Business

Bottom Line

Daqo is a single-product, low-cost commodity manufacturer at the most cyclical point in the solar value chain — high-purity polysilicon — selling into a market that today has roughly 2.4x more capacity than demand and prices roughly 20% below industry cash cost. The thesis is not earnings, growth, or moat. It is balance-sheet survivorship: $2.0B of near-cash, zero debt, a sub-book valuation, and management willing to accept ~3% sales-volume utilization rather than book unprofitable revenue. The business is worth what an investor believes about (a) the timing and depth of polysilicon supply rationalization and (b) how much of Daqo's listed Chinese subsidiary the market is double-counting away.

How This Business Actually Works

The economic engine is a single chemical conversion: turn cheap silicon ore plus electricity into a refined commodity sold by the kilogram. Margin is the spread between a market-set price and a producer-set cost, and that spread swings violently because supply takes 18–24 months to build but customers can defer purchases in days.

Three structural facts drive everything else:

Electricity is destiny. Roughly a third of unit cost is power, which is why Daqo built in Xinjiang and Inner Mongolia (coal-heavy regional grids, ~half the rate of coastal China). This location advantage, not technology, is the company's only durable cost edge versus Wacker (Germany) or OCI (Korea/Malaysia). Inside China, the same advantage levels off — Tongwei, GCL, Xinte, and Asia Silicon all cluster in the same low-power-cost provinces.

Customer power is severe. The top three buyers — vertically integrated Chinese wafer/module conglomerates — were 63.5% of 2025 revenue, up from 53.8% the year prior. Daqo signs framework volumes; price is whatever spot is when the order hits. There are no take-or-pay floors and no premium for product quality beyond a small N-type spread.

Capacity is sticky on the way down. A polysilicon line is a five-to-ten-year payback asset that loses money the instant you idle it (depreciation runs regardless). That is why China entered 2026 with ~600,000 MT of poly inventory, industry utilization at 39%, and prices below cash cost — yet capacity is not exiting. The clearing mechanism is balance-sheet attrition, not rational price discipline.

The Playing Field

Daqo's "peers" split into two completely different businesses, and the most damaging investor error is comparing them on the same ratios. The relevant comparison set is upstream polysilicon producers — Wacker, Tongwei, GCL, OCI, Xinte. Module/cell makers like JinkoSolar, Canadian Solar, First Solar, and Maxeon are customers, not competitors, and trade on different economics (channel, warranty, project pipeline, IRA tax credits).

What the table reveals:

Daqo is the only liquidity-rich pure play. Wacker mixes silicones, polymers, and biosolutions; the poly division is roughly a quarter of group revenue and most of group volatility. JKS, CSIQ are downstream module makers carrying $4–8B of net debt to fund inventory and project working capital — they cannot afford a long downturn. First Solar is a category of one (US thin-film, IRA-protected, premium customer book) and shouldn't anchor expectations for any Chinese silicon-based name. Maxeon has effectively gone to zero.

Every Chinese-exposed name trades below book. P/B of 0.45–0.60x signals that the public market believes a meaningful share of installed capacity will be impaired or written down before the cycle clears. Daqo is the cheapest on book, but also the only one with the cash to outlast the print.

Quality differentiation is real but small. Daqo and Wacker produce N-type polysilicon (the higher-efficiency grade for TOPCon/HJT cells) and command a small premium versus mass-market product. This explains why Daqo's Q1 2026 ASP of $5.96/kg held above the spot RMB 35–37 range — but it does not change the cycle.

Is This Business Cyclical?

Polysilicon is one of the most violently cyclical commodities in the industrial economy. Prices have moved 80% peak-to-trough twice in 14 years, and the 2022→2025 cycle is the second-worst on record.

The cycle does not hit smoothly — it hits in a specific sequence:

1. Price first, instantly. ASP fell from a 2022 peak of ~$36/kg to $5.25/kg by 2025, an 86% decline. This is what new capacity ramps do to a market with rigid downstream demand.
2. Volume second. Daqo's external sales volume actually held up through 2024 (181k MT) before collapsing to 127k MT in 2025 and a near-zero 4.5k MT in Q1 2026 as management chose to stop selling at below-cost prices.
3. Margin third, with a fixed-cost amplifier. When utilization fell to ~40%, depreciation and fixed conversion costs got absorbed across far less product, pushing gross margin to -521% in Q1 2026 — most of which is a $98M inventory write-down because year-end spot price sat below carrying cost.
4. Working capital fourth. Receivables stretch (the company is now selectively reluctant to ship). Inventory builds. $2.0B of operating cash flow in 2022 became a $147M cash burn in Q1 2026 alone.
5. Cash last. The number to watch. Cash + near-cash dropped from $3.5B at the 2022 peak to $2.0B at end of Q1 2026. At the current burn rate (~$400M/quarter including capex), management has roughly 5 quarters of runway before liquidity becomes a question — which is why the June 2026 government anti-involution policy decision matters more than any operating metric.

