What Happened to Northvolt? Why Europe's Battery Champion Failed

Northvolt failed because its first factory could not make enough good battery cells at a competitive cost, while the company spent its capital expanding into new sites before that factory worked. It had customers, with orders it put at more than $50 billion, and it had money, more than $13 billion of equity and debt (reported). What it never had was manufacturing yield. It filed for Chapter 11 in Texas on 21 November 2024 and for bankruptcy in Sweden on 12 March 2025, and its main plants passed to Lyten, a California start-up, in February 2026.

When Northvolt filed in the US Bankruptcy Court for the Southern District of Texas, it had about $5.8 billion of debt and about $30 million of cash, enough to run for about a week (reported). Its co-founder and chief executive, Peter Carlsson, stepped down. Scania, a truck maker that was both a customer and part of its largest shareholder's group, lent $100 million to keep the lights on (reported). Eight years earlier the company had been founded to give Europe its own answer to CATL.

What was Northvolt?

Carlsson had run supply chain at Tesla from 2012 to 2015. He and another Tesla alumnus, Paolo Cerruti, founded the company in 2016 to build a European battery champion in Skellefteå, a town of about 75,000 near the Arctic Circle with cheap hydro and wind power. The pitch had three parts: green cells made with low-carbon power, security of supply for Europe's carmakers, and a domestic industry that the European Commission, Sweden and the European Investment Bank all wanted.

Investors and customers lined up. Volkswagen took about 21%. Goldman Sachs funds, BMW and Nordic pension funds joined. Volkswagen, BMW, Volvo and Scania signed long-term supply agreements that together were reported at more than $50 billion. The first cell came off the line at Northvolt Ett, the Skellefteå plant, on 28 December 2021.

Why did Northvolt fail?

Northvolt's failure comes down to one number that rarely appears in press releases: the share of cells started on the line that come out good enough to sell. That is manufacturing yield.

Yield is the business

A battery plant earns the price per kWh less the cash cost per kWh, multiplied by the good kWh it ships. Materials, the cathode, anode, electrolyte, separator and foil, are most of the cost of a good cell, and a scrapped cell consumes the same materials as a good one. A plant that scraps 30% of what it starts pays about a third more for materials per good cell than one that scraps 5%, before labor, energy and depreciation are spread over fewer cells. In the research's stylized example, $60 per kWh of materials started becomes about $63 per good kWh at 95% yield and about $86 at 70%, a gap larger than a typical cell maker's whole gross margin.

Korean and Chinese cell makers had spent ten to twenty years climbing that yield curve. Northvolt proposed to do it in three or four, in a remote town where it had to recruit thousands of people, with production equipment bought largely from Chinese suppliers and commissioned with the help of their technicians (reported). The process knowledge sat with engineers who would leave. Ett's output stayed a small fraction of its 16 GWh first-phase capacity in 2023–24 (reported).

Too many sites before the first one worked

While Ett struggled, Northvolt kept building. It announced a 60 GWh plant at Heide in Germany and another in Quebec. It added cathode production, recycling, a storage-systems plant in Gdańsk, a joint venture with Volvo and a lithium-metal subsidiary in California. In January 2024 it took on about $5 billion of new debt to expand Ett itself (reported). Each project could be defended alone. Together they spread scarce engineers and cash across sites before the first plant had proved its yield. The group reported a net loss of about $1.2 billion on revenue of about $128 million in 2023 (reported).

An order book that was really a set of options

Northvolt raised capital on its order book. That book was weaker than it looked. An LNG buyer under a take-or-pay contract owes a fee whether or not it takes a cargo. A carmaker's cell contract has no fee for a product that has not been built: if cells arrive late or fail qualification, the buyer can delay, cut or cancel. BMW did exactly that in June 2024, cancelling a contract worth about €2 billion after delays (reported). And because Northvolt's largest shareholder was also its largest customer, the customer's patience and the shareholder's patience were the same patience.

A real advantage that did not matter enough

Northvolt's low-carbon power was a genuine edge. But power is a small share of a cell's cost compared with materials and yield, and in 2023–24 carmakers ranked price and on-time delivery ahead of carbon. A second-order advantage cannot rescue a first-order cost problem.

