Adani Green — five gigawatts delivered, and a grid that couldn’t keep up

The State of Play

Adani Green Energy ended FY26 with 19.3 GW of operating renewable capacity — up 35% in a year — having added 5.1 GW against its 5 GW promise, sold 37.6 billion units of electricity, and posted an EBITDA of ₹10,865 crore at a 91.2% margin. It also ended the year with ₹1,03,545 crore of borrowings against ₹19,965 crore of equity, net profit that fell slightly to ₹1,987 crore despite 15% revenue growth, a December quarter in which profit collapsed to ₹5 crore, and an FY27 capacity plan that was halved within three months — not because the company can’t build, but because the grid can’t absorb what it builds. The stock trades at 137 times those earnings.

The Company

AGEL is a holding company whose subsidiaries develop and operate solar, wind, and hybrid plants — the centrepiece being Khavda in Gujarat’s Rann of Kutch, planned at 30 GW, which management calls the world’s largest renewable energy plant (~9.4 GW operational by March 2026). The economics are the inverse of a thermal generator: fuel is free, so operating margins run above 90%, and essentially all the cost sits below the EBITDA line as interest and depreciation. Roughly 27 GW of the pipeline is locked into 25-year PPAs at tariffs as low as ₹2.42/unit; the rest sells as merchant or “infirm” (pre-commissioning) power. In FY26 the company added two new layers: battery storage (1.4 GWh installed at Khavda — its ~3 GWh total roughly equals all of India’s installed battery capacity) and a 500 MW pumped-storage project nearing completion in Andhra Pradesh.

The capital structure tells its own story. Free cash flow has been negative in five of the last six years (−₹15,857 crore in FY26); investing outflows of ₹26,227 crore were funded by financing inflows; and no dividend has ever been paid. Promoters have raised their stake from 56.3% (mid-2023) to 62.44%, while FII holding halved from 18.3% to 11.1%. ROCE is 7.4% — the price of carrying a third of the asset base as construction-in-progress.

The Story So Far

July 2025 — “fully funded,” and a merchant price problem

The Q1 FY26 call had the swagger of a company mid-sprint: operational capacity up 45% to 15.8 GW, energy sales up 42%, EBITDA up 31% at a 92.8% margin, and CEO Ashish Khanna anchoring the year at 5 GW of additions — “we have already done around one-third” — within a “fully funded” 50 GW-by-2030 plan. The fine print was about prices and wires. Energy sold grew 42% but revenue only 31%, because merchant solar had slumped to ₹2.20/unit from ₹3.00 the prior quarter. And Khavda’s transmission was lagging the build — a drag Khanna sized at “less than 5% of our EBITDA,” promising the gap was “a matter of weeks or months, nothing more than that.” On battery storage, the answer was a teaser repeated twice: “There is a strategy in place… please be reassured.”

October 2025 — committed to 5, silent beyond it

H1 closed with 2.4 GW added (74% of all of FY25’s additions in half a year) and capacity up 49% to 16.7 GW. Asked to raise the 5 GW guidance, Khanna declined with a line worth keeping: “the words come cheap. So, we would like to first demonstrate what we are doing.” The numbers given instead: FY26 capex ~₹30,000 crore, FY27–28 at ₹30,000–35,000 crore a year, leverage at 4.4x operational EBITDA (5.1x including construction debt) to be held at 4–5x before falling “sharply from FY29.” Mohit Kumar of ICICI extracted the most checkable promise of the period — that the guided ₹13,600 crore run-rate EBITDA would be achieved in Q3/Q4: “Yes. Definitely, Mohit.” Meanwhile solar grid availability had dropped from ~97% to ~89%, and merchant prices stayed soft (blended ~₹3.5/unit). BESS strategy: still “shortly.”

January 2026 — the weak quarter, and a 10 GW flourish

Q3 was where the seasonality, the curtailment, and the merchant slump landed at once. An analyst put it bluntly — why is revenue going down while capacity goes up? — and Khanna, to his credit, didn’t dispute the math: “your observation is very right… there has been a dip.” Wind speeds were poor, the expected Khavda grid augmentation slipped from Q3 to Q4, and merchant solar fetched ₹2.20 against ₹2.82 a year earlier. The screener tables show what the call’s percentages didn’t: December-quarter net profit of ₹5 crore, with pre-tax profit actually negative. Leverage had crept to 4.6x/5.6x from October’s 4.4x/5.1x. Pressed three times on what tripled silver prices do to project IRRs, management insisted its old conservative bids had headroom — “aggressive bidders may face certain issues, but that is not the case with Adani Green” — without quantifying.

Yet the same call carried the period’s boldest claim: “another 10 gigawatt to be added in the next year,” predominantly from Khavda, contingent on transmission arriving by end-CY26. Battery storage finally got numbers — 3.5 GWh committed for FY26, more than double that in FY27.

April 2026 — the target met, the flourish walked back

The year-end call (a short fixed-income call — one question, no equity Q&A) confirmed the headline delivery: 5.1 GW added against the 5 GW promise, 19.3 GW cumulative, energy sales up 34%, EBITDA ₹10,865 crore, up 23%. A new Japanese credit rating (JCR BBB+, sovereign-equivalent) was the balance-sheet trophy. But the FY27 plan was now “4.5 to 5.0 gigawatt” — half of January’s 10 GW talk — hedged explicitly “in view of the evacuation constraints.” The Khavda battery target was raised to 10 GWh by year-end. And the moderator’s question about cost of debt, funding mix, and maturities went unanswered in the recorded portion; the call ended without a follow-up.

