BESS in the US Market: Non-FEOC Is Not Where You Build

In the US battery market, non-FEOC is used as if it were a guarantee. A system built in Vietnam, Türkiye or Texas is presented as eligible for American tax credits because the factory is not in China. Since 2026 that reasoning is wrong, and it is wrong in a way that costs tens of millions of dollars per project. Knowing where the product was assembled is no longer enough. The rules ask who controls the seller, and what share of the product's cost came from companies tied to China, Russia, Iran or North Korea.

This article explains the term for readers who have never met it, shows how much money rides on it, walks through the two tests and the calculation, and puts the five common import routes side by side with their tariff and their credit. It ends with the question we are asked most often: can an integrator in Vietnam or Türkiye that buys Chinese cells ever qualify?

What FEOC and non-FEOC actually mean

FEOC stands for Foreign Entity of Concern. In US law it is a company owned by, controlled by, or under the influence of the government of one of four countries: China, Russia, Iran and North Korea. Non-FEOC simply means a company, or by extension a product, that is free of those ties in the ways the law defines.

The term comes from the 2021 infrastructure act and was first used for the electric vehicle credit. The tax law of 4 July 2025, the One Big Beautiful Bill Act, replaced it for storage and manufacturing credits with a broader term, Prohibited Foreign Entity, and added a percentage test on the product itself. Banks, buyers and the press still say FEOC for all of it, and this article does the same. Where a rule applies only under the 2025 definition, the text says so.

Three other abbreviations matter. BESS is a battery energy storage system, the complete container with cells, racks, cooling, fire protection, control electronics and the inverter. Section 48E is the investment tax credit a project owner receives for building a storage plant, a percentage of project cost. Section 45X is the credit a manufacturer receives for each kilowatt-hour of cells and modules produced in the United States.

How much money depends on FEOC status

Under Section 48E a storage project owner receives a credit against federal tax of 30 percent of eligible cost when the project pays prevailing wages, which every utility-scale project does. Two bonuses of 10 points each are available, one for domestic content and one for building in a designated energy community. A well-structured project therefore claims 30, 40 or 50 percent of its cost back.

Take a project of 100 MW and 400 MWh. At NREL's 2024 base-year installed cost of about 334 USD per kWh for a 4-hour system, the project costs roughly 134 million USD; the percentages below scale with whatever a given project actually costs.

• 30 percent base credit: about 40 million USD
• 40 percent with domestic content: about 53 million USD
• 50 percent with both bonuses: about 67 million USD
• FEOC rules failed: zero

The last line is the point newcomers miss. Failing the FEOC rules does not remove a bonus. It removes the entire credit. Three further consequences follow. For projects first claiming the credit from mid-2027, the entire credit is clawed back if, within ten years of starting operation, the owner pays a Chinese, Russian, Iranian or North Korean entity under a contract that gives it effective control. The credit cannot be sold for cash to such an entity. And a supplier that certifies a product as non-FEOC, and knew or should have known it was wrong, owes a penalty of at least 5,000 USD or 10 percent of the buyer's resulting underpayment, whichever is greater. The buyer's own accuracy penalty starts at a 1 percent understatement, and the tax authority has six years to assess.

For a manufacturer the numbers are the same order of magnitude. Section 45X pays 35 USD per kWh for every cell and 10 USD per kWh for every module produced in the United States, 45 USD per kWh for a plant that makes both. A 3 GWh plant running at capacity earns about 135 million USD a year, a large part of its construction cost, every year. The credit stays at full value through 2029, then steps down to 75, 50 and 25 percent in 2030 to 2032 and ends after 2032. A manufacturer that is a FEOC, or whose components fail the ratio, receives nothing.

On a typical project the credit is worth a third of the budget and the bank lends against it. A wrong assumption about non-FEOC status does not produce a smaller profit. It produces a failed company.

