Bernreuter Research projects the Chinese polysilicon industry could swing from chronic overcapacity into a structural shortage by 2028, with polysilicon spot prices having already collapsed from a 2022 peak of US$39/kg toward historic lows as of mid-2026 [S6][S9].
The 2026 floor is policy-driven, not demand-driven: China's GB 29447-2026 energy-consumption standard (effective 2027) and U.S. Section 232 / AD/CVD investigations into solar-grade material are simultaneously squeezing supply-side economics on both sides of the Pacific, with Anza's procurement outlook flagging the Section 232 review of polysilicon and processed critical minerals as a top-2026 watchpoint [S3][S4].
Where the 2026 price floor actually sits
Spot polysilicon has fallen so far that Indexbox reports producers are halting lines and routing inventory through vertically integrated module channels rather than selling into the spot market, with major participants warning that without policy support before May 2026 they may resort to discounted clearance of high module inventories [S5]. Bernreuter's published series anchors the prior peak at US$39/kg in 2022 against 2026 levels near historic lows, the kind of compression that historically clears marginal capacity first [S6]. For buyers still using 2022 price assumptions in long-lead EPC budgets, the comparison is no longer valid; module and BOM contracts being signed in Q2-Q3 2026 are clearing at a different cost basis, and the industrial UPS and switching power supply infrastructure behind ingot/wafer fabs is being sized to that lower-utilization reality.
Why GB 29447-2026 is a 2028 problem, not a 2026 one
GB 29447-2026 caps polysilicon energy intensity and could retire roughly 1.5 million tonnes of PV-grade nameplate capacity on paper, but the practical impact through 2026 is muted because most sub-6.3 kgce/kg lines were already idled on economics alone, leaving China's H1 2026 operating capacity at approximately 1.3 million tonnes [S4]. Bernreuter's shortage call is a forward-curve read, not an in-year shortage: the swing factor is whether China's top-tier producers (the ones below the GB threshold) are willing to keep capex committed when spot is below cash-cost plus a normal return [S6][S9]. Capacity discipline decisions made in late 2026 and 2027 will determine whether 2028 prints a deficit; the rule itself is not a 2026 supply event.
Tariff overlay: Section 232 and AD/CVD stack on the feedstock
Anza's January 2026 procurement outlook tracks three concurrent U.S. actions that hit polysilicon-fed solar supply chains: the Solar IV AD/CVD case on India/Indonesia/Laos cells and modules (preliminary CVD due 2026-02-23, preliminary AD due 2026-03-27), the active-anode AD/CVD case for storage (final ITC determination expected January 2026, with an estimated block-level impact of US$8-15/kWh), and a Section 232 review of critical minerals where a presidential proclamation on 2026-01-14 declined to impose tariffs on critical minerals and derivative products at that stage [S3]. The combination matters because polysilicon is the upstream feed for every c-Si cell under all three proceedings; even when Section 232 itself takes no immediate action, AD/CVD rates exceeding 200% on India-origin cells can flip module sourcing economics overnight, and procurement teams have started pre-positioning supply chains for that contingency [S3].
What changes between now and 2028: a side-by-side of the two regimes
2026 is an oversupplied buyer's market with sub-cash-cost clearing prices, idled sub-threshold capacity, and module inventory being absorbed through vertical channels rather than spot sales [S5]. 2028, per Bernreuter, is a seller's market built on three conditions stacking: GB 29447-2026 retiring the high-energy tail, top-tier Chinese producers running near nameplate, and demand recovering from the 2026 trough [S6][S9]. The transition window is roughly 18-30 months, and that is the same window in which Section 232 remedies, AD/CVD final determinations, and any new critical-minerals action all have to land [S3]. For EPC and module-procurement teams, the hedging problem is asymmetric: a buyer who locks 2026 spot avoids a 2028 spike but accepts tariff risk if Section 232 is imposed mid-contract; a buyer who waits for clearer policy signal carries the deficit risk if 2027 capacity discipline arrives first.
Cross-industrial signal: analog feedstock cycles in 2026
The polysilicon cycle is not unique in 2026. Other industrial-feedstock markets are running the same oversupply-to-tightening pattern, and the procurement playbook is similar: lock forward volume when spot is below marginal cost, hedge freight, and qualify a second geography before the swing. The PVC chain is going through a parallel capacity-crunch, export-pivot, and mercury-cell squeeze in 2026 that is reshaping Chinese export flow and downstream pipe and fitting supply PVC Resin 2026: Capacity Crunch, Export Pivot, and the Mercury-Free Squeeze. For solar EPCs and flow meter skids destined for polysilicon and module plants, the practical lesson is that Chinese nameplate and export licensing are now the dominant price-setters, and equipment orders placed in mid-2026 should be built around a 2027-2028 ramp assumption, not a 2024 baseline. [S3]
Trackable signals to watch through Q4 2026
Three concrete nodes will tell you whether Bernreuter's 2028 call is firming: (1) Q3 2026 Solar IV preliminary AD/CVD rate announcements from the U.S. Department of Commerce, since Indian AD rates are signalled to exceed 200% under adverse-facts-available and that reshuffles c-Si module sourcing overnight [S3]; (2) any Section 232 polysilicon-specific finding from BIS, because the critical-minerals proclamation of 2026-01-14 explicitly left the polysilicon track open [S3]; and (3) Chinese top-tier producer utilization statements in Q4 2026, because if the leaders are running below 70% the 2028 deficit probability rises sharply [S6][S9].