Canadian Solar Inc. (CSIQ) | The Buildout — AI Infrastructure
The Verdict
Canadian Solar is a global solar technology and renewable energy company. Through its Manufacturing segment it produces solar modules and battery energy storage systems, while Recurrent Energy develops utility-scale solar and storage projects. The company's e-STORAGE division is directly positioned to serve the surging electricity demands of AI data centres, making it a key player in the infrastructure buildout supporting artificial intelligence.
| Market Cap | — |
| Revenue (TTM) | $5.5B |
| Revenue Growth | −6.6% |
| EBITDA Margin (TTM) | 5.8% |
| Net Debt | $6.4B |
| Earnings Beats | 3 of 7 |
| P/E (TTM) | — |
| EV/EBITDA (TTM) | — |
What We Like
- U.S. HJT cell factory – the first and only commercial HJT solar cell facility in the United States – produced trial cell on time; 6.3 GWp total planned capacity once Phase II completes.
- Internal battery cell production cost below market prices of third-party cells, providing a structural margin advantage in storage.
- Storage contracted backlog of $3.5 billion, with 34 GWh under long-term service agreements, a shift toward recurring revenue.
- HJT modules already booking a 10–15% price premium over TOPCon in contracts, according to management.
- U.S. solar module ASPs have risen $0.02–0.03/W in 2026, benefiting from tight domestic cell supply.
What We’re Watching
- Jeffersonville Phase II trial production slipped to early 2027; a further delay would push back the full 6.3 GWp cell capacity benefit.
- First Solar ITC Section 337 investigation over TOPCon patents – initial determination expected within ~11 months, potentially disrupting module imports if exclusion order granted.
- Storage gross margin not disclosed; management avoided providing a specific figure, leaving profitability opaque.
- Recurrent Energy posted $60M operating loss in Q1; asset sales are needed to reduce the $6.8B group debt burden.
The thesis that Canadian Solar is transitioning to a higher-margin, U.S.-centric manufacturing and storage model remains intact. The Q1 beat and technological milestones support the direction, but the near-term margin normalisation and Jeffersonville ramp remain critical proving points. The key open question is whether the company can execute the HJT cell ramp on time and convert the data centre pipeline into contracted backlog.
Earnings
Canadian Solar reported Q1 2026 revenue of $1.1 billion, hitting the high end of guidance, with gross margin of 25.1% boosted by an 860-basis-point tariff refund. Energy storage shipments reached 2.1 GWh, and the contracted storage backlog stood at $3.5 billion.
| Metric | Q1 FY2026 | Q4 FY2025 | Q1 FY2025 | YoY |
|---|---|---|---|---|
| Revenue | $1.1B | $1.2B | $1.2B | −9.9% |
| Gross margin | 25.1% | 10.2% | 11.7% | +1340bps |
| EBITDA | $73M | −$64M | −$55M | −233.0% |
| EPS | $-0.47 | $-1.27 | $-0.51 | −6.7% |
We’re very engaged right now on data center opportunities … we expect it to yield some pretty exciting results for us in the next quarters.— Colin Parkin, CEO, 14 May 2026
Management tone: Management’s tone reflected confidence in the strategic direction, tempered by acknowledgment of near-term headwinds. New CEO Colin Parkin emphasised disciplined execution, while Executive Chairman Shawn Qu provided unusually transparent detail on the tariff refund and HJT technology.
Management Guidance
For Q2 2026, management guided revenue of $1.0–$1.2 billion and gross margin of 13–15%, reflecting the roll-off of the one-time tariff refund and shipping congestion. Solar module shipments are expected at 3.1–3.3 GW, and energy storage deliveries at 2.8–3.2 GWh. The full-year U.S. module and storage shipment guidance was reaffirmed, and CapEx was raised to ~$1.3 billion to fund accelerated U.S. manufacturing.
Trajectory
Revenue has been choppy, with the latest quarter at $1.1 billion, down from $1.7 billion in Q2 2025, as the company deliberately scales back low-margin global module volumes. Gross margins spiked in Q2 2025 and Q1 2026 due to one-time items but normally sit in the low teens, while EBITDA has stayed near break-even outside of exceptional quarters. The profit-first strategy is funneling shipments to North America, where pricing is stronger, but the transition is weighing on top-line growth.
The Model
The Buildout model projects FY+1 revenue of $5.0 billion and EBITDA of $455 million, implying a 9.1% margin. FY+2 revenue rises to $5.9 billion with EBITDA of $732 million and a 12.4% margin. The near-term outlook reflects stabilisation in solar module margins and storage growth, while FY+2 benefits from the ramp of U.S. HJT cell production and higher-margin storage service contracts.
| Metric | FY2025 | Next FY (E) | Following FY (E) |
|---|---|---|---|
| Revenue | $5.6B | $5.0B | $5.9B |
| YoY Growth | — | −10.6% | +18.0% |
| EBITDA | $189M | $455M | $732M |
| EBITDA Margin | 3.4% | 9.1% | 12.4% |
Projections are the median of 5 independent model runs. The model’s revenue sits 19.8% below analyst consensus.
