Sociedad Química y Minera de Chile S.A. (SQM) | The Buildout — AI Infrastructure

The Verdict

SQM is a Chilean mining and chemicals producer. It extracts brine at the Salar de Atacama and hard rock at Mount Holland in Western Australia, then sells lithium carbonate, lithium hydroxide and spodumene concentrate alongside iodine, specialty plant nutrients, potassium and industrial chemicals. The only route from the AI buildout to SQM runs through grid-scale battery storage: its lithium compounds go into the batteries that back up data centers and the grid, with the battery-storage slice of demand estimated by management at about 30% of the global lithium market. SQM is a price-taker in a globally traded commodity — management says its contracts are all index-based — and it names no AI customer and makes no AI-specific disclosure.

Earnings

In the June 2026 quarter, reported August 19, 2026, SQM recorded US$2,468.7 million of revenue, a 51.1% gross margin, US$1,314.6 million of EBITDA (53.2% of revenue) and US$660.1 million of net income. The operating headline was a record: lithium sales volumes of more than 84,000 metric tons of lithium carbonate equivalent, from Chile through Novandino Litio and Australia through Covalent Lithium. Management raised its 2026 global lithium demand outlook to more than 2.1 million metric tons and guided Novandino's 2026 production to 280,000-290,000 t LCE. The call itself carried no revenue, margin or segment P&L detail.

Management tone: Across the three calls in the window, tone moved from an inflection call in March 2026 — "we saw the inflection point in lithium prices" — to a confident, delivered Q1 in May, when the average realized price came in at roughly US$18 per kilo against US$10 in Q4 2025. The August call was enthusiastic on volumes and careful on price: "We had a very good second quarter. We're very enthusiastic about how the market is doing," paired in the same answer with "nobody can predict where the indices are going." The consistent pattern is specific on volumes, demand, project scheduling and cost, and unwilling to underwrite price beyond one quarter. Management declined to disclose the China tolling contract and gave no capital-allocation or dividend update at all on the Q2 call.

Management Guidance

Management guided Novandino Litio production to 280,000-290,000 t LCE for 2026, rising to ~300,000 t of capacity in 2027, with the 2027 split stated as 80,000 t of lithium sulfate, 220,000 t from the Antofagasta plant and 20,000-30,000 t of hydroxide. Global lithium demand for 2026 was raised to more than 2.1 million metric tons. For Q3 2026 the guide is flat sequentially: volume in line with Q2, price more or less in line with the first-half average, cost similar to Q2, with full-year 2026 lithium cost reaffirmed as lower than 2025. Iodine production was guided to ~15,500 t with sales volume "similar to what we see last year," and specialty plant nutrition volumes to +~10% versus 2025, with supply-constrained conditions expected to normalize toward the end of the year. Capital expenditure is framed at approximately US$3 billion over 2026-2028. Mount Holland's expansion is guided to start construction mid-2027, commission in 2029 and produce first product in the first half of 2030.

Trajectory

Revenue moved from US$1,036.6 million in the March 2025 quarter to US$1,173.0 million in September 2025, US$1,335.7 million in December 2025, US$1,760.0 million in March 2026 and US$2,468.7 million in June 2026. EBITDA margin over the same stretch went from 23.8% and 24.8% to 38.5%, 47.5% and 53.2%. Two drivers are working together. Price: realized lithium moved from close to US$10 per kilogram in Q4 2025 to roughly US$18 in Q1 2026, with Q2 higher again and Q3 guided to the first-half average. Volume: sales ran above 66,000 t in Q4 2025, ~69,000 t in Q1 2026 and a record above 84,000 t in Q2 2026. Unit costs fell alongside, and management guides full-year 2026 lithium cost below 2025. The one offset is that the Q3 guide is a plateau, not a further step-up.

The Model

The model's locked projections put FY+1 revenue at US$8,776 million with EBITDA of US$4,406 million, a 50.2% margin, and FY+2 revenue at US$9,461 million with EBITDA of US$4,693 million, a 49.6% margin. The near-term anchor is the lithium price and volume level already delivered in the first half of 2026, set against a cost base management guides lower for the full year. FY+2 leans on capacity that lands on stated timelines: Novandino at ~300,000 t of capacity in 2027 and the Mount Holland concentrator doubling, where first product is not expected until the first half of 2030. Dispersion is wide on the out-year — the FY+2 revenue spread across the five runs is 14%, from US$8,800 million to US$10,147 million.

Projections are the median of 5 independent model runs.

