Business

Know the Business

Daqo is a pure-play, low-cost polysilicon refinery whose earnings power is set almost entirely by a single commodity price it cannot control. Two-thirds of its market cap is parked as cash on the balance sheet, which is the right way to read this stock: a debt-free option on the next polysilicon up-cycle, not a normal industrial. The market is probably under-pricing the optionality from China's "anti-involution" capacity discipline and over-pricing the depreciation drag from idle Phase 5 capacity.

1. How This Business Actually Works

Daqo turns electricity and metallurgical-grade silicon into a single product — high-purity polysilicon — and sells every kilogram into the Chinese solar wafer supply chain at the prevailing spot price. Profit is whatever ASP minus all-in cost happens to be that quarter, multiplied by however many tons the market will absorb. There is no recurring revenue, no installed base, no take-or-pay; every quarter resets.

Two inputs drive the cost structure: cheap thermal-coal electricity and metallurgical silicon. Daqo's edge is geographic — its plants sit in Xinjiang and Inner Mongolia where regional coal grids deliver some of the world's lowest industrial power rates. That cost advantage is the entire moat. It produced an all-in cash cost of $4.46/kg in Q4 2025, a company record, while peers in coastal China and Germany run materially higher.

The mechanic that matters is operating leverage on a near-fixed cost base. Once a polysilicon line runs, depreciation, baseline electricity contracts, and skeleton labor are sunk; every additional kilogram drops to gross profit. That is why FY2022 produced a 73.9% gross margin and FY2025 a -20.7% gross margin on revenue that fell only 86% — operating leverage cuts both ways, and idle Phase 5 capacity now amplifies losses through depreciation rather than direct cost. Daqo paused its newest 100,000 MT Phase 5B and the 1,000 MT semiconductor-grade lines in 2025 specifically to stop bleeding through them.

Bargaining power is one-sided. Polysilicon is a fungible chemical input, and Daqo's customers are large vertically-integrated Chinese wafer/module producers who can buy from any of six major Chinese suppliers. Top three customers were 63.5% of FY2025 revenue. The pricing convention — framework volumes, spot pricing at order — means Daqo absorbs price risk on every batch.

2. The Playing Field

Daqo competes in two adjacent industries that the market lumps together but should not. Pure polysilicon has perhaps six globally relevant producers (Daqo, Tongwei, GCL, Xinte, Wacker's polysilicon segment, OCI). The downstream wafer/module makers (Jinko, Canadian Solar, JA Solar, LONGi) are a different business — vertically integrated, brand-and-distribution-led, with their own polysilicon supply choices. First Solar is the outlier: a US thin-film maker insulated from this whole silicon cycle by a different chemistry (CdTe) and by the Inflation Reduction Act.

The peer table reveals the only bull case the silicon-based solar chain has right now: First Solar. FSLR earns a 17% ROE at a 2.9x P/B in the same calendar year that DQ, JKS and CSIQ are sub-book and unprofitable. That is what a moat actually looks like — protected market, differentiated technology, locked-in pricing — and it should be the benchmark you mark Daqo against, not the other Chinese names. Within the silicon group, the right comparable is Wacker Chemie, where polysilicon is one of four divisions cushioned by silicones and biosolutions. WCH stayed profitable through a cycle that wiped out the pure-plays' earnings; that is the cost of running a one-product commodity business without diversification.

The second message: Daqo is the only Chinese silicon-chain operator that is debt-free with a substantial net cash cushion. JKS carries roughly $2.2B of net debt and CSIQ over $6B; both are structurally fragile in a prolonged trough. Daqo can lose money for years and still be standing. That is the survivability premium the P/B 0.45 should partly reflect, and increasingly does not.