Test one: who owns and controls the company

The 2025 law defines two classes of prohibited entity. The first is the obvious one: governments of the four countries, their citizens unless they are also US citizens or permanent residents, companies incorporated or principally based there, companies owned or otherwise controlled by any of these, and companies on four US sanctions and defense lists, including the six battery makers named in the 2024 defense authorization act. Every cell manufacturer incorporated in China is in this class. There is nothing to calculate. Both classes have been barred from the credits for tax years beginning after 4 July 2025.

The second class is the one that surprises people. A company from anywhere becomes a FEOC if, during the tax year, any one of these applies:

• A single Chinese, Russian, Iranian or North Korean entity holds 25 percent of its shares, directly or through intermediate companies.
• Several such entities together hold 40 percent. Three funds at 15 percent each add up to 45.
• Such entities hold 15 percent of its debt. Loans and bonds count, not only shares. Genuine supplier financing, such as a vendor loan or deferred equipment payments structured as debt, counts too; ordinary trade payables need case-by-case analysis.
• Such an entity has the right to appoint a board member or a top executive (chief executive, operating, financial officer, general counsel or senior vice president), whatever the shareholding.
• It pays such an entity under a contract that gives it effective control. The statute lists about a dozen forms: deciding how much is produced and when, where components are sourced, exclusive maintenance or operating rights, control of intellectual property. A licence of intellectual property from such an entity signed or amended on or after 4 July 2025 is itself effective control, unless it is a genuine outright purchase of the technology with no reversion to the seller. The test works through payments: the company is caught for a year if, in the previous tax year, it paid such an entity under the arrangement.

The licence is the trigger that matters most. Ownership can be restructured. A company that is fully owned by local investors, builds its plant in the United States and hires local workers, but makes its cells under a Chinese technology licence signed in 2026, is a FEOC. Its own manufacturing credit is gone, and every customer who installs its cells counts the full cell cost on the FEOC side of the ratio, which in any realistic storage design makes the customer fail the ratio and lose the project credit too. No percentage in the next section can repair this, because the failure is about the company, not the product.

Test two: how much of the product cost is non-FEOC

Even a perfectly clean company can sell a non-compliant product. The material assistance cost ratio looks at the equipment in a storage system and asks what share of its cost did not come from a FEOC. The formula is the total direct cost of manufactured products, minus the cost attributable to FEOC-made products and components, divided by the total direct cost. Installation labour and civil works are not manufactured products and stay outside the ratio.

The project qualifies only if the ratio reaches the threshold for the year in which construction began. For energy storage the minimum non-FEOC share is 55 percent in 2026, 60 in 2027, 65 in 2028, 70 in 2029 and 75 percent from 2030. For cells and modules claiming the manufacturing credit the bar starts at 60 percent and rises to 80.

The first implementing guidance from the Internal Revenue Service, published in February 2026, settles three practical points. The rule looks inside the box: if a rack or module contains FEOC-made components, the share of its price attributable to those components counts as FEOC cost, no matter where the rack was assembled. A buyer may rely on a supplier's written certification of costs and origin, signed under penalty of perjury, unless the buyer knows or has reason to know it is wrong. And components that together make up less than 10 percent of direct cost can be allocated across projects without unit-level tracking; their cost and origin still count.

One detail deserves attention. The IRS default cost tables do break the cell out as a listed component inside the battery pack or module, at 52 percent of a grid-scale system, but they say nothing about where that cell was made. A supplier could be tempted to certify that its Vietnamese-built module was not produced by a FEOC, which is literally true of the module and false of its largest cost item. Because the cell is the largest single component, no buyer with basic industry knowledge can claim it had no reason to know where the cells came from. Cell-level cost and origin have to be documented.

A worked example with three sourcing plans

In 2025 the global average cell price across all applications fell to 74 USD per kWh. Stationary storage packs, cheaper than the all-chemistry average, fell to about 70 USD per kWh and became the cheapest segment in the market; a turnkey storage system was about 117 USD per kWh before installation. Even at these prices the cell is the biggest single cost in a storage system. The example below anchors the cell at 52 percent of equipment cost, the figure in the IRS's own safe-harbour table for grid-scale storage; commercial projects vary between roughly 40 and 60 percent with duration and design. The rest of the equipment splits into the inverter at 14, the container with cooling and fire suppression at 12, rack hardware at 10, controls and switchgear at 8, and the battery management system at 4.

