SK On Q2 2026: KRW 821.8B Operating Profit, None of It Decomposable

SK On reported a KRW 821.8B operating profit for Q2, a KRW 1.17T swing off Q1's loss. SK Innovation named customer compensation and higher AMPC as the drivers and put a figure on neither. The sign of recurring operating profit — whether the cell business made or lost money before credits and settlements — cannot be determined from public disclosure.

Trailing Four-Quarter Decomposition

SK On's battery segment as reported within SK Innovation's consolidated disclosures. Q2 2026 USD equivalents at KRW 1,500.48/USD, the average rate for the quarter.

^1 Originally reported as KRW 1.7912T in the Q1 deck; subsequently presented as KRW 1.870T on a restated credit-presentation basis in Q2 materials. The original filing figure is used here.

^2 Subsequently shown as KRW 440.8B in the Q1 deck; the KRW 0.6B revision was not explained.

^3 Disclosed in Korean-language releases only. English IR decks omitted the quarterly AMPC figure in both periods.

Three of the seven rows — ASP direction, utilization, and EV/ESS mix — are blank across all four quarters, and AMPC moved from quantified in Q3–Q4 2025 to absent from Q1 2026 onward. For comparison: LG Energy Solution, a standalone listed entity, itemizes its 45X contribution and publishes a direct ex-credit operating bridge. Samsung SDI discloses consolidated AMPC without allocating it to the battery segment, and neither company published a comparable utilization percentage this quarter. SK On, reporting as a segment inside SK Innovation, gives none of the three. The gap is wide enough that the decomposition this publication runs on the other two Korean makers cannot be run here. What follows maps the boundary between what the disclosure supports and what it does not.

Pass 1 — Four Economic Vintages, None Quantified

The prior cycle piece left the sign of SK On's recurring operating profit open. This pass tries to assign the KRW 821.8B across the four vintages that make up the reported number.

AMPC/policy credits. SK Innovation's Q3 2025 release disclosed KRW 173.1B of quarterly AMPC and KRW 617.3B cumulatively through September 2025. The Q4 release disclosed KRW 101.3B. Both appeared only in Korean-language materials; the English decks carried neither. From Q1 2026 the quarterly figure disappeared in both languages. Q2 names "increased AMPC" as a driver and stops there.

The quarterly figure was published while the segment was posting losses and withdrawn in the quarters where results began to improve. Whatever the reason, a reader trying to separate credits from operations has fewer inputs now than a year ago, at the point where the separation matters most.

A Hana Securities model referenced in the prior piece put Q2 AMPC at approximately KRW 130B. That is a sell-side estimate, not issuer disclosure, and it stays out of the table.

Customer compensation. The public record neither ties the Q2 compensation item to BlueOval SK nor rules it out. SK Innovation attributed part of the quarter's improvement to customer compensation and disclosed nothing about payer, program, triggering event, covered period, or amount.

The BlueOval SK dissolution supplies a plausible context. The Ford joint venture closed May 20. Ford's closing 8-K shows its membership interest redeemed, Ford assuming a $3.805B DOE promissory note and the full $250M Kentucky forgivable loan, and Ford taking the two Kentucky plants. Ford separately projected roughly $500M in cash expenditures associated with closing the disposition, without naming SK On as a recipient or characterizing the amount as compensation. Ford's Q1 10-Q recorded $103M in charges for cancelled EV program commitments and did not name the counterparty.

If the compensation is dissolution-related, it is a one-time settlement carrying no information about SK On's recurring cell economics. If it is not, then some customer program generated a payment large enough to move a segment from a KRW 349.2B quarterly loss to a KRW 821.8B profit, and nothing in the public record says which program.

Current recurring battery operations. SK Innovation cited higher Asian sales as a Q2 driver with no revenue breakdown, margin, volume, or ASP behind it. With neither the AMPC amount nor the compensation amount known, the residual — what the business earned from making and selling cells — cannot be isolated by subtraction.

Hana Securities modelled that residual at negative KRW 408.2B, which would put the underlying cell business in loss even in its best reported quarter in two years. That figure rests on AMPC and compensation assumptions that cannot be checked against issuer data. Useful as a directional sell-side read; not a substitute for the disclosure that isn't there.

ESS ramp costs. Georgia conversion, GRIDON development, the Tennessee restructuring, the Korean BESS programs — none of it carries a separate expense allocation in the Q2 result. These costs sit somewhere inside the KRW 821.8B and are invisible at segment reporting granularity.

Pass 2 — Utilization: No Figure, No Denominator, Indirect Physical Evidence

SK On has not published a quarterly utilization percentage in any of the four trailing quarters, and it does not state whether its capacity denominator is nameplate, effective, or production-weighted. Annual figures compiled by Seoul Economic Daily showed SK On at 86.8% in 2022 falling to 48.7% in 2025. No named source in the research context carries a Q2 2026 estimate.

The physical record gives a partial read. Commerce consolidated from two plants to one in April, halting Plant 1 (10 GWh nameplate) and shifting volume to Plant 2 (12 GWh). June's layoff of 958 workers, roughly 37% of the Georgia workforce, is consistent with a site running well under nameplate. SK's corporate profile still describes two operating EV plants in Georgia four months after the consolidation.

Revenue rose 64.5% QoQ, from KRW 1.791T to KRW 2.946T. That could be volume, ASP, the compensation item, or some mix. Until the compensation amount is isolated, the revenue line cannot be used as a proxy for either shipped volume or pricing direction.

Pass 3 — Guidance vs. Actuals

The Q1 presentation said only that battery "long-term profitability" would improve on European sales growth and North American ESS expansion — without a Q2 revenue forecast, operating-profit target, ASP direction, or utilization figure. A KB Securities note from the Q1 call described European utilization improving and North American utilization weak, without percentages.

There is no numeric prior-quarter benchmark to test Q2 against. This pass cannot be performed for SK On.

The Income Statement and the Facility Record

The two run on different clocks. Q2's income statement records a quarter in which settlements and credits produced KRW 821.8B of reported profit. The US facility record records a multi-year EV-to-ESS conversion that has not yet produced publicly confirmed accepted output at any site.

Georgia is a single-plant operation since April, with 37% fewer workers since June. The Flatiron Energy agreement supplies a named BESS customer for an initial 1 GWh from converted Commerce capacity, deliveries from H2 2026, with preferential negotiation rights over a further 6.2 GWh through 2030. No public source confirms the production line, LFP cell yield, first accepted output, or customer release.

Tennessee became a standalone SK On company in May, with production projected for 2028 and no committed customer program behind that date.

GRIDON targets US production later in 2026 for its first generation and Q3 2027 for second-generation commercial production, naming no customer and no contracted volume.

Korean BESS tenders are the firmest execution evidence in the set. SK On won 284 MW, 50.3% of the second centralized tender, for supply by end-2027. Contracted volume, defined timeline — and Korean domestic, not separately identified in the Q2 result, and irrelevant to the US supply question.

Sourcing Decision Frame

For a team weighing SK On as a FEOC-compliant US cell source, Q2's result carries no usable signal on recurring cell economics, US production capability, or ESS conversion progress. The reported profit may be substantially or entirely composed of one-time compensation and policy credits; the underlying business may still be loss-making. The US footprint is contracting while the ESS pivot is pre-acceptance across the board: Flatiron's initial 1 GWh has no confirmed first output, Tennessee has no customer, GRIDON has no customer. Qualification work here has to run off facility-level evidence, because the income statement will not support it. And the disclosure gap itself belongs in the risk assessment — a supplier whose segment reporting cannot be decomposed is a supplier whose deterioration a buyer will see later than it would at LG or Samsung SDI.