Read as:Urban planningIndustry & supply chains

Automotive Lithium-Ion Battery Packs: Fragility in a Policy-Supported U.S. Midstream

Updated: 2026-09-04 Scope: Electrical storage batteries. The representative product category is lithium-ion accumulators, about 78% of exports in the classification.

Key takeaway

China accounts for roughly half of exports in this classification (HHI 2,668). But a country-level reading that stops at “China is strong” misses the industry’s actual vulnerabilities. Through loans and grants from the U.S. Department of Energy (DOE), the United States is pursuing large-scale policy-supported reshoring of lithium refining, cathode materials, and cell manufacturing. GM alone plans to reach about 160 GWh of domestic cell capacity when its projects are complete. By comparison, China’s CATL reported global production capacity of 772 GWh in 2025, about five times GM’s target for the entire United States.

More important is where the supply chain is actually fragile. The most financially vulnerable links are not the geological availability of ore or battery metals, but the capital-intensive, policy-dependent midstream stages—cathode materials and recycling. Ascend Elements, a recycling company that received DOE support, filed for Chapter 11 bankruptcy in April 2026, while Redwood Materials, also supported by DOE, is expanding its business. The repository’s nickel-supply-chain-report.md already covers nickel feedstock in detail. This report focuses on lithium, cathode materials, cell manufacturing, pack assembly, and recycling, which that report does not cover.

Lithium ore (spodumene) mining → refining (battery-grade lithium hydroxide)
                                           ├─ Cathode materials (NMC/NCA/LFP)
Nickel · cobalt · graphite (see nickel report) ─┘
                                           └─ Cell manufacturing (cylindrical/prismatic/pouch)
                                                  → Pack assembly · BMS integration
                                                  → Vehicles · electronics · data centers
Used batteries · manufacturing scrap → collection → black mass
                                     → re-refining into cathode-material feedstock (closed loop)

Network review: no new edges added

We reviewed candidate trade classifications for upstream lithium chemicals (lithium carbonate and lithium hydroxide), including categories that cover carbonates, inorganic salts and hydroxides, and other minerals that may include spodumene concentrate. Existing network edges for those categories connect to entirely different uses—such as sheet glass, fertilizers, and stone products—not to batteries. This indicates that most trade in those broad categories is for goods other than lithium products.

For the same reason that the nickel report removed a link from nickel sheet and other nickel products directly to batteries as a false shortcut over the actual chemical stages, we have not added new edges from these broad classifications. Lithium hydroxide and cathode materials are recorded only as possible future process-type nodes, which do not correspond to a trade classification, if the network later needs them.

1. Classification and trade geography

IndicatorBACI 2024
Exports of the product categoryApproximately US$151.1 billion
Export-country HHI2,668
Representative product categoryLithium-ion accumulators, including separators
Representative category’s share within the classification77.6%
Leading exportersChina 50.0%; Poland 5.7%; United States 5.7%; South Korea 5.1%; Germany 5.0%

China’s share is exceptional. Poland, South Korea, and Germany, however, are locations for finished-pack assembly by European- and Korean-based manufacturers, not countries where the raw materials are mined. Export statistics show where finished products are shipped from. They must be read separately from concentration in upstream mining, refining, or cathode materials.

2. Lithium: from mining to battery-grade chemicals

In March 2026, DOE issued an environmental assessment (EA) and a finding of no significant impact (FONSI) for Albemarle’s concentration-facility project at Kings Mountain, North Carolina. The project plans to produce approximately 420,000 tonnes of spodumene concentrate annually. DOE NEPA: Kings Mountain Lithium Mine Project

In an investor announcement in March 2023, Albemarle said its Mega-Flex refining facility in Chester County, South Carolina, would require an initial investment of more than US$1.3 billion and produce about 50,000 tonnes of battery-grade lithium hydroxide annually, with capacity expandable to 100,000 tonnes. The facility is designed to process multiple feedstocks, including lithium recovered from recycled batteries. The company said the initial output would be equivalent to the supply for about 2.4 million EVs. Albemarle: Mega-Flex announcement

3. Cathode materials: a small number of qualified production sites

BASF opened an NMC cathode-materials plant in Elyria, Ohio, with an investment of more than US$50 million and a US$24.6 million DOE grant. At the time of the opening, the company was reported to be one of only two licensees of NMC cathode-material technology from Argonne National Laboratory. BASF: Elyria plant opening

In February 2023, Redwood Materials received a US$2 billion conditional loan commitment from DOE’s Loan Programs Office (LPO). Redwood said it would build what it described as the first U.S. facility to integrate production of anode copper foil and cathode materials from recycled lithium, nickel, and cobalt feedstocks at McCarran, Nevada. Target annual capacity is about 36,000 tonnes of copper foil and 100,000 tonnes of cathode material, which the company said could support more than one million EVs at full capacity. DOE LPO: Redwood Materials

4. Cell manufacturing: U.S. policy-led reshoring and China’s scale advantage

Several large cell factories in the United States are operating or under construction, supported by DOE Advanced Technology Vehicles Manufacturing (ATVM) loans, LPO loans, and grants under the Bipartisan Infrastructure Law (BIL).

