Aneka Tambang

1. 01Mine nickel ore and bauxite

   State-owned upstream operations produce nickel ore, ferronickel and bauxite, the traditional core of the company.

   EconomicsThis is the capital-intensive half: it carries the reserves, the permits and the RKAB quota exposure, and it is where the asset base sits.

2. 02Refine gold, most of which is not your own

   The Logam Mulia refinery processes and sells gold bullion, a large part of it sourced rather than mined, which makes this a processing and trading operation more than a mining one.

   EconomicsBought-in gold passes through at a thin spread, so it inflates revenue far more than profit. This single fact explains the whole margin profile below.

3. 03Sell it to Indonesian retail savers

   Gold volume rose 20 percent to 34.1 tonnes, about 1.1 million troy ounces, in the first nine months of 2025, and the company credits the Logam Mulia digital application for simplifying physical gold purchases. Domestic sales were 96 percent of the total.

   EconomicsRead ANTM as a domestic gold savings channel with a mine attached, not as an exporter. That is unusual among Indonesian miners and it changes which risks matter.

4. 04Carry the bullion while you wait

   Trading physical gold means holding physical gold. The cash conversion cycle stretched from 31.1 days in FY2024 to 42.1 days in FY2025 as volumes surged.

   EconomicsThat is why cash did not keep up with profit in FY2025, and it is the cost of running a trading book rather than a mine.

Cost structureDominated by purchased gold, which is why the margin line and the revenue line tell opposite stories. Revenue more than doubled from IDR 41.0tn in FY2023 to IDR 84.6tn in FY2025, yet gross margin over the same stretch went 16.8 percent, then 10.2 percent, then 17.5 percent. A business whose revenue doubles while its gross margin first collapses is passing volume through at a thin spread, not becoming more profitable per unit. Operating margin of 9.5 percent in FY2025 is still below the 10.5 percent of FY2021 despite revenue being 2.2 times larger. Interest is negligible: IDR 78.5bn against IDR 8.0tn of operating profit, cover of 102 times.

Cash cycleShort in absolute terms but moving the wrong way: 34.0, 30.1, 30.3, 31.1 and 42.1 days. The 11 day extension in FY2025 coincides with the gold volume surge, which is exactly what a growing trading book does to working capital. It is well funded, with IDR 8.4tn of cash against IDR 4.6tn of debt, so this is a margin and cash-timing issue rather than a solvency one. But it is the mechanism behind the single most important number on this page, which is the collapse in cash conversion.

Unit economicsTwo ratios move in opposite directions and the tension between them is the company. Asset turnover rose from 0.96 times in FY2023 to 1.61 times in FY2025 as gold volume surged, which lifted return on equity from 10.0 percent to 20.4 percent, the best in the five-year window. But operating margin over the same period only went 7.8 percent to 9.5 percent, still below FY2021. So the return improvement is a TURNOVER story, not a margin story: ANTM is earning more because it is pushing far more metal through the same asset base, at a spread no better than before. That works while volume grows and reverses immediately when it stops, which is a different risk profile from a miner whose returns come from ore grade or cost position.

The headline is excellent and the cash line is the warning, so both belong on the page. FY2025 delivered record revenue of IDR 84.6tn and record net profit of IDR 7.2tn, nearly double the IDR 3.6tn of FY2024, lifting return on equity to 20.4 percent. Externally verified: gold volume rose 20 percent to 34.1 tonnes in the first nine months of 2025 with nine-month profit up 197 percent, and the company attributes the demand to its Logam Mulia digital application. Now the catch. Operating cash flow covered net income 2.71, 1.08, 1.42, 1.01 and just 0.67 times across FY2021 to FY2025. In the record year, cash generation of IDR 4.9tn came in BELOW reported profit of IDR 7.2tn for the first time in the window, and it was lower in absolute terms than the IDR 5.0tn generated in FY2021 when profit was a quarter the size. The explanation is not accounting: a gold trading book absorbs working capital as it grows, and the cash cycle duly stretched from 31.1 to 42.1 days. Nothing here is improper, and there are no acquisition gains or revaluations. But a doubling of profit that produces less cash than four years ago is a signal about the KIND of growth, and readers should weight the cash line accordingly.

Cash conversionDeteriorating, and it is the number to watch rather than the profit. Operating cash flow covered net income 2.71, 1.08, 1.42, 1.01 and 0.67 times. Free cash flow nonetheless stayed POSITIVE in all five years, at IDR 4.5tn, 3.4tn, 2.9tn, 2.5tn and 4.2tn, because capital spending is modest for a miner at IDR 0.5tn to 1.4tn a year. So the business is not consuming cash; it is simply converting a much larger profit into no more cash than it did five years ago. Judge ANTM on gold volume, spread per unit and the cash cycle together, since the profit line alone now moves for reasons the cash line does not follow.

Conservative to the point of caution, which is unusual for a state miner in an upcycle. Capital spending never exceeded IDR 1.4tn in any year, or about 1.7 percent of FY2025 revenue, while equity grew from IDR 20.8tn to IDR 36.6tn and the group moved to a net cash position of about IDR 3.9tn. Debt fell to IDR 283bn in FY2024, essentially nil, before rising modestly to IDR 4.6tn. The company has funded a doubling of throughput almost entirely from its own cash flow.

DeploymentVery little goes into the ground, which is the strategic question rather than a criticism. Capital spending ran IDR 0.5tn, 0.7tn, 1.4tn, 1.2tn and 0.7tn while revenue more than doubled, so the growth came from moving more traded metal rather than from developing reserves. Working capital took the second call as the cash cycle lengthened 11 days in FY2025. The rest went to the balance sheet: cash rose from IDR 5.1tn to IDR 8.4tn and net debt turned into net cash. That leaves an unresolved allocation question. Indonesia is cutting nickel ore quotas by roughly 30 to 34 percent for 2026 and copper and bauxite exports are already banned in raw form, so upstream value increasingly requires processing capacity, and IDR 0.7tn of annual capital spending does not build much of it.

Returns trendReturns improved sharply and for a specific reason. Return on equity went 8.9, 16.1, 10.0, 11.6 and 20.4 percent, with return on invested capital tracking closely at 9.1, 16.7, 10.2, 11.7 and 19.9 percent because there is almost no debt, so the two measures agree and neither is flattered by leverage. The driver was turnover rather than margin, as set out above. On valuation the model is cautious where the market is not. The forward model is run on a cyclically normalised basis and produces IDR 2,597.60 per share against a price of IDR 2,879.78, a gap of MINUS 9.8 percent, so on normalised assumptions the shares trade slightly above value. Against that, EV/EBITDA of 7.28 times sits 29 percent BELOW the mining peer median of 10.31 times. The two tests disagree, and the reason is instructive: the relative multiple is measured on current earnings, which are at a cyclical high, while the model normalises them. Anyone using the peer discount as the argument is implicitly betting that FY2025 volumes and spreads persist. Both are model outputs, not targets.