• Plan A, Chinese cells, everything else non-FEOC, assembled in Vietnam or Türkiye: non-FEOC share 48 percent. Fails the 2026 line of 55 by seven points and can never catch the rising threshold.
• Plan B, Chinese cells plus Chinese inverter and battery management: 30 percent. Fails by a wide margin.
• Plan C, Korean or US cells with Chinese module hardware, battery management, inverter and container: 60 percent. Passes 2026, meets 2027 exactly, fails 2028 when the bar reaches 65.

Plan A is the case most people believe is safe. Every component except the cell is clean and the system was assembled outside China. It still fails. Plan C shows the opposite: a non-Chinese cell passes in 2026 and passes 2027 exactly even with Chinese balance of system, but by 2028 the rest has to be cleaned up too. Longer storage durations push the cell share toward 60 percent and make Plan A worse. Short-duration systems with heavy inverter content can push the cell toward 40 percent; such a system with Chinese cells and an otherwise clean bill of materials reaches 60 percent, clears 2026 and just meets 2027, but nothing after that. Beyond that narrow design, no cost structure keeps a Chinese-cell system above the threshold, and none survives 2028. That is the result of our cost model, not a legal rule: the statute never bans Chinese cells, it sets a ratio.

Five import routes: the tariff and the credit for each

People ask the question in a practical form: if I import modules and assemble in America, what do I get? If I import cells? If I build in Vietnam or Türkiye? The answer has two halves decided by two different agencies. Tariffs follow the customs origin of the goods. Credits follow the FEOC tests. The origin rule is the part that surprises: in the US customs rulings to date, assembling modules or packs from imported cells has not been treated as changing the origin of the product. The cell's origin has become the module's origin, case by case.

Tariff policy moved three times in 2026. In February the Supreme Court ruled that the emergency powers law does not authorise tariffs, and the reciprocal and fentanyl duties ended on 24 February. A temporary 10 percent duty ran until 24 July, and a new Section 301 action took over at 10 or 12.5 percent depending on the country. The figures below are the stack in force in September 2026 and should be confirmed with a customs broker on the day of import.

• Complete Chinese system, imported and installed. Duty about 41 percent on the battery: 3.4 base plus 25 plus 12.5 under Section 301. Credit zero, the seller is a FEOC by definition.
• Chinese modules, containers integrated in the United States. Duty about 41 percent on the modules, origin follows the cell. Credit zero, cells and modules are about 60 percent of cost.
• Chinese cells, modules and systems made in the United States. Duty about 41 percent on the cells. Project credit zero, the ratio sits at about 50 percent. The module credit of 10 USD per kWh exists on paper, but battery components need 60 percent non-FEOC material cost and the cell is well over 60 percent of a module's material cost.
• Chinese cells, modules or systems made in Vietnam or Türkiye. Duty about 41 percent on the battery portion, because assembly does not confer new origin. Customs decides origin for the article as imported, so a complete container entered as one item takes one origin; only parts entered separately and genuinely local pay the Vietnamese or Turkish rate of about 15.9 percent. Project credit zero. No manufacturing credit, nothing was made in the United States. The company itself is non-FEOC unless the triggers in test one apply.
• Non-FEOC cells from Korea, Japan, the EU or the United States, modules made in the United States. Duty about 12.5 percent on Korean and Japanese cells, about 10 percent on EU cells, zero on US-made. Project credit 30 to 50 percent if the ratio is met; the domestic content bonus additionally needs all structural steel and iron US-made and 50 percent US content in manufactured products for 2026 starts. Module credit 10 USD per kWh, 45 if the cells are also US-made.

