For most of 2024 and 2025, the working assumption in solar underwriting was clean: modules keep getting cheaper. Developers, lenders, and investors had begun treating price deflation as a silent hedge—a structural tailwind that absorbed execution overruns, compressed LCOE, and made contingency calculations feel more comfortable than they deserved. The module price curve had become a proxy for margin of safety.
That assumption is now the risk.
In late December 2025, four of China's leading wafer manufacturers (LONGi, TCL Zhonghuan, Gokin Solar, and Shuangliang Silicon) jointly announced price increases averaging approximately 12% across standard product lines.[1] This was not an isolated move. It was the visible surface of a synchronised shift across the PV supply chain, moving in tandem from polysilicon through wafers, cells, and into module quotes. On January 9, 2026, China's Ministry of Finance and the State Taxation Administration issued a joint notice: VAT export rebates for photovoltaic products (9% at the time, reduced from 13% in December 2024) would be eliminated entirely effective April 1, 2026. Battery export rebates follow on a phased path to full elimination in 2027.[2][3]
If you invest in solar infrastructure across APAC, the significance here is not that panels have become expensive. They have not. The significance is that pricing has become policy-shaped in a way it was not before, and that changes the risk architecture of every project in your pipeline.
Three Forces Are Moving at Once
The pricing shift of late 2025 into 2026 is best understood as the intersection of three forces, each with its own logic and timeline, not a single event.
The first is policy-driven export economics. The April 1 VAT rebate removal is a blunt administrative instrument, but its mechanics are compounding. In a multi-stage value chain, polysilicon to wafer to cell to module, removing a 9% export tax cushion does not produce a 9% price increase at the module level. Each stage reprices relative to its cost floor and negotiating position. Market intelligence provider TrendForce assessed that the cancellation would directly breach the cost threshold for lower-margin producers, accelerating the exit of capacity that had relied on the rebate to remain price-competitive.[3]
The second is pressure on raw material costs. Polysilicon spot prices in China rose above CNY 60,000 per metric tonne in early January 2026 (a weekly increase of approximately 9–10%) driven by reduced output, higher production costs, and tighter downstream supply.[4] But polysilicon's position as the dominant cost variable is being challenged by silver. According to Rystad Energy, silver prices rose sharply through early 2026, driven by structural demand outpacing supply from mining.[5] The solar industry consumes approximately 30% of global industrial silver output, and the transition to high-efficiency N-type cell architectures (TOPCon and HJT) materially increases silver intensity per watt relative to legacy PERC technology.[6] By some estimates, silver now accounts for 16–17% of total module costs, surpassing polysilicon at certain production configurations.[6] This is structural, not cyclical. The shift toward copper electroplating as a substitute is underway at some Tier-1 manufacturers, but widespread adoption is still a multi-year transition.[5]
The third force is supply discipline. China's solar manufacturers have been in severe financial distress from overcapacity-driven price competition since 2023. By 2024, Chinese factories were producing approximately 630 GW of modules annually, nearly double China's own deployment, forcing module prices down to approximately USD 0.10 per watt.[7] Combined losses among listed Chinese solar companies are estimated to exceed 50 billion yuan for 2025 alone.[8] The state has been signalling for over a year that it wants this to stop. In July 2025, China's Ministry of Industry and Information Technology convened leading solar companies, including LONGi, JA Solar, Trina, and Tongwei, urging an end to "disorderly low-price competition."[7] The VAT rebate removal accelerates the same logic: it is industrial policy as attrition mechanism, designed to accelerate the orderly exit of weaker capacity rather than protect the margins of survivors.
China Dominates, But Does Not Price Monopolistically
A narrative has gained traction in some investor conversations: China has monopolized the solar supply chain, and this is the moment it starts extracting rents. The dominance is real. The extraction thesis is not.
According to IEA-PVPS data, China accounted for approximately 91% of global solar cell manufacturing capacity in 2024, and 97% of global solar wafer production.[9] The International Energy Agency has documented that China's share across all manufacturing stages (polysilicon, ingots, wafers, cells, and modules) exceeds 80%, more than double China's share of global PV demand.[10] That level of concentration has no historical parallel in any major industrial supply chain.
But monopoly pricing power requires more than concentration. It requires the ability to sustain above-cost returns. The multi-year margin destruction that triggered these policy interventions is precisely the evidence that extractive pricing was not occurring. Tongwei, the world's largest polysilicon and solar cell producer, projected a loss of up to 10 billion yuan for 2025, on top of a 7 billion yuan loss the year prior.[8] This is not the behavior of a cartel successfully harvesting rents.
