Home
All posts

· Battery manufacturing

Price Wars: East vs. West Battery Production and the SynBatt Response

Why this mattersCheaper, reliable battery production helps clean energy scale. Knowing which costs engineering can change makes the next investment more useful.

A gigafactory announcement is a promise. A cell delivered to a customer is a different thing. The map tracks six years of battery production: the largest Chinese producers expand their global deployment, European startups announce ambitious plants, and some companies fail before those plans become repeatable production.

2020–4 October 2026. Manufacturer circles show company-wide EV battery deployment at their headquarters; factory rings show planned capacity, not achieved output. 2026 observations cover January–August only. Future start years are plans. Data, dates & method ↓

The video omits EnerDel, American Battery Factory (ABF), Our Next Energy (ONE), Factorial, Amprius, and PowerCo's St Thomas project in North America, plus Agratas in Sanand, India. Their milestones remain in the sourced dataset. EnerDel's 2023 Chapter 7 filing has no documented process-cost cause in these sources; ABF's startup and Tucson factory are announcements, with achieved deployment unavailable. [32, 33]

This selected-company view includes operating and ramping factories alongside failures. Chinese, Korean, and Japanese producers also have different growth trajectories. The comparison is how plants reach competitive deliveries, with each company's product and market taken into account.

Company cases: production, financing, and continuation

The cases cluster around three pressures: funding that arrives too late, production that takes longer than financed, and demand that changes after investment. Companies can face several at once; their responses range from administration to product changes and continued operations through restructuring.

Britishvolt and AMTE Power both needed cash before promised funding could arrive. Britishvolt had no material revenue: management planned £800 million of equity in 2022, while EY recorded £167.5 million raised across its equity rounds. Its £100 million government grant required spending milestones that needed substantial additional investment first. Administration followed on 17 January 2023. [25]

AMTE's proposed subscription was delayed in December 2023 by due diligence and the holiday period. The investor declined bridge funding; the board said it could no longer finance trading. Administrators were confirmed in January 2024. FRP later sold the Thurso manufacturing business and assets to LionVolt, preserving twenty jobs. [43, 44, 45]

A funding commitment and available cash have different timelines. Conditional grants require a company to finance progress before receiving support; due diligence can outlast its remaining liquidity. Smaller factory stages and bridge funding aligned with actual closing dates could reduce that exposure. These cases document financing failures, with AMTE's manufacturing assets transferred to a new owner.

Northvolt and Morrow encountered ramp-up and financing pressures together. Northvolt's March 2025 bankruptcy statement cites internal ramp-up difficulties, higher capital costs, supply-chain disruption, and changing demand. Its September 2024 response had concentrated activity on Northvolt Ett, the gigafactory in Skellefteå, Sweden, put upstream cathode production into care and maintenance, and reduced other activities. [23, 24]

Morrow opened its 1 GWh/year Arendal LFP factory in August 2024 for test production and equipment tuning. Installing the dry room, utilities, and production equipment took 366,000 working hours. Its October LinkedIn post said an immigration-rule interpretation prevented South Korean installation specialists from travelling, threatening up to four months of delivery delays. [35, 36]

Morrow's board announced a bankruptcy filing on 6 May 2026, citing oversupply, price pressure, higher capital costs, industrialisation delays, and a restrained investment market. An industrial-investor process could not finish before liquidity ran out, despite recent supply agreements. Engineer Miguel Pereira's May account emphasised manufacturing progress and patient investors; Intercalation Station's June interpretation questioned LFP profitability, next-generation chemistry readiness, and imported-material dependence. [37, 38, 39]

These two cases suggest a compounding problem: delayed deliveries defer income while equipment, staff, and expansion keep consuming cash. Higher financing costs shorten the time available to qualify the product and fulfil orders. Manufacturing progress and new contracts can therefore coexist with a liquidity crisis. Funding has to cover the actual ramp, and the resulting cells must be competitive enough to sustain production.

PowerCo illustrates the infrastructure needed before production starts. A May 2025 Valencia update quoting its Spanish CFO describes 17 clean rooms covering 57,000 m², some requiring dry conditions, installed before production equipment. Salzgitter was commissioned in December 2025 with 20 GWh/year of initial nameplate capacity. Its participation in MiKoBatt, documented in a 2026 Bundestag disclosure, targets smaller conditioned air volumes through factory and machine micro-environments. [26, 27, 28, 19]

FREYR illustrates equipment commissioning risk. It sanctioned the 29 GWh/year Giga Arctic plant in June 2022; by November 2023, casting and cell-assembly commissioning at its qualification plant had slipped beyond the fourth quarter. It increased vendor coordination and technical support, cut spending, and minimised Giga Arctic investment while comparing European support with US and Canadian incentives. In November 2024 it pivoted to US solar manufacturing and terminated its 24M technology licence, recorded as a pivot on the map. [40, 41, 42]

Both examples concern capital committed ahead of saleable output, although their technologies and outcomes differ. Micro-environments can reduce installed infrastructure and operating loads; qualification plants can reveal whether unfamiliar equipment works before a larger commitment. The scale-up budget needs to include the machinery, atmosphere, commissioning knowledge, and time required to make the whole line productive. The ventilation and exposure requirements are discussed below.

ACC and LG Energy Solution show how the market can move during expansion. ACC's 2024 report describes first-factory learning and a slower European EV market. In June its CEO explained pausing German and Italian developments to investigate cheaper products while the French factory ramped up. In February 2026, UILM reported that management had confirmed shelving the Termoli and German projects; these expansion decisions are distinct from the French plant's operation. [11, 30, 46]

In its January 2026 results, LG Energy Solution attributed lower 2025 revenue to slower EV sales at major customers. Its fourth-quarter operating loss was KRW 122 billion including KRW 332.8 billion of North American production incentives, or KRW 454.8 billion before that support. Its announced 2026 response redirects idle EV capacity to storage, expands lower-cost LFP and mid-nickel products, cuts CAPEX by more than 40% against 2025, and includes an LFP dry-electrode pilot. [47]

A faster line helps only if its product fits the buyer's price and specification. Chemistry, demand, utilisation, and incentives can change while a factory is being built. Adapting existing capacity can preserve more value than expanding the original product roadmap. This exposure reaches experienced Asian producers as well as European entrants.