Three conclusions follow. Assembly location buys little on either half. The first four routes give the same result on tariff and credit in any realistic cost structure, they only add labour cost. And the fifth route is the one that reliably unlocks the credits, after which the task is to clean enough of the balance of system to stay ahead of the rising threshold. None of this is a ban: the law never prohibits Chinese cells, it sets a ratio that Chinese cells make very hard to meet.

A real case: a cell plant that lost its business case

In late 2022 a battery manufacturer headquartered outside both China and the United States announced a multi-gigawatt-hour cell and module plant in the US Southeast, with production planned from 2024. The business plan was exactly the one this article describes: domestically produced non-FEOC cells at a moment when nearly every competing cell was Chinese, sold to projects and manufacturers that would earn the credits above.

In May 2026 the parent group and its subsidiary defaulted on bond payments and applied to their banks for financial restructuring, listing asset sales among the measures. The public record attributes this to leverage and financing cost, not to FEOC status. But the case is a useful example. Had that plant been financed in 2026 rather than 2022, the part of its business plan that depended on the credit would have depended entirely on tests nobody runs at the feasibility stage: the ownership chain of every US partner, the technology licensor and the equipment supplier, read to the end. In this industry those chains lead back to China more often than investors expect. A link discovered after the plant is built cannot be added to the financing that built it.

Can a company in Vietnam or Türkiye that uses Chinese cells qualify?

The position splits into the two tests. The company is not a FEOC because it is Vietnamese or Turkish. It becomes one only through the triggers above: Chinese shareholders at 25 or 40 percent, Chinese-held debt at 15 percent, a board seat, or an effective-control contract, of which a post-July-2025 technology licence from a Chinese cell or system maker is the most common. A company that buys cells at arm's length, without a licence, without supplier credit above the debt line and without Chinese shareholders, is a non-FEOC company making a FEOC product. Its customers get no project credit and its battery portion pays the Chinese tariff.

The conditions under which such a company's systems would qualify are few and clear.

• Change the cell. Non-Chinese lithium iron phosphate cells for stationary storage are reaching scale: the largest Korean producer plans more than 50 GWh of lithium iron phosphate cell capacity in North America by the end of 2026 across five sites, a substantial share of it for storage; US start-ups are adding several gigawatt-hours; and Japanese and European producers have announced lines. They cost more than Chinese cells, but the credit at stake is worth far more than the cell premium, and the battery portion drops from the Chinese tariff to the Korean, Japanese or EU rate.
• Run two product lines. Keep Chinese cells for markets without FEOC rules and build a separate, serial-number-tracked, fully documented non-FEOC bill of materials for US-bound systems.
• Sell into the US only where no credit is claimed. Projects that do not claim the credit and some commercial on-site systems can buy Chinese-cell systems today. They still pay the Chinese tariff, and defense-related buyers stop buying from October 2027 under a separate procurement ban.
• Never sign a Chinese technology licence for the US line. A licence turns a ratio problem, which sourcing can solve, into an entity problem, which only unwinding or buying out the licence can solve.

A cell made in Vietnam from Chinese raw materials

This is the question that decides most Southeast Asian cell investments, and the answer is clearer than people expect. For the storage project credit, the cost ratio works at the level of the components listed in the IRS tables: the cell, pack and module hardware, the battery management system, thermal management, the inverter, the container. The cell is its own line, at 52 percent of a grid-scale system in the current table. The test asks who manufactured the cell. It does not look further down into the cathode powder, the graphite, the electrolyte, the separator or the lithium salt inside it. Critical mineral tracing exists only in the manufacturing credit for US producers, not in the storage project credit.

So a cell genuinely manufactured in Vietnam by a company that is not a FEOC counts as non-FEOC for a US storage project today, even if nearly all of its raw materials came from China. Four conditions must all be true.

