1. Core judgment
As of 2026, solar manufacturing profits are shaped by capacity cycles and regional policy; installation growth does not guarantee manufacturer profits. [E01][E04][E06]
2. Industry boundary and revision
The scope covers polysilicon, wafers, cells, modules, inverters and trackers, engineering and construction, developers, operating power plants and grid connections. Thin-film technology provides a comparison. Storage is included only where it changes solar integration or revenues, avoiding duplication of Day02's battery-storage research. Solar thermal power, standalone storage manufacturing and non-energy semiconductors are excluded. The endpoint is electricity actually sold and its price, not just shipped modules.
Boundary revision: installed PV, module nameplate capacity and usable electricity belong to different stages: nameplate capacity is not actual qualifying output, and installed power is not electricity generated or sold. [E02][E07] Grid connection, curtailment, price and financing must enter the profit analysis. US thin-film and Chinese integrated crystalline-silicon manufacturers also differ in technology, markets, tax credits and accounting; a single cost explanation is unsafe. [E04][E05][E06]
Research cutoff: 4 November 2026. GW measures power, not electricity generated; annual nameplate capacity in GW/year is not output. Company figures are consolidated, not necessarily pure-module segments or a global profit pool. Forecasts, company guidance and policies not yet effective are separately labeled.
3. Verdicts on four falsifiable hypotheses
A. Global deployment growth means module manufacturers' profits rise
Verdict: weakened. IEA PVPS estimated in September 2026 that global PV additions were about 690 GW in 2025, 15% above 2024, while module production exceeded installations but remained below manufacturing capacity. [E02] Jinko Holdings recorded FY 2025 operating loss of RMB 8.91bn, a -13.6% margin; Q3 2026 operating loss was RMB 0.08bn, or -0.6%, improved from -11.6% in Q2. [E05][E06] These observations do not establish one cause, but do challenge an automatic link between deployment and profit.
Counterevidence and conditions: First Solar retained a high FY 2025 gross margin; its technology, US sales and accounting for the Section 45X credit differ from an exporter of crystalline-silicon products. [E04][E08] Exits, differentiated products or protected-market premiums can support individual profits. Confidence: medium. Falsification: consistent multi-company, cross-cycle data show profits reliably track deployment after capacity and geography are controlled.
B. Global overcapacity means every stage and region has excess supply
Verdict: varies by stage and geography. IEA estimates China accounted for around 85% of solar supply-chain capacity and 95% of wafer capacity in 2024, not actual production or available supply in each country. [E01] IEEFA estimates India had about 233 GW/year of module nameplate capacity in June 2026 but relatively underdeveloped domestic cell, wafer and polysilicon stages; module surplus can coexist with limited compliant cell supply. [E09] US trade and domestic requirements may limit access to globally cheap products. [E10][E11] Confidence: medium. Falsification: common measures of effective, compliant capacity, demand and prices show a sustained surplus at all key stages.
C. The lowest-cost module maker necessarily captures the largest profit
Verdict: weakened. Profit depends on realized price, manufacturing cost, tax credits, sales mix, duties, depreciation and distribution. First Solar's FY 2025 gross margin was 40.6%; its October 2026 annual guidance assumes $2.07–2.19bn in Section 45X credits. This is a guidance assumption, not Q3 actual credits or pure profit. [E04][E08] Jinko's Q3 2026 consolidated gross margin was 12.3% and operating margin approximately -0.6%; differences in technology, business mix and geography prevent conclusions about unsubsidized unit cost. [E06] Confidence: medium. Falsification: same-technology, same-market comparisons adjusted for policy and capital costs show the lowest-cost producer consistently earning the largest absolute profit across cycles.
D. Cheaper modules always increase returns for the entire power system
Verdict: depends on project and time. Lower module prices generally reduce initial equipment spending, but financing, grid delays, curtailment, daytime power prices, land and system costs jointly determine returns. The IEA describes grid queues and congestion, while negative prices can indicate insufficient system flexibility. [E07][E12] These are regional conditions, not evidence that every market has identical economics. Confidence: medium. Falsification: after controlling for financing, connection, capacity factor and realized power prices, project returns rise proportionately with module-price declines across markets.