・In July 2025, Panasonic Energy announced the start of mass production of cylindrical 2170 cells at De Soto, Kansas. The target annual capacity is about 32 GWh. Together with the existing Nevada plant, which began operating in 2017 and has supplied more than 11 billion cells, employs over 4,000 people, and has annual capacity of about 41 GWh, the planned U.S. total is approximately 73 GWh. Panasonic Energy: De Soto mass production begins
・In August 2026, LG Energy Solution announced that its 226-acre Lansing, Michigan, factory, targeting more than 35 GWh of annual capacity, would make LFP cells for stationary energy-storage systems and NMC cells for Toyota’s 2027-model-year Highlander EV, assembled in Kentucky, at the same plant, with production tailored to each use. LG Energy Solution: Lansing factory
・GM and LG’s joint venture, Ultium Cells, received a US$2.5 billion DOE ATVM loan finalized in December 2022. The three factories in Lordstown, Ohio; Spring Hill, Tennessee; and Lansing, Michigan, produce nickel-cobalt-manganese-aluminum (NCMA) pouch cells. DOE: Ultium Cells
・In April 2023, GM announced an investment of more than US$3 billion with Samsung SDI in a new joint-venture plant targeting more than 30 GWh of capacity and operations beginning in 2026. The plant is planned to make nickel-rich prismatic and cylindrical cells. GM said this would bring its planned U.S. cell capacity to about 160 GWh at full utilization. GM also said the China-built Cadillac Lyriq uses prismatic cells and shares a pack design with the U.S. version. GM: Expansion investment with Samsung SDI

By contrast, CATL disclosed in its 2025 annual report, released in March 2026, global production capacity of 772 GWh as of 2025, another 321 GWh under construction, battery sales of 661 GWh (up 39% year over year), and revenue of RMB 423.7 billion. CATL’s global production capacity alone is about five times GM’s approximately 160 GWh target for the United States. Market-share figures such as 39.2% come from market-research firm SNE Research and are not used here as primary-source evidence. CATL 2025 Annual Report

5. Pack assembly and battery-management systems

In February 2026, BorgWarner announced an expansion of a series-production battery-management-system (BMS) program for an unnamed major automaker. Production has been underway since 2023, and the program is planned to expand to new battery-electric and plug-in hybrid vehicle lines from 2029 onward. The architecture has a modular battery-management unit (BMU) and cell-monitoring units (CMUs) at the cell level, meets ASIL D functional-safety requirements, and supports packs up to 800 volts.

This primary source shows that the stage can be performed as a business by a Tier 1 supplier independent of the cell maker. It is only one confirmed company example; production volume and the overall market structure have not been disclosed. BorgWarner: BMS program expansion

6. Recycling: divergent outcomes among policy-supported companies

In its 2025 annual report, CATL disclosed that it processed 210,000 tonnes of used batteries and produced 24,000 tonnes of recycled lithium salts. CATL 2025 Annual Report

In the United States, two DOE-supported recycling companies have had sharply different outcomes. Ascend Elements, based in Westborough, Massachusetts, produces recycled lithium carbonate and cathode-precursor materials and filed for Chapter 11 bankruptcy in April 2026. An industry-media report citing the company CEO’s LinkedIn post attributed the filing to “years of financial and operational missteps.” The filing came despite the company’s Base 1 facility in Georgia, with annual capacity of 30,000 tonnes, which had operated since 2022 and reached commercial-scale production. Its Apex 1 facility in Kentucky was under construction; the combined investment in the two sites exceeded US$1 billion. The company was also reported to have separately lost a DOE grant. Recycling Today: Ascend Elements files for bankruptcy

By contrast, Redwood Materials, another DOE-backed company described in Section 3, is expanding. This comparison shows how financially vulnerable the policy-supported U.S. midstream—cathode materials and recycling—can be: the capital intensity of these stages makes supply highly dependent on the financial position of individual companies.

7. What creates barriers to entry

1.Dependence on policy support for capital-intensive midstream stages. Many U.S. cell, cathode-material, and recycling facilities rely on DOE ATVM or LPO loans and BIL grants. Supply therefore depends both on policy continuity and on each company’s financial strength.
2.Different chemistries and formats serve different uses. As LG Energy Solution’s Lansing plant illustrates, one factory can make LFP for stationary storage and NMC for vehicles. Its capacity cannot be added up as though it were a single undifferentiated “battery factory.”
3.Asymmetry in cell-manufacturing scale. One Chinese company’s global production capacity can be far greater than the domestic target of a major U.S. automaker.
4.Financial fragility in recycling. Even with technology and permits, capital-intensive closed-loop facilities can struggle to become profitable on their own. Ascend Elements is a concrete example that recycling operations can fail financially despite policy support.

8. Evidence scope and next research steps

Confirmed company and government primary materials include DOE (NEPA environmental review, LPO loan announcement, and the Ultium Cells project page), Albemarle, BASF, Panasonic Energy, LG Energy Solution, GM investor disclosures, CATL’s annual report, and BorgWarner. The bankruptcy account for Ascend Elements comes from industry media citing the CEO’s own statement.

For lithium refining, the only facility reviewed here is Albemarle’s U.S. site; primary sources for refining sites in China, Chile, and Argentina have not yet been collected. The evidence for pack assembly and BMS consists of one BorgWarner example; production volume and market structure remain open questions. For upstream nickel, cobalt, and graphite, see nickel-supply-chain-report.md and existing industry-structure records on Tesla and DOE cases.

Data and primary sources

・Repository data: phys-network.json, trade-concentration.json, primary-hs6-products.json, nickel-supply-chain-report.md