For Q2 2026, management guided revenue of $1.0–$1.2 billion and gross margin of 13–15%, reflecting the roll-off of the one-time tariff refund and shipping congestion. Solar module shipments are expected at 3.1–3.3 GW, and energy storage deliveries at 2.8–3.2 GWh. The full-year U.S. module and storage shipment guidance was reaffirmed, and CapEx was raised to ~$1.3 billion to fund accelerated U.S. manufacturing.
What Could Go Right — and Wrong
- Successful ramp and yield of Jeffersonville Phase I, delivering HJT modules at contracted 10–15% price premiums.
- Major hyperscaler master service agreements for data centre storage, adding multi-gigawatt-hour demand.
- Recurrent Energy asset sales at attractive prices, reducing group debt and restoring segment profitability.
- Storage internal cell cost advantage persists even as lithium prices rise, protecting margins.
- TOPCon 3.0 and zero-silver HJT breakthroughs sustain technology differentiation.
- Jeffersonville HJT ramp encounters technical delays or yield issues, pushing out premium revenue.
- A broad ITC exclusion order blocks imports of TOPCon modules during the transition period.
- Lithium carbonate prices remain elevated and compress storage margins despite internal cell production.
- Hyperscaler storage deals fail to materialise, and Fluence/Tesla capture the AI demand.
- Recurrent Energy debt load becomes unmanageable, forcing dilutive financing or CapEx cuts.
Looking Ahead
The next twelve months centre on the expected commercial start of Jeffersonville Phase I, the first HJT module deliveries, and the potential conversion of the data centre pipeline into named contracts. At the same time, Q2 margin normalization and Recurrent Energy's monetisation will test the company's ability to manage through the heavy capex year without further balance-sheet strain.
- Q3 2026Jeffersonville Phase I commercial op — First commercial HJT cell plant in the U.S. reaches production.
- Q3 2026First HJT module delivery — Tests 10–15% price premium in actual customer invoices.
- 2H 2026Mesquite expansion to 10 GWp — Doubles module assembly capacity to cover all U.S. volume guidance.
- H2 2026Data centre contract announcements — Management expects exciting results from hyperscaler engagement.
- Early 2027Phase II trial production — On-time trial would keep 6.3 GWp cell capacity on track.
Financials
Annual Summary
| Metric | FY2025 | TTM |
|---|---|---|
| Revenue | $5.6B | $5.5B |
| Gross Margin | 17.2% | 21.1% |
| EBITDA | $189M | $513M |
| EBITDA Margin | 3.4% | 5.8% |
| Net Income | −$104M | −$102M |
| Free Cash Flow | −$1.7B | −$3.9B |
| Net Cash | — | — |
Key Ratios (Trailing)
- P/E TTM—
- EV/EBITDA TTM—
- EV/Revenue TTM—
- Price/FCF TTM—
- Gross Margin (TTM)21.1%
- EBITDA Margin (TTM)5.8%
- Net Margin (TTM)-1.9%
- ROIC1.5%
- FCF Conversion-444.8%
- SBC / Revenue0.0%
The Company
Canadian Solar designs and manufactures solar ingots, wafers, cells, and modules, as well as battery energy storage products. Its e-STORAGE division produces utility-scale and behind-the-meter storage solutions that directly serve the power needs of data centres and grid support. The company also develops utility-scale solar and storage projects through its Recurrent Energy segment.
The company operates a vertically integrated manufacturing network: module assembly in Mesquite, Texas (expanding to 10 GWp), and a heterojunction cell factory in Jeffersonville, Indiana, which began producing trial cells in March 2026. It also produces LFP prismatic cells for storage at a cost below third-party market prices, giving it a structural cost advantage.
Business Segments
Competitive Landscape
The company competes with Chinese module giants (LONGi, Trina, Jinko) on a global scale, but is building a U.S.-centric manufacturing moat with its HJT cell factory. In storage, it faces Fluence and Tesla, but its vertical integration into cells sets it apart from integrators that rely on third-party supply.
- Thin-film rival; currently pursuing ITC Section 337 action against CSIQ over TOPCon patents.
- FluencePure-play storage integrator; signed hyperscaler master service agreements and >12 GWh data-centre pipeline.
- TeslaThe scale player with Megapack.
- Building a 2.1 GW TOPCon cell fab in Texas, targeting Q4 2026, a direct domestic competitor.
- LONGiLeading global module manufacturer; competes on cost and scale in international markets.
Supply Chain
CSIQ sits between raw-material suppliers and utility-scale power buyers, assembling modules and storage systems from polysilicon, wafers, and electronic components, and selling to developers and data-centre operators.