The historical analog is 2011–2013, when GCL, REC, and a generation of Chinese poly producers ran the same playbook — capacity additions met a demand pause, prices collapsed 75%, and the survivors emerged when 60%+ of nameplate capacity was rationalized via bankruptcy, idling, or impairment. That cycle took roughly 30 months from peak ASP to clearing. The current cycle is at month 36 from the late-2022 peak with no clear bottom yet.

The Metrics That Actually Matter

Forget P/E, EV/EBITDA, and ROE — they are uninformative for a loss-making cyclical. Five metrics tell you everything about Daqo's position and the probability of survival to the next cycle.

Two of these metrics deserve a direct word. Cash unit cost rose modestly in 2025 (to $6.61/kg from $6.44/kg) entirely because of higher depreciation absorbed across lower production volume — the underlying operating cost in RMB actually fell. So the number that looks like a slip is mostly an accounting artifact of running plants at 41% utilization. Net cash is the metric the equity is priced on: market cap is $2.0B, near-cash is $2.0B, so the stock implicitly says the operating polysilicon business plus the listed subsidiary stake is worth roughly zero. That is the bear case taken at face value.

What Is This Business Worth?

There are two legitimate ways to value Daqo, and they disagree by a wide margin. The right framework is to underwrite balance-sheet floor as the downside and mid-cycle earnings power as the upside, then judge what odds the market is offering.

The first lens is net asset value, with a sum-of-the-parts twist that most analysts miss. Daqo Cayman (the listed ADR) does not own its polysilicon assets directly. It owns 72.8% of Xinjiang Daqo (Shanghai STAR Market: 688303), which in turn owns the Xinjiang and Inner Mongolia plants and most of the cash. So Daqo ADRs are essentially a leveraged claim on the Shanghai-listed sub — and the 27.2% minority interest is a real, non-controllable claim on the same cash and assets.

The mechanical floor: book value attributable to DQ ordinary shareholders is ~$4.4B, market cap is $2.0B → trading at ~45% of book. Even after writing down Phase 5B and the semi-grade plant by half ($1.4B impairment), book is ~$3.0B and the stock is at ~67% of book. This is the downside anchor.

The second lens is mid-cycle earnings power. At 305,000 MT capacity, ~80% utilization (244k MT shipped), an $8/kg ASP, and a $6/kg cost, gross profit is roughly $490M. After SG&A and tax, net income to DQ shareholders (after the 27.2% minority leak) lands in the $200–300M range. At 8–10x mid-cycle earnings, that supports a $1.6–3.0B equity value — overlapping today's market cap. So the stock is not pricing any premium for cyclical recovery.

What this all says: at $29.50, the stock is being valued as if the cycle never recovers and the listed subsidiary stake is worth less than its share of net cash. That is mathematically possible — it requires permanent overcapacity and continued cash burn — but it leaves no room for the upside scenario where the June 2026 policy enforcement does what management expects, ASP returns to $7–8/kg, and Daqo emerges as a dominant low-cost survivor with a cleared field.

What I'd Tell a Young Analyst

Forget the income statement until 2027. A loss-making commodity producer in a trough doesn't have an income statement worth modeling line-by-line. Build a price deck (RMB/kg → USD/kg), apply it to a utilization assumption, run unit economics — that is the entire P&L. Stop trying to forecast SG&A.

Watch four things, in order of importance. (1) China industry inventory weekly print — when it falls below 400,000 MT, the squeeze becomes visible. (2) The June 2026 government cost model and price floor decision. (3) Daqo's monthly production and shipment disclosures — if utilization drops below 40%, depreciation absorption alone wrecks unit cost. (4) Cash balance trajectory; the day this stops being a 2x net-cash story is the day the bull case breaks.

The market may be missing two things. First, the listed subsidiary structure means Daqo ADR holders don't own the assets directly — they own a 72.8% slice of a Shanghai-listed entity. Compare the ADR market cap to 72.8% of Xinjiang Daqo's STAR Market valuation. When the gap widens beyond the historical holding-company discount (~25–30%), there is an arbitrage signal in either direction. Second, management is genuinely refusing to sell below cost, an unusual stance in a Chinese commodity industry where market share preservation usually wins. If they hold this line and the rest of the industry capitulates first, Daqo emerges with disproportionate market share at the recovery — a non-linear outcome the consensus does not model.

What changes the thesis. Bull-case kill: cash falls below $1.0B without ASP recovery, or top-three customers in-source meaningful poly capacity (Tongwei is the obvious threat — they already produce poly internally). Bear-case kill: a Chinese top-5 producer files bankruptcy or announces a multi-billion-RMB impairment, which would reset the supply curve permanently and re-rate every survivor on book.

Don't anchor on First Solar. First Solar is what this industry would look like with IRA tax credits, no Chinese exposure, and CdTe technology insulation. Daqo is what the industry actually is: a brutal commodity sold to four customers in one country, where the only durable competitive advantages are electricity cost and balance-sheet endurance. Both of those, Daqo has — but neither is a moat.