China's EV price war is usually told as a story about sticker prices. For an automaker in Europe, Turkey, Vietnam or the United States, and for the suppliers behind them, it is better understood as a story about cost structure and speed: prices can be copied or blocked by trade policy, a cost structure cannot. What differs from region to region is only the defence each market has built against the overflow.
The price war is a domestic overcapacity problem first
China's passenger car sales fell almost 20% year on year in January 2026, the fastest drop in nearly two years. A few weeks later, on 12 February, the market regulator issued pricing guidelines that ban selling below cost, with a deliberately broad definition of cost that includes administrative, financial and sales expenses. Regulators do not write rules like that for a healthy market.
The rules did not end the discounting. A dealer survey found that 81.9% of dealers were selling below the invoice price, and the industry's average profit margin fell below 4% in the first half of 2026. The guideline is a guideline, not a law: it reaches only sales below cost made to push out competitors, and no public enforcement case had turned up by mid-October. Discounts have mostly changed shape, into long interest-free loans, generous trade-in values and free software features.
Selling below cost is a financing model, not a pricing decision
If a ban was needed, someone was selling below cost. The more useful question is how they can keep doing it, and the answer is that the loss rarely stays on the carmaker's own balance sheet. It is passed along a chain.
- Suppliers act as the bank. Payment terms to suppliers have stretched to months, and supplier-finance instruments let a carmaker's balance sheet look lighter than the chain behind it.
- Dealers absorb the rest. The same survey that found below-invoice selling found that most dealers lost money on new cars. The guideline itself tries to stop manufacturers from pushing dealers into this.
- Capital markets and local governments fund the gap. Chinese carmakers raised large sums through share placements in 2025, including the biggest automotive placement in a decade. Some loss-making start-ups had city investment vehicles as their largest public shareholders, and one brand's home city offered ten years of free rent and a bonus on every car sold. A plant that closes costs the city its jobs, so loss-making capacity stays open.
- Group profits cross-subsidise. A maker with a profitable electronics or battery business can carry a car division that loses money for years.
Not all of it is subsidy, and that matters. Battery pack prices in Europe run about 56% above China's, and industrial electricity in China costs roughly half of what it does in Germany. A price that looks below cost from a European cost base is often a positive gross margin in China. The real losses sit one level down, in research, sales, and financing expenses, which is also why the guideline's broad definition of cost matters so much.
Two questions deserve asking, even though nobody has measured the answers. When a plant, a dormitory and a social insurance gap are carried by someone else, is a very low labour cost per car a wage difference or an accounting arrangement? And when delivery numbers trigger executive share options and the timing of share placements, does management protect the margin or the ranking? The answer is more likely to sit in incentive contracts than in the accounts.
Exports take the volume the home market stopped absorbing
A price war that the domestic regulator cannot stop has to go somewhere. Part of it goes abroad. In July 2026 China exported 553,000 new energy vehicles, up 145.5% year on year, and total vehicle exports passed one million for the second month in a row. September went further: new energy vehicle exports reached about 571,000 units, more than 160% above a year earlier, while total sales over the first nine months, according to the industry association, stayed below last year's level. These shipments are not planned around European demand, or any other single market. They are the volume the home market stopped absorbing. Beijing has started to narrow this escape route as well: in October it decided to cut the export VAT rebate on new energy vehicles from 13% to 9% from 1 April 2027.
For every importing region, the consequence is a steady flow of well-equipped vehicles designed for a market with far thinner margins than the local industry is used to. The question for the local buyer is no longer whether a Chinese car is good enough. The question is how long local makers can compete against a cost base built for a harsher market. Each region answers it with a different defence.
Europe: duties protect the battery car, and the market moves to hybrids
The European Union responded with countervailing duties on battery-electric cars from China, between 7.8% and 35.3% depending on the maker, on top of the standard 10% tariff. In January 2026 the Commission also published guidance on minimum price undertakings and in February accepted the first one, for a single model that a European group builds in China.
Plug-in and conventional hybrids pay only the standard 10%. The result shows in registration data. Chinese brands reached a record 11.7% of European registrations in August 2026, and EU imports of Chinese hybrids rose from about 3,800 in October 2024 to about 50,000 in July 2026. In October the two sides announced an understanding in Beijing. According to press reports, it would cap Chinese hybrid imports at 350,000 a year from 1 December 2026, rising 6% a year for four years, with cars built in the EU by Chinese brands exempt. The legal text has not been published, so the details can still change.
A duty works on one product category at a time. A cost advantage works on all of them. Trade defence therefore buys time but does not change the underlying competition.
Turkey: an additional duty, now under legal review
Turkey protects its market with an additional customs duty on cars. Since September 2025 it applies to every origin without a trade agreement, China included, instead of a China-specific rate: 30% on battery electric cars, or a fixed minimum per car if that is higher, on top of the standard 10%. Relief from the earlier China-specific duty had been tied to local production commitments, and at least one large local production project is reported to be on hold. For the Turkish buyer, this keeps the price gap to the European market wide.
The legal position is not settled. A World Trade Organization panel found in July that Turkey's additional duties on electric cars and hybrids from China, in the form amended in September 2025, exceed its tariff commitments, and that its import licensing rules for these cars are also problematic. Ankara says the report is not final and that it can appeal.
