Market cap sits $655M below the cash on the balance sheet

Enterprise value is negative $655M. The market pays $0.66 per dollar of net cash and assigns the entire 305k MT production platform a value below zero. The $100M buyback authorized August 2025 sat ~1% executed eight months later while the stock traded at sub-0.25x book — the people closest to the arb are not acting like it exists.

Revenue collapsed 86% from peak; the cycle swung 188 points of gross margin

• Revenue freefall. From $4.6B (FY22) to $665M (FY25) to $26.7M (Q1 2026). Polysilicon ASP fell from $30+/kg to $5.25/kg, and Q1 2026 sales volume collapsed to 4,482 MT as management refused below-cost transactions.
• Cash burn masked by accounting. FY25 operating cash flow of $50M nearly equals the $56M SBC add-back — underlying operating cash before equity dilution was negative. Free cash flow was -$917M cumulative over FY24-25. Notes payable of $99M exceeded total CFO.
• Balance sheet still standing. $1.94B cash, zero debt, 5.4x current ratio, book value held above $63/share through two loss years. At Q1 2026’s burn rate (~$270M/quarter), the runway is roughly seven quarters before cash halves.

Accounting is clean enough to underwrite but not clean enough to trust at face value

Big-bath asymmetry. Management concentrated $302M of charges in FY24 (PP&E impairment, inventory write-down, first-ever credit-loss allowance), then booked zero long-lived asset impairment in FY25 despite ASP falling another 7.3%. The recoverability test relied on a Q4 ASP rebound that Q1 2026 just broke.

CFO quality. FY25 notes payable added $99M to operating cash flow — more than total CFO of $50M. Receivables jumped 146% while revenue fell 35%, with DSO rising from 30 to 52 days and a $19M credit-loss reserve tied to a local government entity Daqo had been bankrolling.

Forensic score: 38-54/100. Big 4 auditor (Deloitte, PCAOB-inspectable), no restatement, no factoring, no acquisition tricks. The risk is judgment-driven: impairment timing, notes-payable-dependent CFO, and SBC that held high as revenue collapsed to 8.4% of sales.

Grade C+ to B-: real skin in the game, compromised oversight, dynastic succession

• Family controls 36%. Founder Guangfu Xu (83) holds 18.4%; CEO Xiang Xu holds 11.4% through BVI vehicles. That is real money on the line — but the CEO chairs his own Nominating & Governance Committee, and the Compensation Committee chair is a Daqo Group VP-Finance the company itself calls non-independent.
• Dynastic pipeline. CEO’s 30-year-old daughter went from IR head to Deputy CEO in 17 months, with zero disclosed share ownership. The CFO also holds zero shares after a decade in the role. Four Form 144 sale notices were filed Sept-Dec 2025 with no visible offsetting buys.
• Buyback as revealed preference. $100M authorized, ~$1M executed at sub-cash levels. IR commentary: “the stock had run from $23 to $31, we wanted to purchase more shares, we were waiting.” They waited while it fell to $19. The textbook capital-allocation mistake.

The unexecuted buyback is the family’s verdict on the cash-arb math

Three disagreements with consensus: (1) the $1.94B cash is a deferred test, not a fortress — operating cash before SBC is negative and the Cayman/Xinjiang structure creates capital-trap risk. (2) A fresh PP&E impairment in H1 2026 is more likely than the market prices — the 2025 zero-impairment relied on a Q4 rebound Q1 just broke. (3) The ~1% buyback execution is the family’s revealed preference, not a timing call.

Q2 2026 earnings in late August is the binary event for both sides

• Sales volume recovery. Q1 2026 shipped 4,482 MT against 43,402 MT produced. If Q2 sales rebound toward 30-40k MT with ASP above $5.96/kg, the destocking narrative survives and the floor held. A second sub-$100M revenue quarter triggers the bear’s primary case.
• SAMR policy clarity. On January 6, 2026, SAMR halted antitrust review of the $7B consolidation SPV and barred industry coordination on volumes, pricing, or capacity. The bull’s primary catalyst — forced retirement of marginal capacity — depends on a re-architected mechanism surviving antitrust.
• Buyback execution watch. A single $10M monthly print would be the cheapest disconfirming evidence the family could produce against the capital-trap thesis. Every month of ~$1M execution at sub-cash is the variant view repricing itself.