Samsung SDI Co., Ltd. (KRX:006400) made money at the operating line in the second quarter of 2026. That sentence has not been true since the third quarter of 2024.
The intervening run reads: negative ₩256.7bn, negative ₩434.1bn, negative ₩397.8bn, negative ₩591.3bn, negative ₩299.2bn, negative ₩155.6bn. Six quarters, ₩2.13tn of cumulative operating losses, at a company with ₩26.84tn of equity. Then ₩203.8bn positive.
The shares rose 10.27% on August 27 to close at ₩569,000, for a market capitalisation of ₩45.85tn. It's worth separating which parts of this turn look durable and which parts do not.
Start with the strongest evidence. Gross margin in the quarter was 24.87%, on gross profit of ₩937.3bn against revenue of ₩3.77tn.
Four quarters earlier it was 8.83%. At the trough, in the third quarter of 2025, it was 5.52%, when the company generated ₩168.3bn of gross profit on ₩3.05tn of revenue. A battery maker running a 5% gross margin is selling cells for barely more than the materials cost.
Sixteen points of gross margin in four quarters is not a cost-cutting story and not an accounting artefact. It is a mix and pricing change, and given what the company has been saying about energy storage and what its Indiana decision implies, storage cells are the likeliest source. Revenue also grew, to ₩3.77tn from ₩3.18tn, up 18.5%, so this is not margin bought by shrinking.
For scale against history: FY2023 revenue was ₩21.44tn with a 17.64% gross margin. FY2025 was ₩13.27tn at 11.02%. The current quarter is running above the good-year margin on a much smaller revenue base, which is what a genuine mix improvement looks like.
Now the part that deserves scepticism. Pretax income was ₩479.2bn. The income tax charge was ₩7.6bn. That's an effective rate of 1.6%.
Samsung SDI's normal effective rate is nowhere near that. In FY2022 it paid ₩612.9bn on ₩2.65tn of pretax income, about 23%. In FY2023, ₩403.1bn on ₩2.38tn, about 17%. Apply even 20% to this quarter and net income would have been roughly ₩383bn rather than the ₩471.6bn reported.
There is a benign and probably correct explanation. A company emerging from a run of losses typically recognises deferred tax assets against those carried-forward losses once profitability returns, and that recognition shows up as a credit that offsets the current-period charge. Look back and you can see the losses being taken the other way: the tax line was negative ₩135.5bn, negative ₩165.5bn, negative ₩166.6bn and negative ₩21.6bn in the four quarters before Q1 2026's negative ₩99.5bn.
So nothing improper. But it is not repeatable. A 1.6% tax rate is a one-time benefit of having lost a great deal of money, and anyone annualising ₩471.6bn of quarterly net income into a valuation is capitalising it.
At ₩569,000 the market capitalisation of ₩45.85tn sits at about 1.7 times the ₩26.84tn of book equity. Annualise the quarter's reported net income and it's roughly 24 times. Annualise it at a normal tax rate and it's closer to 31 times. For a company that lost money for six quarters, either number is a lot of faith in one print.
The cash flow statement is the third thing to check and it does not corroborate the income statement.
Cash from operating activities for the entire first half was ₩109.4bn, against ₩701.6bn in the same period of 2025 when the company was losing money at the operating line. More profit, less cash.
The balance sheet says where it went. Inventories went from ₩2.94tn at December 31 to ₩3.68tn at June 30, up 25.2% in six months. Trade receivables went from ₩3.08tn to ₩3.74tn, up 21.6%. Revenue in the half grew 15.6%.
Both lines growing faster than sales is what you expect from a company ramping production into a demand recovery, and it is also what you see when product is being built ahead of orders. The difference matters and the half-year summary does not settle it. Energy storage systems are shipped in large container-scale units against project schedules, which lumps both inventory and receivables in ways that a monthly-shipment business would not, so some of this is structural to the segment.
Cash stood at ₩1.51tn at June 30, down from ₩1.80tn at December 31, against total liabilities of ₩20.79tn.
None of the above would matter much if the company had run out of room, and it did not, because it raised money at the bottom.
Share capital went from ₩356.71bn at March 31, 2025 to ₩415.82bn at June 30, 2025. At the ₩5,000 par value that implies 11,821,000 new shares against roughly 71.3m existing, dilution of about 16.6%. Financing activities brought in ₩1.98tn in that quarter, though the figure includes borrowing as well as equity.
Diluting existing holders by a sixth, in the middle of the worst stretch in a decade, is expensive. It also meant the company entered 2026 with a balance sheet that could absorb the Indiana decision and fund an Ulsan programme without a second raise. Total equity is ₩26.84tn today against ₩21.14tn in the third quarter of 2024, despite ₩2.13tn of operating losses in between.
Capital spending has come down hard too: ₩6.27tn in FY2024, ₩3.07tn in FY2025, ₩1.13tn in the first half of 2026. Management stopped building while demand was absent, which is the right sequence and not one every battery maker managed.
The third quarter, reported in late October, and two numbers in it. Gross margin, which should hold near the mid-twenties if the mix change is real. And the effective tax rate, which will tell you how much of the second quarter's ₩471.6bn was operating recovery and how much was the accounting catch-up on prior losses.
The second marker is the operating cash flow line for the nine months. If it stays near ₩100bn while reported profit accumulates, the working capital build is absorbing everything the turn generates, and the recovery is on paper before it is in the bank.
kstock reads DART every morning and writes up what moved — the contract, the buyback, the number that does not add up. The daily post and a Saturday roundup, by email.