3. Is This Business Cyclical?

Polysilicon is the single most violently cyclical large-cap commodity in the listed solar supply chain. Daqo's gross margin has swung from +80% in Q3 2022 to -108% in Q2 2025 — a 188-point range over eleven quarters. Revenue per quarter ran from $1.28B to $75M over the same window. Both extremes were driven almost entirely by a single variable: polysilicon ASP, which moved from peaks above $30/kg in 2022 to roughly $5/kg through 2024 and the first half of 2025.

Where does the cycle hit? Three places, in order: ASP first, then utilization, then working-capital impairment. ASP collapses when industry capacity laps demand — China's nameplate polysilicon capacity is now over 3 million MT against roughly 1.4 million MT of demand. Utilization follows: Daqo ran 33% capacity utilization in Q1 2025 and pulled it to 55% by Q4 by sequentially idling older Xinjiang lines. Inventory and PP&E impairments come last; Daqo took a $176M long-lived-asset impairment in 2024 and an $18M-19M credit-loss reserve each of the last two years for receivables from a local-government infrastructure entity that cannot pay because the regional tax base collapsed with the industry.

This is not the first cycle. The 2011-2013 polysilicon glut bankrupted multiple Western producers and forced LDK Solar into receivership; ASP fell from $80/kg to under $20/kg. Daqo survived that round by relocating from Chongqing to Xinjiang for cheaper power, and emerged as a low-cost winner. The current trough is structurally similar but quantitatively worse on the supply side, and the policy response is more aggressive.

4. The Metrics That Actually Matter

Standard ratios are mostly noise here. P/E is meaningless when earnings are negative; ROE is meaningless when the business is mid-cycle; book value is suspect when 53% of total assets are PP&E that may face further impairment. The five operating metrics below explain almost everything:

The chart above is the entire investment thesis on one axis. When the orange ASP line is above the blue cost line, Daqo prints money disproportionately to revenue. When ASP is below cost — which it has been for two consecutive years — every kilogram sold loses money. The bull case is not that demand explodes; it is that the orange line crosses back above the blue line and stays there. That requires marginal Chinese capacity to exit, not just idle.

5. What I'd Tell a Young Analyst

Stop modeling this like a normal industrial. There is no through-cycle ROE to anchor on; there is only the cost curve, the policy regime, and the balance sheet. The right framework is: estimate Daqo's net cash and at-cost optionality value, and treat the production business as a free call on the next polysilicon upturn.

Three things to actually watch through 2026:

1. The credibility of the RMB 53-54/kg policy floor. If sales-below-cost enforcement holds and the consolidation SPV (formed December 2025) starts forcing capacity exits — not just idling — the cycle will turn faster than consensus thinks. If enforcement slips, the floor breaks and the trough extends.

2. Daqo's Q4 cash cost gap to the industry. A $4.46/kg cash cost against an industry marginal cost near $7/kg means Daqo gets paid first in any recovery. Watch whether peers (especially Tongwei and Xinte) close that gap, because that is where the moat erodes.

3. What management does with $2.27B of cash. The buyback hesitation in the Q4 2025 call ("waiting for policy clarity") is rational but not free — every quarter of inaction at sub-book is shareholder value left on the table. The ideal use is opportunistic M&A of distressed Chinese capacity at scrap value during the consolidation phase. The worst use is another greenfield expansion into oversupply.

What the market is probably underestimating: the asymmetry. Net cash plus liquid investments is roughly $2.0B; market cap is roughly $2.0B; the entire production business is being valued near zero. If polysilicon ASP recovers to even $9-10/kg (well below the 2021 average) on stable 200k MT volume, Daqo could earn $400-600M annually. That is not a forecast — it is a sensitivity that explains the option.

What the market may be overestimating: the speed and durability of the recovery. Anti-involution is real but slow; capacity exits take years, and Daqo's own Phase 5B and semi-grade lines are still on the book waiting to be turned back on. Be patient and skeptical in equal measure. The thesis breaks if (a) marginal capacity does not exit, (b) cash burn resumes from policy reversal, or (c) management deploys cash into new capacity instead of share repurchases or distressed M&A.