How 50 GWh of FEOC-Compliant ESS Capacity Becomes 11

Clean Energy Associates estimated in January 2026that FEOC-compliant ESS cell capacity in the US would carry a roughly 10% surplus this year, with Korean suppliers representing more than 80% of the compliant base. Wood Mackenzie and the American Clean Power Association placed 2025 US storage installations at 49 GWh. LG Energy Solution, Samsung SDI, and SK On have collectively targeted over 50 GWh of US-based ESS cell capacity by year-end. That 10% surplus figure has become the default planning number in procurement circles, the figure that shows up in board decks and sourcing recommendations. The arithmetic behind it is straightforward.

The arithmetic is also wrong, in the specific sense that matters to a buyer entering the market this quarter for 2026-2027 delivery. Four constraints sit between announced nameplate capacity and procurable, credit-eligible supply. Each one removes volume. Applied sequentially, they leave credit-dependent buyers facing a functional shortage and everyone else facing a contested market. The constraint accounting follows, sources named and inference methods visible.

The utilization discount

The CEA surplus estimate implicitly assumes full utilization of nameplate capacity. Korean battery makers' own disclosures make this assumption indefensible.

Seoul Economic Daily reported in March 2026 that all three Korean majors saw global utilization collapse below 50% in full-year 2025. LG Energy Solution disclosed 47.6%, down from 73.6% in 2022 and 57.8% in 2024. Samsung SDI fell from approximately 84% to roughly 50% over the same period. SK On dropped from 86.8% to 48.7%. The Korea Herald's half-year data from August 2025 confirmed the trajectory mid-year: all three hovering around 50-52%.

An important caveat: these are global figures blending geographies, chemistries, and end markets. ESS-dedicated US lines may run hotter than idled EV lines in Korea or Hungary. But no public facility-level utilization disclosure exists for any US ESS plant from any of the three companies. I have looked. The StarPlus Energy facility in Indiana was estimated by industry sources at below 60% utilization in H1 2025. SK On's US plants ran near full capacity in the same period, but that was EV production for Hyundai, not ESS.

Absent direct plant data, I am bracketing. A newly ramping ESS line converting from EV tooling, in a demand environment where the parent company is running sub-50% globally, plausibly operates in the 65-80% range. At 75%, the 50+ GWh nameplate figure drops to approximately 38 GWh of actual output. This is an assessed midpoint. No sourced figure exists for US ESS-specific utilization. Treat it accordingly.

Anchor contracts have claimed the output

The 38 GWh of utilization-adjusted capacity is not sitting on a spot market. It is substantially allocated through multi-year agreements signed before any new buyer enters the queue.

The publicly disclosed commitments:

• LG Energy Solution signed a $4.3 billion LFP cell deal with Tesla for Megapack production, with start of production targeted for August 2027. LGES's Holland, Michigan facility began ESS LFP production in May 2025; the Lansing facility is being built to serve this contract.
• Samsung SDI committed approximately 10 GWh/year to Tesla ESS, valued at $2.1 billion, reported November 2025. Samsung SDI also disclosed over KRW 2 trillion ($1.4 billion) in a three-year supply agreement with an unnamed US buyer starting 2027.
• SK On signed a $1.4 billion agreement with Flatiron Energy for up to 7.2 GWh through 2030.

Summing the disclosed annual volumes where they can be estimated: conservatively 20-25 GWh per year is spoken for. Against 38 GWh of utilization-adjusted output, that leaves roughly 13-18 GWh uncommitted. No public data quantifies the residual with more precision. The signal from the contract announcements is that uncommitted volume is thin, and broader availability may not materialize until 2028.

The analysis splits at this point. A buyer indifferent to 45X or 48E credit eligibility faces a tight but navigable market. The 13-18 GWh residual is contestable, pricing will reflect the scarcity, but physical cells exist. For a buyer whose project economics depend on tax credit eligibility, two additional constraints bind, and they bind hard.

The graphite cliff at year-end 2026

The exposure is structural. China controls over 90% of global graphite anode material production. BTR, the world's largest anode producer, was classified as a FEOC by DOE in January 2025, with the designation extended to overseas subsidiaries in Indonesia and Morocco. This closed the most obvious non-Chinese scale workaround.