VARTA Microbattery is also my employer, so my take may carry a little extra charge of optimism. Operations continue during the group's restructuring: its July 2026 statement reports unrestricted business activity and a long-term continuation objective. Its manufacturing expertise remains part of Germany's battery industry. [71]

CoinPower cells are rechargeable lithium-ion batteries in a round, flat button-cell format: anode, cathode, and separator are wound into a coil inside stainless-steel housing. The CP1254 A4X measures 12.1 mm diameter × 5.4 mm height, with 3.7 V nominal voltage and 74 mAh nominal capacity. These miniature cells sit outside the map's automotive EV-deployment measurements. [57, 58]

VARTA's November 2022 statement describes capacity and cost adjustments, including short-time working at Nördlingen, in response to higher energy and raw-material prices, customer production interruptions, semiconductor shortages, and weaker true wireless stereo headset demand. Cost increases reached customers only partly and with a delay. [59] The 2024 annual report shows a recovery in Lithium-Ion Small Cells, representing CoinPower: €101.1 million revenue, €64.2 million personnel expenses, and €14.9 million adjusted EBITDA; the adjusted EBITDA margin rose from −25.0% in 2023 to 14.7%. VARTA attributed this to higher call-offs from a major customer and more flexible production. EBITDA excludes depreciation, interest, and taxes; cell-level cost comparisons require separate data. [60, p. 44]

The 2 October 2026 update reports stabilised operations and a sales process progressing as planned towards long-term continuation. Formal insolvency proceedings opened on 1 October for VARTA AG, Microbattery, Storage, and Micro Production. The planned Nördlingen closure concerns VARTA Micro Production after a major customer loss, separately from ongoing Microbattery operations in Ellwangen. Consumer Batteries is unaffected. The recovery and continuation efforts illustrate operational flexibility alongside customer-concentration risk; the public reports leave CoinPower's unit-cost gap unquantified. [61]

Verkor provides a useful comparison: its 2023 financing supported an initial 16 GWh/year Dunkirk project with a Renault partnership. In August 2026 it reported commissioning a second line and delivering its first batteries made entirely on-site. Financing and a customer link reduce two risks; competitive production still requires qualified output, yield, and utilisation. The announcements leave annual yield and cost parity unquantified; the map shows the nameplate target. [13, 31]

The recurring risk is a mismatch between the production, funding, and customer schedules. A delayed ramp can exhaust cash and miss a vehicle launch; a completed ramp can still face falling orders or an uncompetitive price. Preserving industrial capability therefore requires coordinated financing, realistic commissioning time, and customers for the eventual product. A smaller, productive first stage can be more valuable than a larger announced capacity that remains unfunded or underused.

Why not fund it until it works?

Volkswagen announced an additional €500 million equity investment in Northvolt in June 2021. BMW participated in the 2019 equity round and separately signed a €2 billion battery-cell supply contract in 2020, with deliveries planned from 2024. Equity finances the company; a supply contract commits future purchases. Its headline value is not cash already invested. [48, 49, 50]

BMW cancelled the order in June 2024 while expressing continued interest in sustainable European manufacturing and next-generation cells with Northvolt. It had already announced CATL and EVE contracts for sixth-generation round cells in 2022. An automaker can continue electrification through alternative suppliers even when one factory project falls behind. [51, 52]

A rescue decision has to value the equipment, knowledge, and relationships already built, then assess the remaining investment against three conditions:

  • Technical progress: measured improvements in yield, throughput, and qualification, with a credible completion timetable.
  • Enough runway: funding through commissioning and qualification, including delays.
  • Customer fit: a cell that meets the buyer's architecture, price, and vehicle-launch schedule.

Northvolt received further support: its November 2024 restructuring announcement identified approximately $145 million in cash collateral and a customer's $100 million new-financing commitment. Bankruptcy followed in March 2025. The attempt shows willingness to fund a further bridge, but leaves open how much funding and time a viable recovery would have required. [53, 23]

Industrial capability has benefits that an individual buyer cannot fully capture. Europe gains skills, supplier diversity, and resilience from a surviving plant, while the next investor bears concentrated risk and a carmaker can source elsewhere. That gives public and private stakeholders a reason to coordinate financing. Patient capital still needs an independently assessed technical recovery, committed customers, and enough runway; prior spending alone cannot determine the next investment.

The EU's enacted 100% fleet CO₂ reduction target for new cars and vans from 2035 concerns new sales, leaving existing combustion cars in use. The Commission proposed 90% plus compensation mechanisms, still under Council discussion in June 2026. Both favour electrification, but policy-driven battery demand does not guarantee orders for a particular supplier. [54, 55]

Industrial assets can also continue under new ownership. Lyten completed its acquisition of Northvolt Ett, Ett Expansion, and Northvolt Labs in February 2026, announcing a commercial restart plan. The map records Northvolt's failure and the Lyten/Ett project separately; continuation preserves an opportunity to demonstrate viable production economics. [56]

First, compare the same thing

BloombergNEF's 2025 survey puts the global average lithium-ion pack selling price at $108/kWh, China's average at $84/kWh, and Europe's average 56% above China's. That implies roughly $131/kWh in Europe. These are market averages, not identical-cell factory cost estimates. Its stationary-storage pack average is $70/kWh; LFP and NMC packs average $81 and $128/kWh respectively. Product mix changes the comparison substantially. [1]

A bare stationary-storage cell and an automotive pack have different packaging, thermal management, electronics, qualification, and applications. Moving from $131 to $84 requires about a 36% reduction in the European price: the 56% premium uses China as its denominator.

The IEA's 2026 review confirms China's manufacturing dominance and the shift towards LFP. Regional selling prices reflect supply origins and product mix; identifying achievable process savings requires a factory-level comparison. [2]

Cheap inputs, fast lines, and delivered volume drive the competition. China's advantage includes established supply chains, cheaper materials and equipment, scale, and manufacturing efficiency. Europe's response has to reach the purchasing contracts, machine prices, line speeds, and production schedule. Better analysis is useful only when it changes those economics. [3, 4, 67]

A manufacturer can meet an output target by producing more, including extra production to cover rejects. Whether that is competitive depends on what each additional kWh costs. Cheap inputs and high throughput can make a lower-yield line more economical than a slower, expensive line with fewer rejects. Safety and customer specifications still apply to every cell shipped.

Two compounding effects: currency and trade policy

A declining dollar selling price and a weaker dollar are separate pressures. In an illustrative, unhedged example, a $100 order brings €90.91 at $1.10 per euro, but €83.33 at $1.20 per euro. If the dollar price also falls to $90, receipts become €75.00, a 17.5% reduction from the starting point, while euro-denominated factory costs may remain. The exchange rates and order values in this example are hypothetical.

Dollar-denominated inputs, local costs, currency hedges, and contract terms determine net exposure. VARTA's 2024 report provides a useful example of this balance: procurement and sales currency effects largely offset across the group, with different effects in individual segments. CoinPower cannot be assigned the full gross-revenue effect from the example above. [60, pp. 56, 157]

The official notices cover export VAT rebates, export licensing controls, and domestic battery taxes. None introduces a blanket tariff on lithium exports.