What a slow ramp costs

The book's chapter on building an energy company works through a stylized cell plant to show why yield is a capital problem. Take a 20 GWh plant costing $1.6 billion, with $60 per kWh of materials, $300 million a year of fixed cash costs and good cells selling at $90 per kWh. The numbers are illustrative, not Northvolt's.

The slow plant needs about $660 million more capital before it turns positive and, five years in, still has a higher cost per good kWh than the fast plant had in year three. A slow ramp consumes capital, delays the learning that would lower cost and gives customers time to cancel. Building a second and third plant during that ramp multiplies the funding need before anyone knows whether the first will ever reach the leader's cost.

How did CATL and First Solar do it differently?

The most successful manufacturers in the book reversed Northvolt's order. CATL's founders, led by Robin Zeng, had spent more than a decade making lithium-polymer cells for phones and laptops before spinning out the power-battery business in Ningde in December 2011 (reported). Its first big customer, BMW Brilliance in 2012, is widely told as the moment it learned automotive-grade quality, because BMW's audits forced it to build processes it later sold to every carmaker (reported). By 2025 it was selling about 661 GWh a year.

First Solar followed a similar path in solar. Its founder, Harold McMaster, began developing cadmium-telluride panels in Toledo, Ohio, in 1990, and patient owners funded years of work on the hardest step, depositing an even layer at speed across a full panel, before the first line reached commercial output around 2002. In the fourth quarter of 2008 it became the first module maker to report manufacturing costs below $1 a watt (company). Its later position as the largest US-headquartered module maker rested on twenty years of process learning.

What they shared was a sequence: prove the process at one site, with a demanding customer, under an owner who could wait, and only then multiply.

What happened to Northvolt's factories?

The assets were split up. The industrial-battery unit went to Scania, completed in April 2025. Lyten agreed in August 2025 to buy Ett, its expansion and the Västerås research site, and completed the Swedish deal on 26–27 February 2026 (Lyten). The German Heide plant was left out of that closing; Lyten signed an exclusive memorandum for it on 1 July 2026, with a definitive agreement targeted for the third quarter, unconfirmed as of early November 2026. Lyten said it would restart Ett line by line and targeted commercial cell sales from it in the second half of 2026 (reported).

Equity holders were largely wiped out, and Volkswagen and Goldman funds took write-downs (reported). Lenders including the EIB and Nordic banks faced losses (reported). Several thousand Skellefteå jobs were lost, and Europe's battery policy lost its flagship.

Which factor of the power equation failed?

The Power Equation holds that an energy company creates durable value only when four factors hold at once: supply at a cost the market will pay, a route to market, capital matched to the asset, and a licence to operate. Northvolt is the clearest modern case of the first factor failing while the other three looked strong.

It had a route to market in an order book few start-ups have matched. It had capital from carmakers, pension funds, banks and two governments. It had a political licence most industrial companies could only envy. What it never had was a cost per good kWh that could compete with the leaders, and its abundant capital let it multiply that problem across sites before solving it at one. Plenty of capital did not make up for the missing cost advantage. It made the loss bigger.

The pattern is older than Northvolt. A123 Systems had a better chemistry and a federal grant and still went bankrupt in 2012; Solyndra had a federal loan guarantee and a cost curve moving against it. Our roundup of the biggest energy bankruptcies sets them side by side, and our guide to how the energy industry makes money shows where manufacturing sits in the value chain. For the wider battery market, see the book's study of energy storage and batteries and Chapter 10 on batteries, storage and electric vehicles. Company detail is in the Northvolt profile.

The takeaway

Northvolt did not fail for lack of demand, money or political support. It failed because cell-making is a yield business, yield is learned on the line over years, and the company tried to buy that learning on a timetable set by its fundraising. It then multiplied the problem by building new plants before the first one worked, and it treated cancellable orders as if they were committed revenue. The lesson for any hardware company is that the order in which the factors are proved decides whether the money raised becomes an asset or a loss. Chapter 24 of The Power Equation sets out that founder's sequence in full.