conundrum Jaffe describes is real and not obviously avoidable, since very few companies can supply equipment for every one of those stages and very few of those are outside China. Wuxi LEAD was to build the equipment, install it and make it work, and Jaffe’s judgment, explicitly hedged, is that the third step broke. Kann’s independent sourcing points the same way: contacts with as much battery manufacturing experience as anyone universally focused on that relationship, and a Northvolt employee’s blog post, written in Swedish and later translated, put it at the center too. Jaffe adds two pressures that are not about equipment at all. Northvolt committed to a high-nickel NMC cathode, which he describes as not wrong but less of a no-brainer once lithium iron phosphate re-emerged around 2018 and 2019 and European carmakers began reconsidering their chemistry choices; a built factory cannot simply switch cathodes. And European electric vehicle demand slowed, which he pointedly declines to call a decline in demand. Against that backdrop, Kann says, BMW canceling a $2 billion order was what visibly accelerated the company’s downturn, for quality reasons as far as he understands it.

The turn into the lessons comes through institutional knowledge, and Jaffe’s answer is more forgiving than the framing around it. Nobody knew how to build a battery factory at gigafactory scale until Tesla built one in 2014 and 2015; that knowledge had to be invented, and now exists mostly in China and Korea. So the easy criticism, that faking it until you make it fails if you fail while faking it, is one he rejects: someone has to invent the knowledge, and he points at the fusion industry as a field in exactly that position today, where he thinks the Silicon Valley model is probably the right way to shepherd one or two companies to success. His complaint is about focus, not ambition, and his counterfactual is explicit: if the investors had pushed Northvolt to do anything necessary to make that one factory work and leave everything else for later, he thinks it could have worked. The wrong lesson, which both reject flatly, is that batteries cannot be built in Europe or outside China. Europe has close to 100 gigawatt-hours of manufacturing today and the US will have close to 300 within a few years; labor, energy and capital cost more outside Asia, and that is simply what commerce costs. Jaffe’s right lesson is the venture model: investors who wanted a 150 gigawatt-hour company instead of a 40 gigawatt-hour one, and six factories at once, were the fundamental problem. Kann accepts the description of events and rejects the diagnosis. Taking big swings, he argues, does not inherently mean pushing a company to do many things simultaneously, and he can imagine the same investors funding the same company on a build-it-prove-it-then-scale path. His own lesson is narrower and about the sector: battery manufacturing is hard, so the more novel what you are doing is, whether novel in technology or merely the first of its kind in Sweden, the more focus is the thing that exposes you to compounding risk if you lose it. He also puts some agency back on management, noting they clearly believed they had manufacturing in hand well before they did. Jaffe narrows his own claim in response rather than abandoning it: he is not saying venture capital does not work, but that at those funding levels its weaknesses in shareholder expectations and definitions of winning get exposed.