• The cell maker must not be a FEOC. Chinese equity below 25 percent single and 40 percent combined, Chinese-held debt below 15 percent, no Chinese board appointee, and no licence of intellectual property belonging to a Chinese entity signed or amended after 4 July 2025. Equipment contracts need the same scrutiny: exclusive maintenance or operating rights for a Chinese line supplier are on the statute's list of effective control. Buying Chinese production equipment is fine. Signing a contract under which only the Chinese supplier may run or service it is not.
• The cell must really be made in Vietnam. Electrode coating, winding or stacking, assembly and formation have to happen there. Importing finished electrodes or jelly rolls from China and doing formation and packing in Vietnam is not manufacturing under either the tax rule or the customs rule.
• The paperwork must exist. The buyer will ask for a certification of cost and origin signed under penalty of perjury, and the "reason to know" standard means production records, bill of materials and process descriptions have to back it up.
• The rule can tighten. The February 2026 notice is interim guidance; Treasury has said it will issue regulations and new tables. Raw materials are not traced today, but Congress is moving toward mineral tracing. Long-term supply agreements should already plan a non-Chinese second source for cathode material and graphite, at least for deliveries after 2030.

One pattern recurs in real projects. A cell plant licenses its electrode recipe from a company presented as American, and the licensor later turns out to be controlled by, or sub-licensing the technology of, a Chinese company. The statute treats any post-July-2025 right to use intellectual property that belongs to a Chinese, Russian, Iranian or North Korean entity as effective control in itself, with one carve-out for a genuine outright purchase of the technology; the licensee becomes a FEOC and every cell it makes is a FEOC product. Read the licensor's ownership chain and the chain of title of the technology to the end before signing, not after the plant is built.

Moving cell production out of China is not enough on its own. A cell plant in Vietnam, Türkiye or the United States still fails if the cell maker's shareholders, lenders or technology licensor turn out to be FEOC. The cell then counts as FEOC cost for every customer, and the plant itself loses its own manufacturing credit.

Why these rules will stay

China produces about 80 percent of the world's lithium-ion cells and the overwhelming majority of lithium iron phosphate cells, plus roughly 85 percent of cathode active material, more than 90 percent of anode material and the majority of separator and electrolyte capacity. Chinese lithium battery exports to the United States were about 15 billion USD in 2024, a quarter of China's total battery exports. The US rules are aimed at exactly this dependency, and the thresholds rising every year to 2030 tell the market to build the alternative itself. Until non-FEOC cell supply in North America reaches tens of gigawatt-hours a year, every non-FEOC claim deserves the same two questions: whose cells, and what share of the cost?

eMOBINO screens suppliers and partners against the ownership, debt, board and licence tests, models the cost ratio of a bill of materials against the 2026 to 2030 thresholds, and sources non-FEOC cells and balance of system through eMOBINO Gateway. If a US project or a US-bound product line is on your table, write to info@emobino.com.

The full carousel

1 / 8

References

• One Big Beautiful Bill Act, Public Law 119-21, 4 July 2025, Internal Revenue Code sections 7701(a)(51) and (a)(52), 48E, 45X, 6418 and 6695B
• Internal Revenue Service, Notice 2025-08 (domestic content safe harbor tables for battery energy storage) and Notice 2026-15 and accompanying newsroom release on material assistance from prohibited foreign entities, 12 February 2026
• US Department of Energy, Interpretation of Foreign Entity of Concern, final interpretive rule, Federal Register, May 2024
• Supreme Court of the United States, Learning Resources v. Trump, 20 February 2026; US Trade Representative, Section 301 actions on lithium-ion batteries, 2024 and 2026
• US Customs and Border Protection, ruling N329847 on origin of modules assembled from imported cells (H316389 consistent) and N335526 on classification of containerised storage systems
• US Department of Commerce, final determinations on active anode material from China, February 2026, and US International Trade Commission negative injury determination, March 2026 (no duty orders issued)
• National Renewable Energy Laboratory, Cost Projections for Utility-Scale Battery Storage: 2025 Update; BloombergNEF energy storage system cost survey 2025
• BloombergNEF, Lithium-Ion Battery Price Survey, December 2025
• International Energy Agency, Global EV Outlook 2026 and Global Critical Minerals Outlook 2025
• State government announcement of the plant, late 2022; public disclosures on the parent group's bond default and restructuring application, May 2026