Management guided Novandino Litio production to 280,000-290,000 t LCE for 2026, rising to ~300,000 t of capacity in 2027, with the 2027 split stated as 80,000 t of lithium sulfate, 220,000 t from the Antofagasta plant and 20,000-30,000 t of hydroxide. Global lithium demand for 2026 was raised to more than 2.1 million metric tons. For Q3 2026 the guide is flat sequentially: volume in line with Q2, price more or less in line with the first-half average, cost similar to Q2, with full-year 2026 lithium cost reaffirmed as lower than 2025. Iodine production was guided to ~15,500 t with sales volume "similar to what we see last year," and specialty plant nutrition volumes to +~10% versus 2025, with supply-constrained conditions expected to normalize toward the end of the year. Capital expenditure is framed at approximately US$3 billion over 2026-2028. Mount Holland's expansion is guided to start construction mid-2027, commission in 2029 and produce first product in the first half of 2030.

What Could Go Right — and Wrong

Looking Ahead

The next twelve months turn on three things the source states plainly. First, production beyond ~300,000 t, where management has said only that future initiatives remain under study and that more detail will come at a future call, and where the brine-reduction constraint on growth becomes visible. Second, the demand question: whether BESS installations catch up with shipments, or whether the inventory gap management is monitoring widens. Third, project plumbing — Salar Futuro's permit path, the undecided Mount Holland refinery expansion, Kwinana's ramp into 2027, and the ENAP gas contract expiring December 31, 2026 and the Copec fuel contract. On the cost side, management has flagged marginal facilities, the Maria Elena ramp-up and fuel costs as pressures on iodine.

Financials

Annual Summary

Key Ratios (Trailing)

The Company

SQM describes itself in its 20-F as the world's largest producer of potassium nitrate and iodine and one of the world's largest lithium producers. It sells lithium carbonate, lithium hydroxide and spodumene concentrate; iodine and iodine derivatives; specialty plant nutrients; potassium chloride and potassium sulfate; and three industrial chemicals. The mix matters more than the tonnage: iodine and lithium together produced about 86% of FY2025 gross profit from about 73% of revenue, while specialty plant nutrition, potassium and industrial chemicals carry most of the physical volume. Iodine's demand is tied to medical imaging, pharmaceutical and industrial uses rather than computing, so the AI-infrastructure link runs through lithium and, within lithium, through grid-scale battery storage.

SQM operates two resource complexes and a conversion network. In Chile, brines are extracted at the Salar de Atacama, 210 km east of Antofagasta, to produce potassium chloride, lithium sulfate and lithium chloride solutions; the chemical plant near Antofagasta holds 210,000 tons per year of lithium carbonate capacity and 40,000 tons per year of lithium hydroxide capacity. That Chilean operation now runs through Nova Andino Litio, the association with Codelco that closed at the end of 2025. In Australia, SQM holds a 50/50 joint venture with Wesfarmers, Covalent Lithium, covering the Mount Holland open-pit mine and spodumene concentrator on the Earl Grey deposit plus the Kwinana hydroxide refinery, whose ramp has moved into 2027. Conversion in China runs through SQM's own plant, expanding toward ~30,000 tons, plus tolling with a supplier whose contract SQM declines to disclose. Products are sold in more than 100 countries, and 96.5% of 2025 sales came from outside Chile.

Business Segments

Competitive Landscape

The 20-F gives estimated market shares only for lithium: Albemarle at 12%, Jiangxi Ganfeng Lithium at 6% and Tianqi Lithium at 5%. Iodine's competitor set is short and mostly Chilean — S.C.M. Cosayach, ACF Minera, Algorta Norte and Atacama Minerals, the last owned by Chinese company Tewoo. In 100% soluble potassium nitrate the main competitor is Haifa Chemicals of Israel, with Kemapco of Jordan and ACF also named. Potassium competes against Uralkali, Belaruskali, Nutrien and Mosaic, and industrial sodium nitrate against BASF. SQM is a price-taker in these markets: management says its contracts are all based on indices, and the filing extract contains no sole-source disclosures. On the record's own read, SQM is hardest to replace in iodine, where the resource base is scarce and the competitor list short, and least differentiated in specialty plant nutrition and potassium, where the product is substitutable and the FY2025 gross margins were 11% and 2%.

Supply Chain

SQM sits at the resource end of the chain, converting Chilean brine and Australian hard rock into lithium compounds, iodine and fertilizers. Its largest named cost relationship is Corfo, which leases the Salar de Atacama rights that the lithium business depends on. No supply-chain neighbour transcript names SQM directly.