Non-Chinese anode supply at commercial scale:

Syrah Resources operates the Vidalia, Louisiana facility, the only commercial-scale vertically integrated natural graphite active anode material plant outside China, with nameplate capacity of 11.25 ktpa. At roughly 1 kg of graphite anode material per kWh of cell capacity, 11.25 ktpa supports approximately 11 GWh of cell production at full nameplate. Syrah has not demonstrated full nameplate output. It has not delivered conforming AAM samples to its anchor customer, Tesla. The dispute resolution deadline has been extended four times, most recently to June 1, 2026. Whether that deadline has been met is not publicly confirmed as of this writing. The expansion to 45 ktpa remains pre-FID. Syrah's CEO stated in 2024:

Eleven GWh is a ceiling, and one Syrah has not demonstrated it can reach. Against 38 GWh of utilization-adjusted Korean ESS capacity, the gap is 27 GWh of cell production with no verified non-Chinese anode source.

NOVONIX is building synthetic graphite capacity in Chattanooga but disclosed at its April 2026 AGM that battery-grade anode material for Panasonic will not reach mass production until H2 2027. Industrial-grade graphite, which began production in 2025, does not meet battery cell specifications. The Stellantis offtake has been terminated.

No other non-Chinese graphite anode supplier has confirmed commercial-scale battery-grade production for the US market in 2027. Japanese producers serve established Panasonic supply chains with no publicly disclosed US ESS-directed agreements.

After December 31, 2026, any Korean cell maker sourcing graphite anode from Chinese or Chinese-affiliated suppliers produces cells that do not qualify for 45X or 48E credits. The FEOC-compliant capacity figure contracts to whatever can be supplied with non-Chinese anode, which at verified commercial scale is approximately 11 GWh at best.

The LFP cathode sourcing gap and MACR arithmetic

All three Korean majors are converting to LFP chemistry for ESS. All three face the same problem: LFP cathode supply chains are overwhelmingly Chinese, and MACR thresholds demand increasing non-PFE content.

The MACR math makes this concrete. Anode and cathode materials together typically represent 60-70% of a cell's applicable critical mineral content value. The 2026 threshold requires 60% non-PFE content. If both anode and cathode are PFE-sourced, a cell cannot reach 60% non-PFE regardless of what other components are clean. If one is PFE-sourced and the other is not, compliance is possible at the 60% threshold depending on other mineral sourcing, but tight. The trajectory doesn't loosen:

Year	Non-PFE threshold	Compliance feasibility with one PFE input
2026	60%	Tight but possible
2028	70%	Marginal at best
2030	85%	Effectively impossible

A partial workaround that barely clears 60% in 2026 fails at 70% two years later. A buyer signing a multi-year offtake needs to understand that the compliance problem compounds.

As of May 2026, the sourcing routes:

Samsung SDI has the cleanest path. In March 2026, it signed a binding three-year contract with L&F, a South Korean cathode producer, for KRW 1.6 trillion ($1.1 billion) of LFP cathode material. L&F is building 60,000 metric tons of annual LFP cathode capacity in Daegu. L&F is not a FEOC. The contract starts in 2027. Samsung SDI plans to begin mass LFP ESS production at StarPlus Energy in Q4 2026. What cathode material the Q4 2026 production uses is not publicly disclosed.

SK On signed an MOU with L&F in July 2025 covering LFP cathode for the North American market. It remains an MOU. Volume, pricing, and start date have not been publicly confirmed. SK On's H2 2026 ESS production ramp at Commerce, Georgia does not have a publicly identified FEOC-compliant LFP cathode supplier.

LG Energy Solution is sourcing through LG Chem's Morocco joint venture with Huayou Group, announced September 2023 with a 50,000 metric ton annual capacity target and 2026 mass production timeline. The MOU language acknowledged the FEOC problem: LG Chem and Huayou's subsidiary were "to adjust share in compliance with the FEOC regulations." No public confirmation of the final equity structure, FEOC compliance determination, or operational commissioning has been found. If Huayou retains a stake that triggers PFE classification, LG Chem's LFP cathode is MACR-noncompliant. A binary unknown with no publicly available resolution.