4. Asset one: profit map
The table compares observable company results, not global profit pools. Different currencies, technologies and business scopes are not summed. FY is fiscal year; Q3 is a quarter.
| Stage / sample | FY 2025 | Q3 2026 | Mechanism and limitation |
|---|---|---|---|
| Polysilicon, wafers and cells | No comparable pure-segment operating-profit sample | No comparable sample established | Capacity additions, utilization, technology shifts and power costs matter; consolidated group profit is not segment profit [E01][E13] |
| Jinko integrated group | Revenue RMB 65.50bn; operating loss RMB 8.91bn; margin -13.6% | Revenue RMB 13.620bn; operating loss RMB 0.080bn; margin -0.6% | Integrated business includes more than modules [E05][E06] |
| First Solar thin-film modules | Revenue $5.219bn; gross profit $2.120bn; gross margin 40.6%; operating income about $1.597bn, or 30.6% | Revenue $1.231bn; gross profit $0.713bn, or about 57.9%; operating income about $0.546bn, or 44.4% | Section 45X reduces cost of sales; comparisons need adjustment [E04][E08] |
| Enphase microinverters and storage group | Revenue $1.472985bn; operating income $0.157526bn, about 10.7% | Revenue $0.293329bn; operating income $0.006466bn, about 2.2% | Consolidated results include batteries and other business; FY 2025 benefits from manufacturing credits [E18][E19] |
| Trackers and EPC | No consistent verified financial sample | No consistent sample | Product mix, services and location change economics; ranking remains undetermined |
| Project development and operation | No consistent segment sample | No consistent sample | Contract price, financing, land, connection and curtailment govern returns over many years [E07][E12] |
First Solar operating income is calculated from disclosed gross profit less operating expenses: FY 2025 $2.120339bn minus $0.523475bn; Q3 2026 $0.713060bn minus $0.166952bn. Tax credits reduce reported cost of sales, so the margin cannot be attributed wholly to technology. [E04][E08] Jinko's consolidated business includes storage and other activities; operating loss differs from net loss attributable to shareholders. [E06]
Global profit-pool shares: undecided. Private companies, intercompany trade, cross-ownership and incomplete regional data prevent reliable aggregation. Gross margin, operating margin, absolute profit and return on invested capital are distinct metrics. A high margin does not establish the largest profit pool.
5. Asset two: bottleneck urgency
The following qualitative ranking assesses conversion of capacity into deliverable products and saleable electricity in 2026. It is not a standardized shortage rate; regional judgments are inferences with incomplete comparable delay data.
| Rank | Bottleneck | Transmission and evidence | Regional differences and disconfirming signal |
|---|---|---|---|
| 1 | Grid connection, transmission and flexibility | Congestion can cause waiting, curtailment or low prices; IEA cites 5–15 years for grid planning and completion [E07] | Europe, US and Chinese provinces require separate queue and dispatch data; falling waits and curtailment would lower the rank |
| 2 | Compliant and financeable regional supply | Origin and import policy can prevent global low-price capacity from entering a market frictionlessly; announcement is not effectiveness [E10] | US policy and traceability are especially relevant; other markets differ |
| 3 | Financing, permitting and land | Higher financing costs and delays can prevent low-cost modules from becoming orders; no globally standardized evidence was assembled | Currency and credit conditions matter especially in developing markets; no quantified ranking claimed |
| 4 | Local upstream capacity | India's module expansion is uneven relative to upstream inputs [E09] | China has a fuller chain; new qualified Indian cell capacity can ease constraints |
| 5 | Nominal module capacity | A global surplus does not mean every qualified product is available everywhere [E01][E02] | Specific technologies and certifications can create localized tightness; global shortage is not established |
The IEA's 5–15 years covers planning through completion of grid infrastructure, not a universal project timeline. [E07]
6. Asset three: timing of profit migration
During the expansion around 2023, Jinko's operating margin was 5.1%; it became -3.6% in 2024 and -13.6% in FY 2025. [E05] This is one integrated producer's reversal, not proof every stage simultaneously lost money. IEA documented pressure on polysilicon prices and utilization in 2024; module output remained below available capacity in 2025. [E02][E13]
In FY 2025 First Solar's gross margin was 40.6% while Jinko recorded an operating loss. Those metrics differ and cannot be subtracted or called a direct transfer of profit. US Section 45X credits and different technologies are material conditions. [E04][E08] In Q3 2026 Jinko's operating loss margin narrowed to 0.6%, while First Solar's gross margin was about 57.9%; the difference remains multi-causal. [E06][E08]
As of November 2026, cheaper modules can reduce equipment costs and potentially shift value toward developers, buyers or system services. Grid, price and financing constraints can offset this. Insufficient cross-chain financial data prevent a conclusion that global profits migrated linearly from manufacturing to power plants. [E07][E12]
7. Asset four: counterfactual stress tests
These are scenarios, not forecasts or realized outcomes.