China's battery-related measures, checked against official notices as of 4 October 2026
InstrumentVerified scope and timingWhat it means for the comparison
Export VAT rebateListed battery products: 9% → 6% from 1 April through 31 December 2026; cancellation from 1 January 2027. [62]A reduction in exporter tax support, rather than a blanket tariff on raw lithium. My inference: if passed into export prices, it could narrow the gap against European cells; the pass-through is not measured here.
Export licensing controlsOctober 2025 Announcement 58 covers selected batteries, manufacturing equipment, technology, and materials. Announcement 70 suspended it through 10 November 2026. [63] [64]Licensing is distinct from taxation. These suspended provisions cannot be counted as a continuously applied 2025–2026 lithium export levy.
Battery consumption taxListed categories, including lithium-ion batteries: 2% from 1 September 2026, rising to 4% in September 2027. Export refund/exemption policies remain applicable. [65] [66]The headline domestic tax rate cannot simply be added to a European buyer's imported-cell cost as an export tariff.

VARTA's 2022 production adjustments predate these notices, which do not establish a CoinPower lithium-export-tariff effect. Company-specific attribution requires the input, origin, contract date, tariff classification, and price change reaching the factory.

How much support do Chinese manufacturers receive?

MIT CEEPR/Rhodium estimates Chinese EV and battery manufacturing support at €67.3 billion for 2018–2024 and €16.2 billion for 2024 alone. Batteries received 42% of the 2024 total, approximately €6.8 billion. These estimates include grants, tax incentives, and below-market financing benefits, excluding vehicle-purchase incentives. [72, pp. 4, 8–9]

Company-level support: estimates and audited disclosures
Recipient and periodAmountBasis and scope
CATL · 2024Nearly €1.20 billion in grants plus about €0.74 billion in tax incentives: roughly €1.94 billion combined.MIT/Rhodium estimates in 2023 euros; reported grants can include other jurisdictions. [72, pp. 9, 16]
BYD · 2025€1.56 billion of grants recognised in profit or loss, equivalent to 1.57% of group revenue.Audited group accounts, including overseas operations, with no battery-only or donor-country split. Asset-related and income-related grants include interest support and releases of earlier deferred grants; this is not cash newly received in 2025. [73, pp. 144, 311]

Qualifying Chinese high-tech enterprises pay 15% corporate income tax instead of 25%, a 40% rate reduction. Manufacturing received 46% of China's EV-and-battery support in 2018–2024, versus 11% in Europe (EU-27 plus UK): 4.2× the share directed to producers, not 4.2× the support per kWh. [72, pp. 4, 9, 12–14]

Supporting manufacturers through construction and ramp-up gives them time to ship enough cells to spread fixed costs. Buyer incentives can increase demand without reserving orders for a European factory. Grants and cheaper finance can therefore reinforce the advantage from cheap inputs, equipment, and high output; these figures do not establish how much of the pack-price gap subsidies alone explain.

Coverage includes Chinese local aid but excludes European local aid, R&D, and below-market inputs. Study amounts retain 2023 purchasing power, converted at the ECB's 1.08127 dollars per euro. BYD's nominal 2025 amounts use 8.11850 yuan per euro; grant income is divided by group revenue. CATL is included in the industry estimates; figures with different periods and bases should not be added. [72, pp. 2, 15–16] [73] [74]

Twenty reasons the gap persists, in numbers

Each comparison uses a European reference or conventional route and a cheaper or more efficient benchmark. Parts, staff, time, and power have different units and cost shares. Private purchasing data are represented by published models or labelled scenarios; these benchmarks do not identify the cheapest possible factory.

The product and its supply chain

  1. Chemistry mix: roughly 1.67× or more in pack price. The IEA reports LFP packs were over 40% cheaper than NMC in 2025: NMC/LFP > 1/0.60. EU EV demand remained predominantly nickel-based. Storage applications affect the average; this is not a matched-cell experiment. Market comparison · [2]
  2. Specification: 1.24× the pack mass for the same energy. The IEA's 2025 examples reach 205 Wh/kg for LFP and 255 Wh/kg for nickel-based packs: 255/205 = 1.24. Choosing cheaper chemistry changes mass and packaging requirements. This physical ratio is not a manufacturing-cost penalty. Product benchmark · [67, p. 123]
  3. Electrode processing: 1.10× the cell cost. A 2026 LFP study models an optimised dry route 9.3% below its water-based reference: 1/0.907 = 1.10. The corresponding cell costs are approximately €45.1 versus €49.8/kWh. This is a process scenario, not an EU–China measurement. Cost model · [5]
  4. Pack architecture: 1.67× as many parts. CATL claimed its first cell-to-pack design removed 40% of conventional pack parts: 1/0.60 = 1.67. That reduces assembly work, but does not establish an equal cost saving or describe every current European pack. Manufacturer claim · [68]
  5. Lithium purchasing: 1.27× the input price. The IEA assumes approximately €61.9/kg of lithium content for EU buyers versus €48.7/kg for major Chinese buyers. The inputs reflect bargaining power and integration, not a uniform tariff; they are not prices per kilogram of lithium carbonate. Cost-model inputs · [67, Annex A]
  6. Materials sourcing: 1.24× input price. Modelled discount: 19.5%, calibrated to reported Chinese prices; purchasing contracts are undisclosed. Model assumption · [4, §3.1.2]
  7. Manufacturing experience: roughly 10× the cumulative volume in China. By 2024, China had cumulatively produced about ten times Europe's EV battery volume. That is a much larger base of operating lines, deliveries, and production experience; ten times the volume does not imply ten times lower costs. Historical production · [67, p. 119]
  8. Late rejection: 5× scrap. Modelling 5% versus 1% end-of-line rejects lowers cell cost 4.2%. Research scenario · [4, §3.3.3]

The plant that actually runs

  1. Equipment: about 2× machine price. Chinese-supplier equipment is assumed to cost half the European benchmark. Model inputs · [4, §3.1.2]
  2. Scrap and downtime: 3× the inefficiency allowance. The IEA assumes 15% for the EU and 5% for China, combining rejects and downtime. For equal input and other conditions, 95%/85% implies about 1.12× the cost per good unit, rather than 3× total cost. Model assumption · [67, Annex A]
  3. Capacity utilisation: 1.70× the fixed cost in a half-used plant. At 50% versus 85% utilisation, fixed cost per kWh scales as 85/50 = 1.70, holding yield constant. The IEA uses 85% as a planning ceiling; 50% here is an illustration, not a measured European average or Chinese benchmark. Illustrative calculation · [69]
  4. Factory planning: 1.80× time. German model: 4.5 years; Chinese expert estimate: 2.5 years before production. Model / estimate · [4, §3.1]
  5. Staffing and automation: 3.57× the labour intensity. The IEA assumes 125 workers/GWh for the EU versus 35 in China: 125/35 = 3.57. This measures staffing for equivalent output, before differences in hourly pay. Cost-model inputs · [67, Annex A]
  6. Specialist pay: 5.86× hourly wage. German/Chinese inputs: €45.4/€7.7 per hour. Model inputs · [4, Table 1]