POSCO Future M is building additional Korean LFP cathode capacity, with mass production expected H2 2027. Its investment vehicle is a joint venture with China's CNGR. No public FEOC determination on this entity has been found.

Binding, FEOC-clean LFP cathode supply at scale begins in 2027, not 2026. Samsung SDI's L&F contract is the first credible non-Chinese LFP cathode commitment, and it doesn't deliver material until next year. NOVONIX's battery-grade anode reaches mass production H2 2027 at the earliest. Both the graphite constraint and the cathode constraint bind simultaneously through roughly the first half of 2027. No verified non-Chinese anode supply at scale, no verified non-Chinese LFP cathode supply at scale. For credit-dependent procurement targeting H1 2027 delivery, this is the tightest point in the funnel.

Counterparty risk as a supply variable

The three companies constituting 80%+ of FEOC-compliant ESS cell supply posted combined losses of KRW 3.2 trillion ($2.2 billion) in 2025, excluding US production tax credits. Worth restating: without AMPC revenue from 45X credits, the entire Korean cell manufacturing base serving the US ESS market is loss-making. LGES has disclosed AMPC revenue in its quarterly earnings as a separate line item; the gap between reported operating income and operating income excluding AMPC is the measure of dependency. I do not have the specific AMPC figure from LGES's most recent quarterly filing in the data I can verify for this piece, and I will not estimate it. The KRW 3.2 trillion combined loss figure, sourced from Seoul Economic Daily's analysis of annual reports, establishes the structural point: profitability is contingent on a US policy variable.

SNE Research data through February 2026 shows all three posting negative year-on-year growth in global EV battery usage: LGES down 2.7%, SK On down 12.9%, Samsung SDI down 21.9%, driven by a 29.8% plunge in US EV sales. SK On laid off 958 workers at its Commerce, Georgia facility in March 2026, approximately 37% of the workforce, though the cuts targeted EV lines rather than ESS conversion.

For a procurement lead evaluating 2027 delivery commitments, this matters directly. A supplier whose manufacturing operations are loss-making absent a single US tax credit is a supplier whose capacity commitments carry policy risk the contract language does not price. The confidence interval around 2027 delivery volumes is wider than the signed terms imply, regardless of any view on 45X's legislative durability.

The assessed balance

The funnel, with each constraint applied:

Constraint stage	Volume	Basis
Nameplate FEOC-compliant ESS capacity	50+ GWh	CEA estimate, company targets
After utilization discount (75% assessed)	~38 GWh	Inferred from global utilization data
After anchor contract commitments	13–18 GWh uncommitted	Disclosed offtake agreements
Credit-dependent: after graphite exemption expiry	~11 GWh ceiling	Syrah Vidalia nameplate, undemonstrated

For credit-indifferent buyers, the market at the 13-18 GWh stage is tight but not short. Cells exist. Pricing reflects scarcity. Physical procurement is feasible if the buyer accepts competition for residual volume.

For credit-dependent buyers, the additional constraints bind. After the graphite exemption expires December 31, 2026, the FEOC-compliant base contracts to whatever volume Korean cell makers can produce using non-Chinese anode material. At verified commercial scale, that ceiling is roughly 11 GWh from Syrah's Vidalia plant, undemonstrated at nameplate, with its anchor customer relationship unresolved. The LFP cathode gap introduces a parallel compliance risk for 2026 production and early 2027 that Samsung SDI's L&F contract partially addresses starting mid-2027 but that LGES and SK On have not publicly resolved. The MACR trajectory from 60% to 85% non-PFE means partial workarounds available in 2026-2027 do not survive into the 2028-2030 contracting horizon.

The nominal 10% surplus dissolves. For credit-dependent procurement at scale in 2027, the market is functionally short. The price has not yet reflected this because the headline capacity figures continue to circulate without the constraint stack applied. Whether Korean cell makers have secured non-Chinese anode and FEOC-clean cathode supply through undisclosed channels is the single largest variable in this assessment. If they have, the tightness moderates. If they haven't, the gap between quoted availability and procurable, credit-eligible supply is the widest it has been since FEOC rules took effect.

That gap is where the sourcing risk sits. The data available today cannot close it. A buyer making a commitment this quarter should price accordingly.