For Turkish makers and the supplier base, the duty means something different anyway. It protects the domestic showroom, not the export markets, where Turkish-built vehicles compete with Chinese ones in Europe and neighbouring regions with no duty advantage. Suppliers that depend on export programmes feel the price war first.
Vietnam and Southeast Asia: Chinese brands took 47% of electrified sales, and three governments answered with local-production rules
Southeast Asia is where most of the exported cars land. In the first half of 2026 Chinese brands accounted for 47% of electric and hybrid car sales across the six largest markets, and for 16% of all cars sold, up from about 11% in 2025, according to a consultancy's half-year review. Three governments have answered with a production condition rather than a duty.
Thailand opened first. Electric and hybrid cars were 53% of its new car sales in the first half of 2026, up from 32% two years earlier, and Chinese makers hold most of the electric segment. Under the current incentive scheme an importer must build two cars locally for every car it imports in 2026, and three in 2027. Indonesia ended duty-free imports of built-up electric cars at the end of 2025. Companies in the scheme must now produce locally as many cars as they imported, with local content announced at 40% for 2026 and rising afterwards, and the first Chinese plant opened in September. Malaysia set a value floor and a motor-power floor on imported electric cars from July 2026, which excludes entry models and leaves locally assembled cars untouched.
Vietnam is the exception. A domestic maker, not a Chinese brand, sold more than a third of all cars in the first half and nearly all of the battery-electric sales that are disclosed, helped by a special consumption tax of 3% on electric cars against 35% to 150% for comparable petrol cars, and by heavy tariffs on built-up imports. In April 2026 the National Assembly extended the 3% rate to the end of 2030, with 11% from 2031, instead of the rise that had been scheduled for March 2027. Chinese car brands are still small in Vietnamese showrooms, so the pressure arrives first through parts, cells and two-wheelers: electric motorcycles were reportedly about a quarter of two-wheeler sales in the first seven months of 2026, and Chinese electric two-wheeler brands are building local plants.
The pattern is simple. Where cars enter without a local-production condition, prices fall and local makers lose share. Where a local-production condition applies, Chinese makers build plants, and a battery and parts supply chain follows.
United States: the closed door that still feels the pressure
The United States is the control experiment. A 100% tariff on Chinese-made electric cars is still in force, and the connected-vehicle rule bans Chinese software from model year 2027 and Chinese hardware from model year 2030, reaching cars built outside China as well. A pending Senate bill would go further, with a 15% Chinese-ownership threshold, and has been postponed until after the elections. Chinese cars are practically absent from American showrooms, so there is no visible car price war.
The neighbours chose differently. Mexico raised its tariff on cars from countries without a trade agreement to 50% from January 2026, while Canada agreed to admit up to 49,000 Chinese electric vehicles a year at a 6.1% duty. In September the administration signalled that it would accept Chinese makers building plants in the United States if they hire American workers, which shows the line is about data and ownership, not only price.
For a battery engineer in the United States, the exposure is the cell. Foreign-entity rules shape where cells may come from, while the cell price is set in China. And that price is not a one-way street: the latest industry survey showed pack prices at a record low at the end of 2025, down 8%, but in the first months of 2026 the mainstream storage cell in China reportedly rose by about a fifth as lithium carbonate more than doubled. Cheap cells are not the carmaker's advantage. Integration, scale and yield are, and a pack that carries unmeasured margin gets more expensive every time raw materials swing.
What the price war changes for an engineering team
The practical lesson is not "cut prices". The price a Chinese maker can offer is the visible end of a chain of decisions made years earlier, and a European, Turkish, Vietnamese or American programme can still act on its own link in that chain.
- Battery cost is the single biggest lever. Pack design, cell format choice and the validation margin carried in the design determine a large share of vehicle cost. A pack that carries margin because nobody measured the real cell spread is expensive twice: in material and in warranty provision.
- Development speed decides which cost curve a programme lands on. A programme that finishes validation late reaches the market against a competitor whose cost has already fallen again. Test planning, abuse testing and software validation sit on the critical path, not at the end as a formality.
- Supplier qualification should rest on measured data. A certificate for one batch says little about the next. Programmes that compare cell and pack data across batches find problems before the supplier invoice turns into a warranty claim.
- Local content is a design decision, not a logistics decision. A duty or a local production rule changes which pack, which cell and which assembly step make sense, in Turkey, Vietnam, Indonesia or the United States alike. Deciding it after the platform is frozen is expensive.
No one can yet say how long the price war lasts
One month of data does not settle it. Five developments are worth watching: whether the below-cost rule starts to bite in China, whether the financing chain behind it can carry the same load after the 2027 export rebate cut, whether the European hybrid cap becomes law, what Turkey does with the WTO report, and whether the United States opens the door to Chinese-owned plants. Each changes the timing, and none changes the direction.
Duties can slow a price war down, but only a lower cost base ends the competition.
For a programme in any of these regions, the useful question for the next planning round is where the cost in the vehicle actually sits and how much of it can still be changed before the platform is fixed. It is an engineering question before it is a commercial one, and it is usually cheaper to answer early.