Buildings, energy, and the operating schedule

  1. Construction: 1.33× cost/m². German/Chinese inputs: €3,129/€2,347; factory areas vary independently. Model inputs · [4, Table 1]
  2. Electricity: 2.46× tariff. German/Chinese inputs: €0.183/€0.074 per kWh, before energy-demand differences. Model inputs · [4, Table 1]
  3. Gas: 1.23× tariff. German/Chinese inputs: €0.051/€0.042 per kWh. Model inputs · [4, Table 1]
  4. Dry-room scheduling: 4.59× the power in the active setting. Buck's model reduces power by 78.23% in a lowered, unoccupied operating mode: 1/0.2177 = 4.59. This is an idle-period opportunity, not evidence that occupied production can run at the same saving or that MiKoBatt has already achieved it. Operating-mode scenario · [6]

The cash needed to reach competitive production

  1. Ramp-up: 1.99× spending. Doubled duration raises modelled German ramp-up cost from €170 million to €339 million. Delay scenario · [4, §3.2]
  2. Output support: 1.57× breakeven price without aid. A historical US-credit scenario applied to a German factory models €31/kWh of support; German levelized cost: €85.8 versus €54.8/kWh. This illustrates the effect of aid, separately from the documented Chinese support above. Policy scenario · [4, §3.3.1]

The factors overlap, so their ratios cannot be multiplied: staffing, scrap, throughput, and utilisation interact, and some comparisons change the product. Scrap reductions must be weighed against the costs and deliveries of extra production, inspection, and process changes. The output-credit scenario is a modelling experiment, with no current EU eligibility implied.

Monetary model inputs are shown in euros using the ECB's 2025 annual average of 1.12998 US dollars per euro, without an inflation adjustment; factors use the unrounded original inputs. [70] Supplier clusters, qualification, financing terms, customer commitments, and margin compression also matter, but the public sources do not provide comparable company-level cost factors for each.

The dry room is a lever, not the whole explanation

Solvent control protects workers during processes such as NMP coating; moisture control protects sensitive materials and cell assembly. A factory may need both, with different enclosure and air-handling requirements.

NMP has reproductive toxicity hazards; ECHA's REACH guidance addresses exposure control, and Germany's TRGS 402 covers inhalation assessment. Gefahrstoffverordnung §10 restricts recirculation from relevant hazardous work areas, with an exception for adequately cleaned air using recognised procedures or equipment. Required enclosure, extraction, air handling, and dryness depend on the process and exposure assessment, rather than a universal whole-factory volume or dew point. [14, 15, 29]

The 2025 EDAG coverage gives €120.7 million CAPEX and €18.1 million annual OPEX for an improved factory-layout scenario. Those are layout-specific figures, with no separate energy-only OPEX figure. They cannot be assigned wholesale to a universal dry-room baseline or regional price premium. [16]

A pressurised, air-recirculating machine enclosure can protect the product and reduce conditioning loads if containment, air cleaning, and exposure control also protect workers. Pressure alone does not permit hazardous vapour to enter occupied space. Compare designs handling the same substances and account for leakage, product entry, people, interventions, transfers, maintenance, and safe extraction. Any relaxed dew point must remain compatible with cell quality.

What the 2026 literature is trying to change

Five journal papers and one policy modelling brief address different parts of the cost gap. Their results are specific to the technology, operating conditions, and scenario studied; gigafactory-wide savings require industrial validation.

  1. April 2026 · Factory economicsBridging the global cost gap in battery cell manufacturing

    Lechner and colleagues connect location, factory planning, ramp-up, and subsidy design. The implication is to optimise plant performance and policy together, rather than choose a country from wages alone. The reported outcomes depend on the modelled factory and assumptions. Paper [4] ↗

  2. January 2026 issue · Dry LFP electrodesCosts and carbon footprint analysis of dry manufacturing battery LiFePO4-based cathodes

    Greitemeier and colleagues extend CellEst with laboratory and industry data. Their optimised dry-electrode scenario reduces total cell cost by 9.3% against a water-based reference. This is a chemistry- and scenario-specific estimate, not a claim that dry coating erases the regional pack-price gap. Paper [5] ↗

  3. March 2026 · Factory atmosphereModel based evaluation of dry rooms in battery production

    Buck and colleagues model dry-room loads and air handling to size infrastructure for actual operating conditions. Each process retains its own atmospheric requirements. Paper [6] ↗

  4. June 2026 · Technology selectionA multi-criteria decision analysis approach for evaluating production technology in battery cell manufacturing

    Wicke, Neef, and Tübke compare technology through cost, throughput, quality, and sustainability criteria. The useful contribution is an explicit decision framework; no one process wins independently of the application and assumptions. Paper [7] ↗

  5. January 2026 issue · Dry-process maturityInsights into dry battery electrode manufacturing: Unveiling the patent landscape

    Greitemeier and Lux map the competing dry-process approaches and their patent activity. This helps distinguish a broad technology label from specific routes and ownership. Patents document development activity, not production yield or an achieved cost advantage. Paper [17] ↗

  6. March 2026 · Scale and policyThe falling cost gap between EU and Chinese batteries

    Transport & Environment explores how learning, scale, and policy could narrow the gap. It is an advocacy organisation's scenario analysis, with assumed improvements in scrap and automation, rather than a measurement of factories that have already converged. Brief [8] ↗

SynBatt: change the process, then prove the economics

The BMFTR-supported SynBatt transfer cluster connects research and industrial implementation in battery cell production. Its projects address the machinery, conditioned space, solvent demand, and processing steps that determine what a line costs and how much it can deliver. Their industrial value depends on cheaper installed capacity, faster production, or lower running costs while meeting the required cell specifications. [18]

The BMFTR-funded MiKoBatt project, with PowerCo as an industrial partner, investigates micro-environments around machinery and material handling. The idea is to condition a smaller volume near the product instead of assuming that every surrounding cubic metre needs the same specification. Its 2026 project poster also makes the hard parts visible: transfers, interventions, and integration with the equipment. A smaller enclosure is useful only if it preserves product quality and safe operation. [19, 28]