Scenario one: module capacity exits faster than demand growth
Shutdowns, restructuring and reduced capex → less effective supply → if inventories, imports and demand stay unchanged, producers may gain bargaining power → margins and cash flows may improve. Weak or leveraged producers may exit before any recovery. Inventory release, cheap new lines, redirected trade or weaker demand can offset the effect. Track utilization, inventory, cash cost and transaction prices rather than announced nameplate closures alone. [E01][E13]
Scenario two: grids and flexibility improve faster than expected
Transmission, storage and flexible loads → greater connection and usable generation → possible reductions in curtailment and low-price hours → potentially stronger project revenues and financeability → demand for modules and system services. Grid investments have costs; location, congestion and load patterns determine beneficiaries, and prices do not automatically improve. [E07][E12]
Scenario three: regional trade barriers grow and 45X incentives change
Higher import costs or minimum prices → higher relative prices for compliant domestic production and possibly higher project module costs → manufacturers may benefit while developer economics weaken. If credits fall, eligibility changes or ramp-up is delayed, domestic producers may also suffer. Effects depend on contracts and actual legal scope and timing. [E08][E10] At the cutoff, the US proclamation specifies 4 December 2026 for the relevant measures; it would be wrong to portray them as fully effective in November. [E10]
8. Asset five: evidence gaps and unresolved questions
- Comparable worldwide profits and invested capital are missing for polysilicon, wafers, cells, inverters, EPC and power generation, preventing a global profit-pool or ROIC estimate.
- A consistent measure of shipments, effective capacity, inventories and regionally compliant supply is absent; India's module nameplate estimate lacks an independent corroboration. [E09]
- First Solar's credit treatment is established, but annual guidance must not be subtracted from quarterly earnings; its cost, contracts and market mix differ from Jinko's. [E06][E08]
- No comparable global project-level sample of queue times, curtailment, negative prices and financing cost was assembled; the bottleneck ranking is conditional rather than measured. [E07][E12]
- Two IEA PVPS releases report different estimates of 2025 global additions. The later release is not automatically a formally explained correction.
Conflict record [conflict E02-vs-E03]: global PV additions in 2025. IEA PVPS's September 2026 Trends report says about 690 GW [E02]; its 2026 Snapshot says an estimated 698 GW [E03]. The publication dates and estimates differ; no explicit explanation of revision or boundary was found. Verdict: use the later report's approximately 690 GW as a time-anchored estimate while retaining the earlier figure; the reason remains unresolved.
9. Evidence ledger
The table identifies key evidence and independent verification. A second source must independently check the same measure and scope; a company's press release repeating its own filing is not independent verification. Single-source confidence is capped at medium.