The January 9, 2026 SAMR intervention sharpens the point further. On the same day as the rebate announcement, China's antitrust regulator convened a closed session with the six largest polysilicon producers and ordered full cessation of coordinated capacity and pricing measures, halting the industry's proposed USD 7 billion consolidation fund on antitrust grounds.[11] Polysilicon futures fell 9% the same day. Beijing wants cheap inputs for its own energy transition, cheap panels for domestic deployment, and trade friction reduction for exports. Those objectives are in tension with each other, and the policy set that emerges is the compromise between them, not a strategy of extraction.
The Bankability Question Is Being Asked Wrong
Most commentary on this shift focuses on module price forecasts: up 10%, up 20%, back down by year-end. That framing is the wrong question for infrastructure investors.
For project finance and bankability, the relevant variables are cost certainty, schedule certainty, performance enforceability, and contractual allocation of policy risk. A 5–8% module price movement is manageable within most well-structured EPC contingencies. What changes the risk architecture is a shift in the distribution of outcomes: what was previously a tail risk (price uplift, procurement disruption, policy-triggered cost changes) becomes a live planning scenario requiring explicit treatment in documentation and contract structure.
The pre-deadline export rush created by the April 1 cut-off generated exactly that. Analysts reported severe strain on maritime shipping logistics as manufacturers accelerated outbound shipments ahead of the deadline, compressing carrier capacity and extending customs processing timelines.[2] For projects expecting Q2 2026 or later deliveries, this is a material schedule risk and should be included in the sensitivity analysis, not in the footnotes. The pull-forward in Q1 procurement also left the second half of 2026 facing softer buying interest and sustained pricing uncertainty, with buyers reluctant to commit to forward deals amid volatility.[12]
On counterparty quality: "25-year warranties" are only as bankable as the manufacturer's financial resilience. In a margin-compressed environment where second and third-tier producers are being shaken out, the question of who actually absorbs a warranty claim, and across what jurisdiction, is more consequential than the headline warranty term. Counterparty credit quality in module supply agreements is an underwriting variable, not a checkbox.
What This Means for Investors
The adjustments I would make to underwriting and documentation now are specific rather than directional.
Remove continued deflation from the base case. If your financial model quietly assumes another step-down in module pricing, that assumption is now the risk, not the hedge. The appropriate treatment is a pricing distribution: a flat-to-mild uplift as the base, mild deflation as the upside if oversupply reasserts, and policy-plus-input-cost persistence as the downside. Industry data from OPIS indicate that TOPCon spot module prices had risen by more than 30% since the start of 2026 by late January, with forward curves for H2 2026 settling around USD 0.12 per watt.[12] That remains globally competitive, but it is not the USD 0.07–0.08 per watt that was anchoring assumptions a year ago.
Rebuild EPC contingency logic to map to specific risk drivers. Aggregate contingency percentages obscure where the risk actually lives. Explicitly separate policy-driven cost changes (rebate removal, tariff exposure), commodity-driven inputs (silver, polysilicon), and execution risk (logistics, interconnection, civil works). Each category has different correlation structure and different hedging options.
Tighten "change in law / tariff / tax" contract provisions. The January 9 announcement is a template for how quickly export policy can shift with limited transition notice. Herbert Smith Freehills Kramer noted that suppliers with contracts executed before the announcement may seek change-in-law relief, and entitlement turns entirely on specific contractual terms, including notice provisions and mitigation obligations.[13] For projects sourced through intermediaries, ensure that policy risk does not leak to project cost via uncapped change orders. "Change in law" provisions drafted as boilerplate need to be read as active risk-allocation instruments.
Revisit price validity windows in procurement contracts. In a period of genuine cost-push from input prices and policy change, counterparties who offered long validity windows at fixed prices are under margin pressure. Expect shorter validity periods, embedded premiums for extended price locks, or indexation provisions in new contracts. Procurement strategies that assume historical availability windows remain in place are likely to encounter surprises.
Closing View
Solar remains an attractive infrastructure asset class. The economy has not inverted. What has changed is the shape of the risk and the competencies that will separate good execution from poor execution in this next phase.
The era of "free deflation", where declining module prices quietly absorbed contingency shortfalls and helped investors reach return targets without procurement discipline, is over. Pricing is now policy-shaped at both ends: China managing its industry structure domestically through a combination of anti-involution intervention and antitrust oversight, and importing markets hardening their procurement rules through local content requirements, trade defence measures, and supply chain diligence obligations. The EU's Net-Zero Industry Act, which came into force in June 2024, builds resilience criteria into public procurement and targets 40% domestic self-sufficiency in solar manufacturing by 2030.[14] Similar dynamics are visible across ASEAN procurement politics, India's PLI-driven sourcing preferences, and US tariff enforcement. The directional pressure is consistent across geographies.
In that environment, the investors who perform are those with cleaner risk-allocation frameworks, tighter counterparty due diligence, and procurement strategies with built-in redundancy. The cost curve still works in your favour. The volatility regime around it has changed.
Sources
Primary sources and market references cited above.