SkaleD focuses on bringing direct extrusion coating towards industrial production, including double-sided electrodes. Using less solvent can reduce the drying burden, but high-viscosity material must still be fed, distributed, and coated uniformly. The cost opportunity lies in reducing drying equipment and energy demand while producing enough electrode per hour to compete. [20]

E-MoPEd develops simultaneous, automated double-sided multilayer coating and inline process control. Processing both sides together targets more electrode output from the production line. Its economic test is the cost per accepted square metre at the achieved speed, including rejects and downstream constraints. Inline control supports that production target; fewer defects alone would not establish a competitive line. [21]

GranuGoIn is the project I work on personally. It runs from January 2025 to December 2027 and develops granulate-based semi-dry electrode production, including extrusion, calendering, dosing, and inline quality control. Its public description is more specific than “remove every solvent”: the objective is to reduce solvent and drying requirements and compare the resulting route with conventional wet processing. [22]

Fully dry binder-fibrillation routes are a separate research direction from GranuGoIn's semi-dry process. Both must control dust, moisture, and other hazards. Their industrial test is repeatable electrode performance and lower cost per accepted kWh at production speed.

For my work on machine learning, a useful model might shorten a material changeover, enable a higher line speed, or avoid an expensive trial. Its value has to exceed measurement and integration costs, measured through time saved, additional output, and total cost. If extra production fulfils the order more cheaply, it belongs in the comparison. Machinery, purchasing terms, and utilisation still determine whether the factory makes economic sense.

The target is affordable cells, delivered at volume

European plants need equipment and inputs they can afford, lines that produce enough to fulfil orders, and financing that lasts through ramp-up. Mature wet coating remains part of that competition. Each alternative has to justify its investment through the cost and volume of cells it can actually deliver.

Micro-environments, reduced-solvent coating, multilayer processing, and semi-dry granulation offer ways to reduce infrastructure, energy demand, and processing work. They need to deliver those savings at production speed. SynBatt contributes to that effort alongside materials sourcing, utilisation, customer demand, and financing.

Analysis can help choose and operate a production route. Its value has to appear in the factory's costs, delivery schedule, or output. The next European dot on this map needs to grow because customers can buy its cells at a competitive price and receive them on time.

Map data, milestones & method

Evidence cutoff: 4 October 2026. The animation uses one continuous 2020–2026 timeline. Deployment observations are retrospective annual snapshots; startup, commissioning, and insolvency milestones use their recorded dates. Smooth circle changes are visual interpolation, not measured monthly production.

GWh of batteries installed in EVs is a proxy for commercial activity, not cells manufactured, total company revenue, or all stationary-storage sales. Filled area uses GWh divided by 12 for complete years and by 8 for January–August 2026, so a partial year does not look like an annual collapse. This is an observed monthly average, not a forecast of full-year 2026 output.

Dashed rings represent selected factory nameplate targets, normalised to GWh/month by dividing by 12. A blue centre marks a documented ramp-up; an ochre centre marks a plan or construction milestone. An unscaled grey hollow marker means comparable volume is unavailable, including Samsung SDI in the selected 2026 top-ten release. Missing data is never zero. A small blue ring with a blue centre marks continuing operations during restructuring; it carries no volume scale. Black insolvency markers also carry no output scale and do not assert permanent factory closure.

Company markers use approximate headquarters or the named project site, with leader lines to offset labels. Brief halos highlight new milestones without encoding output, and dates beside project names are planned start years. Established makers' deployment is global; it is not output at that dot's geographical location. The choice of companies is illustrative and incomplete. A marker's first appearance can be the first selected announcement, rather than the company's founding date. VARTA AG enters with its 2022 production-adjustment statement; no EV-deployment volume is assigned to its miniature CoinPower cells. Its blue restructuring marker records continued operations, including VARTA Microbattery; the dataset retains the legal proceedings separately. Manufacturing assets can be acquired or restarted under new ownership. Project phases follow the milestones included here, not a live status feed.

Historical deployment values use SNE's following-year comparison column where available, because reporting coverage is revised. CATL's 2020 value comes from its corporate brochure; the other 2020 observations are left unavailable. Amber diamonds distinguish funding distress, project spending pauses, and strategy pivots; grey outlined targets retain the paused plan, rather than represent output. The Morrow marker uses its board's filing announcement; AMTE uses its regulatory confirmation of administrator appointments. Sources and reporting windows accompany the table. Download the sourced dataset (.json) ↓ · Deployment table (.csv) ↓