The FY26 annual report, read alongside, shows the strategic tilt the calls telegraphed: storage, hybrid power, C&I and merchant offtake elevated to formal strategy pillars next to the PPA model — a hedge against exactly the curtailment-and-merchant-price problem the year exposed.

Where Things Stand

On its core promise — building — AGEL delivered exactly: 5 GW promised, 5.1 GW built, the fourth consecutive year of record additions, at unit costs (₹4–4.5 crore/MW solar) and tariffs that lead the country. The misses live everywhere downstream of the meter: merchant prices fell all year, curtailment hit the infirm book, wind didn’t blow in Q3, and the transmission grid — owned by others, including a sister company — became the binding constraint, formally halving the FY27 plan. The balance sheet carries the strain of the sprint: borrowings up ₹23,500 crore in the year to over ₹1 lakh crore, interest of ₹6,484 crore consuming 60% of operating profit, and net profit flat for two years running while capacity nearly doubled.

The equity case management offers is that this is timing, not structure: leverage holds at 4–5x through the build, then falls sharply from FY29 as ~₹17,000 crore of run-rate EBITDA matures into reported EBITDA and capex tapers toward the 50 GW finish line. The thing to watch in FY27, by the company’s own framing, is none of its own numbers — it’s whether the grid shows up: ~10 GW of Khavda evacuation promised by end-FY26, the rest “by the end of this calendar year.” Every forward number now bends around that.

The Four Checks

1. Quality and moat. A formidable builder in a business with no real moat. Power generation is a commodity — a unit from Khavda is identical to a unit from anyone else’s plant, and new capacity is won at competitive auction, which hands the economics to the buyer. What AGEL has instead is execution advantage: country-leading unit costs (₹4–4.5 crore/MW solar), the scale of a 30 GW single site, ~27 GW of pipeline locked into 25-year PPAs that make existing cash flows durable, and the Adani group ecosystem around land and transmission. That’s a cost-and-scale edge, not a structural one — the year itself showed the binding constraints (the grid, merchant prices at ₹2.20/unit) sit outside the company’s control. The remaining checks should be read with that in mind.

2. Returns on incremental capital and runway. The runway is enormous — 19.3 GW operating against a 50 GW-by-2030 target, with ₹30,000–35,000 crore of capex a year planned — but the returns on it are thin. ROCE is 7.4% and ROE 11.4% (13.1% over three years), and while a third of the asset base sitting as construction-in-progress flatters neither number, even the mature math is levered-utility economics: tariffs as low as ₹2.42/unit were bid to win, and FY26 net profit of ₹1,987 crore actually fell despite 15% revenue growth because interest of ₹6,484 crore ate 60% of operating profit. Lots of road, modest mileage per rupee.

3. Capital allocation for the stage. Internally consistent, externally aggressive. For a build-phase company, reinvesting everything and paying no dividend (zero payout every year on record) is the right shape; the company has also delivered what it funded — 5.1 GW against a 5 GW promise, four consecutive record years. The quibbles are the scale of the bet: borrowings up ₹23,500 crore in a year to ₹1,03,545 crore against ₹19,965 crore of equity, free cash flow negative in five of six years (−₹15,857 crore in FY26), leverage at 4.6x/5.6x with the deleveraging promise pushed to FY29 — all deployed at single-digit ROCE into projects gated by transmission the company doesn’t own. Promoters raising their stake from 56.3% to 62.44% is a real signal of conviction; the screener flag that interest may be capitalised is a real reason to keep checking the denominator.

4. Price. Demanding by almost any yardstick. As of the June 2026 snapshot the stock trades at ₹1,512 — 136 times FY26 earnings, 12.5 times book, zero yield — and adding the ₹1.03 lakh crore of borrowings to the ₹2.49 lakh crore market cap puts the enterprise at roughly 32 times FY26 EBITDA. That price assumes the 50 GW gets built, the grid shows up on schedule, merchant prices recover, and leverage unwinds from FY29 exactly as promised — for a business earning 7.4% on capital today. Priced for a flawless decade, in a year that just demonstrated how much of the outcome sits outside the company’s hands.

Sources

• Concall transcripts read: Q1 FY26 (July 29, 2025), Q2 FY26 (October 29, 2025), Q3 FY26 (January 23, 2026), Q4/FY26 fixed-income call (April 24, 2026 — partial: opening remarks and a single question only; no equity-analyst Q&A took place on it). An Oct 2025 entry was PPT-only. No Q4 FY25 call was retrievable, so the chronicle starts at Q1 FY26.
• Annual reports read: FY24, FY25, FY26 (high-signal sections — these extracts were strategy/ESG/risk-heavy with almost no financial tables; financials above come from the screener snapshot instead).
• Financial tables: screener.in snapshot fetched 2026-06-05 (logged-out session).
• Research dumps: vault/Sources/Earnings/Adani Green Energy Ltd/ (not published).