| ID | Data point / judgment | Value or conclusion | Time anchor | Source type | Source and link | Independent second source? | Confidence | Conflict flag |
|---|---|---|---|---|---|---|---|---|
| E01 | China capacity concentration | Around 85% of solar supply-chain capacity, 95% of wafers; capacity, not output | 2024 basis, 2026 report | International organization | IEA ETP 2026 | No; S&P supports concentration but differs in year and scope [E14] | Medium | No |
| E02 | Global additions | About 690 GW, up 15%; module production below capacity | 2025; published September 2026 | International organization | IEA PVPS Trends | No | Medium | Yes, E03 |
| E03 | Alternative additions estimate | Approximately 698 GW | 2025; 2026 Snapshot | International organization | IEA PVPS Snapshot | No | Medium | Yes, E02 |
| E04 | First Solar annual financials | Revenue $5.219376bn, gross profit $2.120339bn, expenses $0.523475bn; gross margin 40.6% | FY 2025 | Audited company filing | SEC 10-K | No | Medium | No |
| E05 | Jinko cycle | FY 2023, 2024 and 2025 operating margins 5.1%, -3.6%, -13.6%; FY 2025 revenue RMB 65.50bn and operating loss RMB 8.91bn | FY 2023–2025 | Company filing | SEC 20-F | No; its release repeats same source [E15] | Medium | No |
| E06 | Jinko Q3 | Revenue RMB 13.61974bn, operating loss RMB 0.079974bn, operating margin -0.6%, gross margin 12.3% | Q3 2026 | Company release | Jinko Holdings | No | Medium | No |
| E07 | Grid construction and congestion | Grid planning to completion can take 5–15 years; congestion raises curtailment | 2026 IEA report | International organization | IEA Electricity 2026 | No | Medium | No |
| E08 | First Solar Q3 and guidance | Q3 revenue $1.230936bn, gross profit $0.713060bn, operating expenses $0.166952bn; annual 45X assumption $2.07–2.19bn | Q3 actual; October 2026 annual guidance | SEC filing/company guidance | SEC 10-Q; company results | No, both company disclosures | Medium | No |
| E09 | Indian module nameplate capacity | Around 233 GW/year, with weaker upstream stages | June 2026 estimate | Industry research | IEEFA/JMK | No | Medium | No |
| E10 | US import measures | Minimum import prices and additional duty announced; applicable from 4 December 2026, subject to scope and exceptions | August 2026 announcement; future effective date | Regulation | Federal Register | No | Medium | No |
| E11 | US regional supply | Around 66 GWdc nameplate module capacity; distributed-module Q2 average $0.37/Wdc | Q2 2026; Q3 2026 report | Industry research | SEIA | No; direct page access blocked, search excerpt only | Low | No |
| E12 | Negative prices and flexibility | Negative prices can reflect technical, regulatory or contractual inflexibility | 2026 report | International organization | IEA Electricity 2026 | No | Medium | No |
| E13 | Upstream surplus and pricing | Polysilicon price fell as low as $4–5/kg | Q2 2024, 2024 report | International organization | IEA Renewables 2024 | No | Medium | No |
| E14 | Supply-chain concentration | China expected to exceed 90% of polysilicon and wafer capacity | 2026 estimate/forecast | Industry research | S&P Global | No; supports E01 direction but not same period or scope | Medium | No |
| E15 | Jinko annual loss | FY 2025 operating loss RMB 8.91bn | FY 2025 | Company release | Jinko correction | No, same company as E05 | Medium | No |
| E16 | Regional prices and consolidation | Oversupply, price divergence and consolidation coexist | 8 October 2026 | Professional media / industry interview | pv magazine interview with S&P | No | Medium | No |
| E17 | Policy and prices | Support and excess capacity affect prices and profitability | 15 January 2026 | International organization | OECD | Yes, IEA supports surplus direction [E02], not the same quantified statistic | Medium | No |
| E18 | Enphase annual consolidated financials | Revenue $1.472985bn, operating income $0.157526bn; includes manufacturing credits | FY 2025 | Company filing | SEC 10-K | No | Medium | No |
| E19 | Enphase quarterly consolidated financials | Revenue $0.293329bn, operating income $0.006466bn | Q3 2026 | Company release | Enphase | No | Medium | No |
10. References and source tiers
Primary financial: First Solar SEC 10-K and 10-Q, Jinko SEC 20-F and quarterly releases [E04][E05][E06][E08][E15][E18][E19]. Industry and international organizations: IEA, IEA PVPS, IEEFA/JMK, S&P, SEIA and OECD [E01][E02][E03][E07][E09][E11][E12][E13][E14][E17]. Professional media: pv magazine [E16]. Regulation: US Federal Register [E10]. All accessed 2026-11-04; dynamic pages and forecasts are time-bound. Direct access to E11 failed and the search excerpt alone was used; confidence is low.
This is the Day11 research draft. Evidence dates and definitions govern the figures; forecasts and scenarios are not realized facts.