Reported EV battery deployment (GWh). 2020–2025 are complete years; 2026 is January–August only. n/a means unavailable.
Company2020202120222023202420252026
Jan–Aug
CATL36.299.5184.4257.7342.5464.7333.0
BYDn/a26.470.5111.8152.6194.8127.9
LG Energy Solutionn/a59.471.695.197.8108.868.3
Panasonicn/a36.335.642.834.644.229.7
SK On / predecessor battery businessn/a17.330.134.739.744.524.9
Samsung SDIn/a14.523.933.131.128.9n/a
CALBn/a8.018.533.841.262.844.4
Gotionn/a6.713.916.429.353.541.5
EVE Energyn/an/an/an/a18.731.329.5
SVOLTn/an/an/an/a17.428.522.0
REPT BATTEROn/an/an/an/an/an/a20.3
Sunwodan/an/an/an/an/an/a19.3
  1. · Morrow: Morrow launched in May 2020 (month-level date). Morrow: company history, May 2020 launch.
  2. · Verkor: Verkor publicly announces its launch. Verkor celebrates one year since its July 2020 launch.
  3. · ACC: ACC formed in August 2020 (month-level date). ACC formed in August 2020.
  4. · Northvolt: Volkswagen invests a further EUR 500 million; Ett expansion target rises from 40 to 60 GWh/year. Volkswagen: additional EUR 500 million Northvolt investment; 60 GWh expansion target.
  5. · Northvolt: First Ett cell assembled; expansion target now 60 GWh/year. Northvolt assembles its first Ett cell; 60 GWh expansion target.
  6. · Verkor: Dunkirk selected; initial target 16 GWh/year. Verkor selects Dunkirk; initial 16 GWh target.
  7. · American Battery Factory: ABF announces a US LFP factory network. ABF announces US LFP factory network.
  8. · Northvolt: Commercial deliveries reported; nameplate target is not achieved output. Northvolt Ett commences commercial deliveries.
  9. · FREYR: Giga Arctic construction sanctioned with a 29 GWh/year target. FREYR: Giga Arctic construction sanction, 29 GWh annual target.
  10. · PowerCo: PowerCo founded; factory ramp-up remains ahead. PowerCo founded in July 2022.
  11. · Britishvolt: UK final grant offer; planned Blyth factory capacity 30 GWh/year. UK Government: final grant offer to Britishvolt; capacity: UK Parliament: Britishvolt's planned 30 GWh Blyth factory.
  12. · CATL / Debrecen project: 100 GWh/year cell factory announced across successive phases; total target, not first-phase output. CATL: 100 GWh/year Debrecen target across phases.
  13. · VARTA AG: Capacity and cost measures respond to CoinPower demand and higher input prices; miniature cells, not EV deployment. VARTA AG: production adjustments in response to CoinPower demand and input costs.
  14. · American Battery Factory: Tucson selected for first factory and headquarters; no comparable capacity or output entered in this dataset. ABF selects Tucson for its first factory and headquarters.
  15. · Britishvolt: Funding shortfall before material revenue; Britishvolt enters administration. EY administrators proposals: financing shortfall before revenue.
  16. · PowerCo / Sagunto: Construction begins: 40 GWh/year design, production initially targeted for 2026. Volkswagen: Sagunto construction starts; 40 GWh/year design.
  17. · PowerCo / St. Thomas: Up to 90 GWh/year in the final expansion phase, not at its envisaged 2027 start. Volkswagen: St. Thomas, up to 90 GWh/year at final expansion.
  18. · ACC: Douvrin inaugurated: initial nameplate over 13 GWh/year, shown conservatively as 13; eventual 40 GWh target is excluded. ACC inaugurates Douvrin: initial capacity over 13 GWh, 40 GWh target for 2030.
  19. · EnerDel: EnerDel files for Chapter 7 liquidation. Indianapolis Business Journal: EnerDel files Chapter 7.
  20. · Verkor: Financing secured for the initial 16 GWh/year factory. Verkor secures financing for its 16 GWh gigafactory.
  21. · Amprius / Brighton project: Rezoning approved; 500 MWh/year initial target and a projected 2025 opening. Amprius: Brighton approval; initial 500 MWh/year target.
  22. · Factorial / Methuen: Solid-state development facility opens, designed for up to 200 MWh/year. Not measured deployment. Factorial press release: Methuen pilot facility, up to 200 MWh design.
  23. · Our Next Energy / Michigan: Pilot cells produced; 20 GWh/year describes the factory design, not pilot output. ONE: Michigan pilot production begins; 20 GWh factory design.
  24. · FREYR: Qualification equipment commissioning delayed; Giga Arctic spending minimised to preserve liquidity and await competitive policy conditions. FREYR: Q3 2023 commissioning delay and Giga Arctic spending reduction.
  25. · AMTE Power: Funding completion delayed and bridge advance refused; intention to appoint administrators announced. AMTE: funding delay and intention to appoint administrators.
  26. · AMTE Power: Regulatory announcement confirms administrator appointments (confirmation date). AMTE: regulatory confirmation of administrator appointments.
  27. · Agratas / Bridgwater: Somerset selected: 40 GWh/year factory target, not achieved output. Agratas: Somerset site confirmed; 40 GWh/year target.
  28. · AMTE Power: Administrators report Thurso manufacturing assets sold to LionVolt; insolvency does not mean permanent factory closure. FRP: AMTE Thurso manufacturing assets sold to LionVolt.
  29. · ACC: Slower EV demand and a lower-cost product strategy: German and Italian development paused; French ramp-up continues. ACC: German and Italian project development paused; French ramp-up continues.
  30. · Northvolt: BMW confirms cancellation of its EUR 2 billion supply order; a purchase contract is distinct from equity investment. Reuters: BMW confirms Northvolt supply-order cancellation and next-generation focus.
