Variant Perception

Where We Disagree With the Market

The Market's Implied View

At $19/share (0.26x book, 0.69x liquid assets), the market is pricing three assumptions simultaneously:

1. The cash is largely inaccessible. A 31% discount to liquid assets implies the market treats at least $600M of the $2.0B as permanently trapped behind the VIE structure.
2. PP&E is worth close to zero. At 0.26x book value, the market assigns minimal value to $3.4B in manufacturing assets — implying they will either be impaired or never generate adequate returns.
3. The cycle trough extends for years. The stock implies no recovery to mid-cycle profitability within a 2-3 year horizon.

Variant Perception #1: The Market Underestimates Government Enforcement Probability

Why the market may be wrong: The April 2026 symposium was not a generic policy statement — it involved MIIT, NDRC, SAMR, NEA, and other agencies, with specific enforcement mechanisms discussed (license revocation, electricity shutdowns). Chinese government has strong tools for industrial policy enforcement in sectors it considers strategic. Solar/clean energy is explicitly a strategic priority. The specificity of mechanisms (not just guidelines) is new relative to prior government statements.

Why the market may be right: Chinese policy announcements frequently fail at provincial implementation. Local governments protect local employers. Enforcement mechanisms may be selectively applied, benefiting connected producers over Daqo.

Test: Are polysilicon spot prices sustained above RMB 40/kg by August 2026? If yes, enforcement is working. If no, the market was right to be skeptical.

Conviction: 6/10. Higher than the market implies, but with genuine uncertainty about implementation.

Variant Perception #2: The Market Overestimates Cash Burn Duration

The gap: The market appears to extrapolate Q1 2026's catastrophic $147.5M operating cash burn as the new normal. But Q1 was uniquely awful: Daqo deliberately withheld 90% of production from sale as a protest against below-cost pricing. This is not a sustainable cash burn rate — either sales resume (reducing burn) or production curtails further (also reducing burn).

FY2025 operating cash flow was positive at $50M despite a $170M net loss, because D&A ($240M) and working capital improvements offset cash losses. Even in a sustained trough, quarterly cash burn is more likely $30-75M than $147.5M.

Why this matters: If cash burn is $50M/quarter (not $150M), Daqo has 10+ years of runway, not 3. The survival question that dominates the bear case becomes much less urgent.

Test: Q2 2026 operating cash flow. If OCF is $(50M) or better, the Q1 burn was anomalous. If OCF exceeds $(100M) again, the market is right.

Conviction: 7/10. The Q1 2026 sales volume was clearly a deliberate choice, not forced selling.

Variant Perception #3: The Market Ignores the Baotou Expansion Optionality

The gap: The market prices Baotou expansion capex as value-destroying (spending cash into an oversupplied market). But the expansion is nearly complete, and the capex cliff means FY2026+ free cash flow improves by $100-200M relative to FY2023-2024, even at flat operations. If the cycle eventually turns, the new capacity earns returns on capital already spent — the option is nearly fully paid for.

Conviction: 5/10. Timing risk is real — more capacity into overcapacity is dangerous. But the capex is sunk, so the option cost is zero going forward.

Summary: Evidence vs. Market Consensus

What Would Prove Us Wrong

All three variant perceptions fail if:

• Government enforcement is announced but not implemented at the provincial level (variant #1)
• Q2 sales volumes remain under 10,000 MT because market conditions truly prevent sales at any price (variant #2)
• Baotou expansion requires additional capex beyond current projections (variant #3)

The meta-risk: this is a Chinese company with VIE structure. Even if all three variant perceptions prove correct, value realization depends on the governance structure permitting returns to flow to ADS holders. That structural question is not a variant perception — it is an unresolvable uncertainty that justifies a permanent discount, just not necessarily a 74% discount.