Daqo's Pivot: From Polysilicon Glut to AI Data Center Power Infrastructure

Amid a brutal polysilicon downturn, Daqo bets on anti-involution enforcement and a bold expansion into solid-state power equipment for AI data centers.

DQ · Earnings Call · 2026-08-20

The Polysilicon Crossroads

Daqo New Energy's second-quarter 2026 results read like a war diary of a company under siege. Revenue recovered to $62.7 million from the disastrous $26.7 million in Q1, but that was still down from $75 million a year ago. The gross margin, while improving to negative 132% from negative 520%, remains deeply in the red. “Despite these headwinds, we resumed sales in June, delivering a sequential increase in revenue” — Anita Zhu, Translator · 2026-08-20, yet the average selling price crumbled to just $4.04 per kilogram. The company's balance sheet remains a fortress—$1.9 billion in readily convertible assets and zero debt—which is precisely why it can afford to sit out the worst of the price war and plot its next move. The core problem is an industry drowning in inventory and below-cost pricing. N-type polysilicon prices fell from RMB 35–37 per kilogram in late Q1 to RMB 31–34 by mid-year. “Polysilicon market prices came under further downward pressure during the second quarter with N-type polysilicon prices falling from RMB 35 to RMB 37 per kilogram at the end of the first quarter to RMB 31 to RMB 34 per kilogram at the end of the second quarter.” — Anita Zhu, Translator · 2026-08-20 In response, Daqo initially held back sales, adhering to voluntary self-regulation, but eventually capitulated to a sell below cost reality in June to rebuild volume. Sales volumes tripled to 15,190 metric tons, but only at the price of unprofitable transactions.

Anti-Involution: A New Enforcement Regime

The hope now rests on a more muscular version of anti-involution policy. In recent months, Beijing has moved from rhetoric to rule. A mandatory energy consumption standard—6.3 kgce/kg by January 1, 2027—will force inefficient plants to close. The China Photovoltaic Industry Association published a cost accounting model pegging industry average production cost near RMB 50 per kilogram. The State Administration for Market Regulation issued price compliance guidelines, explicitly threatening enforcement against below-cost dumping. And on August 6, Daqo joined seven other polysilicon manufacturers in signing a pledge to eliminate below-cost sales. Management now sees a structural shift. “there is strong consensus within the industry for self-discipline and also with the urging of the government and the related departments that the Industry consensus is that it's no longer viable to sell below cost.” — Ming Yang, Chief Financial Officer · 2026-08-20 They expect prices to recover gradually as weaker players are forced out—effective capacity, they argue, is already below 2 million tons against nameplate of nearly 3 million. This is not the first time Daqo has promised government-led relief. On the prior quarter's call, CFO Ming Yang acknowledged the rollout had been slower than expected: “our understanding is that the government, especially at the most recent industry meeting with the Ministry of Industry Information Technology and NDRC and NEA and the Market Regulation Agency – so basically, there is a consensus from the government that at the minimum, while maintaining some market competition, there's a need to enforce the price law.” — Ming Yang, Chief Financial Officer · 2026-04-29 The difference today is the concrete legal and administrative scaffolding. Last February, the company guided to a floor of RMB 53–54 per kilogram based on the Pricing Law: “As part of the Pricing Law, sales should not be below the industry-level cost. I would say the lower bound will be at least RMB 53-54 per kilogram” — Anita Xu, Senior Management (speaking on behalf of Chairman) · 2026-02-26. That floor collapsed in Q2, but the new standards may finally force the industry to obey.

A Bold Pivot into AIDC Power

The most significant change on the call, however, is not about polysilicon at all. Daqo unveiled a diversification plan into AI data center power infrastructure. The company signed an investment agreement to build a manufacturing base for next-generation energy solutions, including solid-state transformers, solid-state circuit breakers, and energy storage systems designed for 800V DC architectures championed by NVIDIA. The venture is anchored by affiliate Daqo Group, which brings four decades of power equipment expertise. Management frames this as a second growth engine. “We do see that the AIDC related power infrastructure and equipment market is actually a very viable sector where it's going to be a significant growth driver for the company, and it's the second sector that the company is entering into.” — Ming Yang, Chief Financial Officer · 2026-08-20 The initial phase is RMB 2 billion out of a total RMB 6 billion plan, with only $30–40 million of capex this year. A prototype is expected by year-end, with sales beginning in 2027 and meaningful growth in 2028–2030.

So in terms of products as well. So with the growing power demand and especially for the next-generation power infrastructure for AIDC, where – led by NVIDIA, the future development of a new – next generation of equipment under the 800-volt DC infrastructure for – so we're targeting initially in the solid-state transformer and solid-state circuit breaker market... and then we expect to see very significant growth from 2028 to 2030
Ming Yang, Chief Financial Officer · 2026-08-20

This is a company-unique pivot that goes beyond the usual data center talk. While the broader market is fixated on AI data center power demand—global keyword lists are saturated with phrases like "critical IT load" and "high power density"—Daqo's move into actual power equipment manufacturing is distinctly its own. It leverages its parent's manufacturing know-how and its own war chest, essentially trading a cyclical commodity for a structural growth theme. The company also quietly noted progress in semiconductor-grade polysilicon, a higher-margin niche, with an additional RMB 1.2 billion invested and qualification ongoing. What remains to be seen is execution. Daqo has no track record in power electronics, and the timeline is ambitious. But the strategic rationale is clear: with solar polysilicon mired in oversupply, this is a hedge against an otherwise bleak near-term outlook. The market will likely reward the optionality—if not the current financials. The stock has no price tape in our dataset, but the narrative alone marks a decisive inflection point in the company's history.