  31. · Morrow: Arendal inaugurated: test production and equipment tuning; 1 GWh/year target is not achieved output. Morrow: Arendal inauguration, test production, 1 GWh annual target.
  32. · Britishvolt: Britishvolt moves from administration into liquidation. EY: Britishvolt administration and liquidation.
  33. · FREYR: US solar acquisition announced; 24M licence terminated. A strategy pivot, not a bankruptcy. FREYR: US solar pivot and termination of 24M licence.
  34. · Northvolt: Chapter 11 restructuring provides access to cash collateral and a USD 100 million customer financing commitment; operations planned to continue. Northvolt: Chapter 11 restructuring and USD 100 million customer financing commitment.
  35. · Northvolt: Production ramp-up and financing pressures culminate in bankruptcy in Sweden. Northvolt files for bankruptcy in Sweden.
  36. · AESC / Sunderland: UK Export Finance supports plant 2: 15.8 GWh/year nameplate, not achieved output. UK Export Finance: Sunderland plant 2, 15.8 GWh/year nameplate.
  37. · Tiamat / Boves: Financing round for a future 5 GWh/year factory; first phase targeted for mid-2027. Its power rating is not energy capacity. Tiamat: financing round and future 5 GWh factory.
  38. · Gotion InoBat / Surany: Groundbreaking: initial 20 GWh/year target, production expected in 2027. Gotion: Surany groundbreaking; 20 GWh/year target, 2027 production plan.
  39. · AESC / Sunderland: Start of operations announced; nameplate remains distinct from achieved annual production. AESC statement reported by Mysteel: Sunderland start of operations.
  40. · PowerCo: Salzgitter commissioned: first-stage nameplate 20 GWh/year; possible later expansion to 40. PowerCo commissions Salzgitter: initial 20 GWh capacity, expandable to 40.
  41. · Agratas / Bridgwater: Building One steel frame reaches 90% completion; UK operations scheduled for 2027. Agratas: Bridgwater construction progress.
  42. · Our Next Energy / Michigan: EV investments paused; infrastructure and defence batteries prioritised. Not an insolvency. ONE: automotive EV investments paused; infrastructure and defence focus.
  43. · LG Energy Solution: EV customer slowdown: planned response reallocates idle lines to storage and shifts product mix; production incentives support earnings. LG Energy Solution: 2025 earnings and 2026 initiatives.
  44. · Amprius / Brighton project: Lease termination agreed; contract manufacturing chosen for capital-efficient expansion. Not a bankruptcy. Amprius 2025 annual report: Brighton lease termination and contract-manufacturing strategy.
  45. · ACC: UILM says management confirmed shelving German and Italian expansion; the mapped French factory is a separate project. UILM: union account of ACC shelving Italian and German plans.
  46. · ProLogium / Dunkirk: Groundbreaking: 0.8 GWh/year first phase targeted for 2028; full Fab 1 at 4 GWh/year targeted for 2030. ProLogium: Dunkirk groundbreaking and 2028-2032 phased roadmap.
  47. · ElevenEs / Subotica: Series B first closing funds initial works; 1 GWh/year factory planned, first cells expected in 2027. ElevenEs: Caterpillar investment and planned 1 GWh factory.
  48. · Northvolt: Lyten completes acquisition of Swedish manufacturing assets; the original company failure and the new Ett project are shown separately. Lyten: completed acquisition of Northvolt Swedish assets and restart plans.
  49. · Lyten / Ett: Swedish Northvolt assets acquired; 16 GWh/year existing nameplate and planned restart, not measured achieved production. Lyten: completed acquisition of Northvolt Swedish assets and restart plans.
  50. · Agratas / Sanand: Steel frame completed; production expected in 2027. Tata reports a 20 GWh/year plant target. Agratas: Sanand steel frame completed; 2027 production target; capacity: Tata Sons annual report FY2024: Sanand 20 GWh target.
  51. · PowerCo: PowerCo discloses MiKoBatt participation: researching smaller conditioned air volumes with micro-environments. PowerCo Bundestag disclosure: MiKoBatt participation.
  52. · Morrow: Board decides to file: price pressure, capital costs and industrialisation delays; financing could not close before liquidity expired. Morrow: board decision to file for bankruptcy and contributing pressures.
  53. · CATL / Debrecen project: Adjacent module assembly begins. This does not establish achieved cell output. CATL: Debrecen module assembly begins, distinct from cell production; capacity: CATL: 100 GWh/year Debrecen target across phases.
  54. · Verkor: Second line commissioned; first entirely on-site batteries delivered. Achieved annual output not reported here. Verkor: second line commissioned and first on-site batteries delivered.
  55. · Tiamat / Boves: CEO confirms construction postponed pending private finance. Contract production in Asia continues; no bankruptcy. Amiens Metropole: Tiamat construction postponed pending private financing; capacity: Tiamat: financing round and future 5 GWh factory.
  56. · PowerCo / Sagunto: PowerCo and Gotion plan a joint venture: 30 GWh/year initial capacity, production from 2027. Volkswagen, PowerCo and Gotion: Valencia joint venture and 2027 start.
  57. · ProLogium / Dunkirk: Site infrastructure and grid work advance; the phased 2028-2032 roadmap remains a plan. ProLogium: Dunkirk infrastructure update; capacity: ProLogium: Dunkirk groundbreaking and 2028-2032 phased roadmap.
  58. · VARTA AG: Operations stabilised during restructuring; investor process on track. VARTA Microbattery continues operating. VARTA AG: stabilised operations and sales process progressing during restructuring.

Land outlines: Natural Earth, public domain. Figure and rendering by Julius Störk.

Sources

  1. BloombergNEF. 2025 battery pack price survey, 9 December 2025.
  2. IEA. Global EV Outlook 2026: Electric vehicle batteries, 2026.
  3. IEA. Pathways to global EV cost-competitiveness, 2025. Controlled NMC811 direct-cost comparison; not the same scope as the regional pack survey.
  4. Lechner, Nanz, Diller & Daub. Bridging the global cost gap in battery cell manufacturing: From locational factors to factory planning, and policy design. Next Energy 11, 100564, April 2026. Open PDF.
  5. Greitemeier et al. Costs and carbon footprint analysis of dry manufacturing battery LiFePO4-based cathodes. Cell Reports Physical Science 7(1), 103058, 21 January 2026 issue. First available in 2025; the DOI year is not the issue year.
  6. Buck, Fischer, Dilger, Zellmer & Herrmann. Model based evaluation of dry rooms in battery production. Journal of Cleaner Production 551, 147923, March 2026.
  7. Wicke, Neef & Tübke. A multi-criteria decision analysis approach for evaluating production technology in battery cell manufacturing. Journal of Power Sources Advances 39, 100211, June 2026. Open PDF.
  8. Transport & Environment. The falling cost gap between EU and Chinese batteries, March 2026. Policy scenario analysis.
  9. Volkswagen Group. PowerCo supply-chain plans and cathode-material partnership, 2022.
  10. IEA. Global EV Outlook 2025: Electric vehicle batteries, 2025.
  11. ACC. 2024 sustainability report. First-factory learning, quality, and financing.
  12. UK Government. Final grant offer provided to Britishvolt, 27 July 2022.
  13. Verkor. Gigafactory financing and customer partnership, September 2023.
  14. ECHA. Advice on how to comply with the NMP restriction, 17 July 2019.
  15. BAuA. TRGS 402: Ermitteln und Beurteilen der Gefährdungen bei Tätigkeiten mit Gefahrstoffen: Inhalative Exposition.
  16. EDAG. Battery-cell manufacturing white paper, 2025. The layout figures are reported in Battery-News's 7 November 2025 coverage; they are reported scenarios, not independently verified plant costs.
  17. Greitemeier & Lux. Insights into dry battery electrode manufacturing: Unveiling the patent landscape. Next Research, 101118, January 2026 issue; available online in 2025.
  18. SynBatt. Transfer cluster and projects. Official programme information.
  19. PEM, RWTH Aachen. MiKoBatt project poster, 2026.
  20. EAS Batteries. SkaleD: direct extrusion coating at industrial scale, 11 February 2025.
  21. Kampf / Jagenberg. E-MoPEd transfer project, 25 April 2025. Three-year project started in March 2025.
  22. iPAT, TU Braunschweig. GranuGoIn: granulate-based semi-dry electrode production, January 2025–December 2027.
  23. Northvolt. Bankruptcy announcement and stated challenges, 12 March 2025.
  24. Northvolt. Strategic review: sequencing Ett ramp-up and narrowing scope, 9 September 2024.
  25. EY. Britishvolt administrators' proposals, 13 March 2023, pp. 4–7: funding, grant conditions, liquidity, and circumstances of administration.
  26. ICEX / Invest in Spain. PowerCo Valencia construction update quoting CFO Javier Rivera, May 2025. Reported infrastructure and planned schedule, not a finding of delay.
  27. Volkswagen Group. PowerCo commissions Salzgitter, 17 December 2025.
  28. PowerCo SE. Bundestag lobby-register disclosure, 9 April 2026. Primary company disclosure identifying MiKoBatt project participation.
  29. German federal legislation. Gefahrstoffverordnung §10, particularly paragraphs 1 and 2(5), including the exception for adequately cleaned air.
  30. ACC. CEO on affordability, factory pauses, and French ramp-up, 13 June 2024.
  31. Verkor. Second-line commissioning and on-site deliveries, 4 August 2026.
  32. Indianapolis Business Journal. EnerDel Chapter 7 filing, 25 July 2023. Event reporting, not an engineering root-cause analysis.
  33. American Battery Factory. Network launch, 9 March 2022, and Tucson selection, 6 December 2022.
  34. IEA. Energy Technology Perspectives 2026: Executive summary, 2026. Accumulated industrial, policy, and financing advantages.
  35. Morrow Batteries. Arendal factory inauguration, equipment tuning, and installation effort, 16 August 2024.
  36. Morrow Batteries. Company LinkedIn post on subcontractor access and commissioning delays, October 2024, ahead of the 29 October public meeting. Norwegian-language company account.
  37. Morrow Batteries. Board decision to file for bankruptcy and stated contributing pressures, 6 May 2026.
  38. Miguel Pereira. Engineering perspective on Morrow and patient investment, LinkedIn, 7 May 2026. Personal account.
  39. Intercalation Station. There is no (to)Morrow, 3 June 2026. Attributed industry commentary; its interpretation is distinct from the board's statement.
  40. FREYR Battery. Giga Arctic construction sanction and 29 GWh target, 29 June 2022. Company release distributed by Cision.
  41. FREYR Battery. Q3 2023 results: commissioning delay, reduced spending, and incentive comparison, 9 November 2023.
  42. FREYR Battery. Solar acquisition announcement and termination of the 24M licence, 6 November 2024.
  43. AMTE Power. Funding delay and intention to appoint administrators, RNS, 19 December 2023.
  44. AMTE Power. Appointment of administrators, RNS, 19 January 2024. This confirmation date is used on the map.
  45. FRP Advisory. AMTE Power sale preserves Scottish battery manufacturing, 4 March 2024.
  46. UILM. Union statement on ACC management's shelving of Termoli and German projects, 7 February 2026. Italian-language participant account.
  47. LG Energy Solution. 2025 financial results and 2026 initiatives, 29 January 2026. Revenue, quarterly loss and incentives, capacity conversion, and planned investment reduction.
  48. Volkswagen Group. Additional €500 million investment in Northvolt, 9 June 2021.
  49. Northvolt. 2019 equity round including BMW, June 2019. Participation is documented; BMW's individual contribution is not specified here.
  50. BMW Group. €2 billion Northvolt supply contract, 16 July 2020. Purchase commitment, distinct from equity investment.
  51. Reuters. BMW confirms contract cancellation and comments on future cells, 20 June 2024. Company statement reported by Reuters.
  52. BMW Group. Round cells and CATL/EVE contracts for Neue Klasse, 9 September 2022.
  53. Northvolt. Chapter 11 restructuring, cash collateral, and customer financing, 21 November 2024.
  54. European Commission. Current CO₂ standards for cars and vans and proposed revision. Separates enacted targets from the December 2025 proposal.
  55. Council of the EU. Environment Council discussion of the proposed revision, 25 June 2026.
  56. Lyten. Completed acquisition of Swedish Northvolt assets and restart plans, 27 February 2026. Corporate announcement distributed by Cision.
  57. VARTA Microbattery. CoinPower technical handbook, pp. 6 and 9. Manufacturer-authored document hosted by DirectIndustry: round button-cell format and wound electrode construction.
  58. VARTA Microbattery. CoinPower product brochure, specification table, p. 2. CP1254 A4X dimensions, nominal voltage, and capacity; linked from the manufacturer's downloads page.
  59. VARTA AG. Strategic realignment in response to challenges, 15 November 2022, pp. 1–3. Company statement on CoinPower demand, input costs, price pass-through, and production adjustments.
  60. VARTA AG. Annual report 2024, pp. 44, 56, and 157. CoinPower segment results, customer concentration, and currency exposure. Segment figures are not a cell-level manufacturing-cost calculation.
  61. VARTA AG. Stabilised operations and sales process progressing as planned, 2 October 2026. Company statement on continuation, legal proceedings, and the separate Nördlingen decision.
  62. China, Ministry of Finance / State Taxation Administration. Announcement 2026 No. 2: export VAT rebates, 8 January 2026. Official notice and product annexes reproduced by Hong Kong's Trade and Industry Department.
  63. China, Ministry of Commerce / General Administration of Customs. Announcement 2025 No. 58: selected battery and artificial-graphite export controls, 9 October 2025. Read together with its suspension, reference 64.
  64. China, Ministry of Commerce / General Administration of Customs. Announcement 2025 No. 70: suspension through 10 November 2026, 7 November 2025. Includes Announcement 58.
  65. China, Ministry of Finance / General Administration of Customs / State Taxation Administration. Announcement 2026 No. 20: battery consumption tax, 16 July 2026. Rates and effective dates for the listed categories.
  66. China, State Taxation Administration. Battery consumption-tax Q&A, part three, 27 August 2026. Explains export refund/exemption treatment alongside the export VAT rebate change.
  67. IEA. What Next for the Global Car Industry?, 18 November 2025. Product and production benchmarks: pp. 119–123; regional modelling inputs: Annex A, p. 176. Inefficiencies include both scrap and downtime.
  68. CATL. Cell-to-pack platform unveiled at IAA, 10 September 2019. Manufacturer's comparison with a conventional modular pack, including 40% fewer parts.
  69. IEA. Battery production and expansion scenarios, 5 May 2026. Assumes 85% maximum average utilisation; does not report 85% as China's observed average.
  70. European Central Bank. Annual USD/EUR reference rate, 2025. Conversion convention for the monetary model inputs in the twenty comparisons.
  71. VARTA AG. Business operations continue during restructuring, 28 July 2026. Company statement explicitly includes VARTA Microbattery and continued service to customers and suppliers.
  72. Deese, Sidner, Reese, Berit, Wistreich & Bermel, MIT CEEPR / Rhodium Group. Global Clean Investment Monitor: Government Support for Electric Vehicles and Batteries, July 2025. Estimates for 2018–2024; monetary values originally in inflation-adjusted 2023 dollars. Manufacturing support includes estimated financing benefits, not just cash grants.
  73. BYD Company Limited. Annual report 2025, published 27 March 2026, pp. 144 and 311, note IX. Government grants recognised in profit or loss and consolidated operating revenue; amounts in thousands of yuan. Audited accounting recognition, distinct from new grant receipts and vehicle-purchase subsidies.
  74. European Central Bank. Annual USD/EUR reference rate, 2023 and annual CNY/EUR reference rate, 2025. Currency conversions for the subsidy comparison; each retains its source's price-year basis.

What should I call you?

Choose a display name for your comments. No email or account signup.

Your commenting identity

Use at least 12 characters. You’ll need this passphrase to restore the file.