Adani Power — the self-funding story meets a ₹2 lakh crore bill

The State of Play

Adani Power is India’s largest private thermal generator — 18.15 GW of coal-fired capacity, 7.3% of the country’s coal-based generation — midway through a plan to more than double itself to 42 GW by FY32. The year under review was the awkward middle chapter: revenue dipped to ₹54,241 crore, plant utilisation slid as an endless monsoon suppressed power demand, merchant prices fell well below what management had guided, and borrowings jumped ₹15,000 crore in a year the company had opened by saying it wouldn’t need to borrow at all. Profit still came in at ₹12,971 crore, the expansion is visibly ahead of schedule, and not a rupee has ever been paid as dividend.

The Company

The business model is simple to state: build or buy coal plants, sell the electricity through a mix of long-term power purchase agreements with state distribution companies (about 90% of capacity), short-term contracts, and merchant sales on the exchanges (the remaining ~10%, which management wants down to 3–4%). Operating margins run in the high-30s. A house specialty is buying broken plants out of bankruptcy and reviving them — the FY24 annual report lists M&A as risk number one on its own register, which tells you how central acquisitions are; the Butibori plant bought in July 2025 had been shut for nearly a decade and was generating within two months.

The promoter is the Adani Group, holding 74.96% — parked just under the 75% regulatory ceiling for eight straight quarters. DIIs have climbed from effectively zero (0.04% in mid-2023) to 3.69%; FIIs have drifted from 17.5% to 11.7%. The screener page carries two standing caveats: the company has never paid an equity dividend despite years of profits, and it “might be capitalizing the interest cost.”

The three annual reports trace a clean arc: FY24 was harvest (capex ₹2,602 crore, below depreciation; liabilities down ₹6,800 crore), FY25 was the turn (capex 4.4x to ₹11,485 crore), and FY26 is full build (capex doubled again to ₹23,350 crore, segment assets up ₹28,600 crore, capital work-in-progress at ₹35,053 crore). The FY26 report also adds a new phrase to the strategy: “diversifying into nuclear and international projects.”

The Story So Far

May 2025 — a strong year, and a famous last word on debt

The Q4 FY25 call reported a year of genuine operating strength: 102 billion units generated at 91% availability, sales volume up 20.7%, continuing EBITDA up 14.8% to ₹21,575 crore. The stated plan: capacity to 30,670 MW by 2030, FY26 expansion capex of ₹13,307 crore, roughly ₹1.2 lakh crore over six years. On funding, Anupam Misra of group corporate finance was unambiguous:

“We actually do not even need to have any additional borrowing.” — capex to be met “from our internal accruals only”

Net debt to EBITDA stood at 1.4–1.5x; the credit rating was AA stable from four agencies. The known soft spots: merchant realisation had fallen to ₹5.93/unit from ₹6.92 the year before, and Bangladesh — sole customer of the 1,600 MW Godda plant — still owed roughly $900 million, though management insisted the old dues would be “liquidated.”

August 2025 — the monsoon arrives early, and so does the debt

Q1 FY26 ran into weather: all-India power demand shrank 1.6% as the monsoon landed in May, and Adani Power’s EBITDA fell to ₹5,744 crore from ₹6,290 a year earlier. Two quieter developments mattered more. First, total debt rose ₹5,600 crore in the quarter to ₹44,372 crore — described as “interim bridge funding,” three months after the no-borrowing declaration. Second, the company fully extinguished the unsecured perpetual securities owed to the promoter (₹2,579 crore principal plus ₹1,146 crore of distributions), cleaning up a related-party layer of the balance sheet. Bangladesh wired in over $500 million across June–July, deflating that worry. A 1,600 MW PPA was signed with Uttar Pradesh, and a new pollution-control ruling exempted most of the pipeline from installing flue-gas desulphurisation — a cost saving management declined to quantify.

October 2025 — the target nearly doubles

The Q2 call was the inflection: the capacity ambition jumped from ~30 GW by 2030 to 42 GW by 2032 — 23.72 GW of new build, all equipment 100% advance-ordered (eight 800 MW machines to L&T, the balance to BHEL), all land owned, total capex now “approximately ₹2 lakh crore.” The funding language evolved accordingly: a “significant portion” from internal accruals, with the gap bridged from domestic capital markets and banks. Debt reached ₹47,254 crore. The quarter itself was soft — PLF 62.8% against 66.9%, merchant realisation ₹5.37 against ₹5.88 — but management guided merchant rates back to “around ₹6 or so” for the second half. When a media report claimed a ₹70,000 crore six-year EBITDA guidance, IR disowned it on the call, adding, with some swagger, “we hope to actually achieve a better EBITDA than what they had projected.” The PPA machine kept running: L1 on Assam’s 3,200 MW tender, a bid pipeline of roughly 22 GW.

January 2026 — the merchant guidance misses, the build accelerates

Q3 was the weakest print of the period. Merchant realisation came in at ₹4.37/unit — against the ~₹6 guided three months earlier — as a monsoon that “started in May and extended to October” kept hydro and renewables running and pushed thermal’s share of the national mix down to 73% from 76%. EBITDA slipped 3% to ₹4,636 crore; nine-month PAT of ₹8,700 crore trailed the prior year’s ₹10,150 crore. Management called the renewable surge “transient” and pointed to the structural pivot embedded in the new contracts: in every new PPA, fuel is a 100% pass-through and EBITDA comes entirely from fixed capacity charges — paid for availability, not generation. The plants only have to be ready to earn.

The funding picture got its most explicit framing yet: ₹1.4 lakh crore of expected operating cash over the build period, leaving “approximately a ₹60,000 crore gap” from markets and banks. ₹7,500 crore of AA-rated NCDs had been raised at 8.0–8.4%. Construction ran ahead of plan — Mahan Phase-II at ~80%, with management volunteering that parts of the second 12 GW tranche would commission early. One governance footnote: CEO S.B. Khyalia was absent from the call, leaving the CFO to carry it.

The FY26 close — what the books say

The fourth quarter wasn’t discussed on any call in this set (the June 2026 transcript wasn’t retrievable), but the snapshot fills in the year: Q4 net profit of ₹4,271 crore — helped by ₹1,766 crore of other income and a 3% tax rate — took FY26 PAT to ₹12,971 crore on revenue of ₹54,241 crore, down 3.5% in the company’s first revenue decline of the period. Borrowings ended at ₹54,670 crore, up from ₹39,495 crore. Free cash flow turned negative (−₹2,817 crore) for the first time in years. ROCE has stepped down from 32% (FY24) to 23% (FY25) to 17% (FY26) as capital piles up in plants that won’t generate until FY27–FY31. The FY26 annual report also records that SEBI concluded two show-cause notices in September 2025, a late chapter of the 2023 short-seller episode.

Where Things Stand

The chronicle splits cleanly into what management controls and what it doesn’t. What it controls — construction, equipment, land, PPAs — has consistently run ahead of its own promises: targets raised twice, commissioning pulled forward, 9+ GW won from 14.5 GW of state tenders. What it doesn’t control — weather, merchant prices, demand — has consistently undershot: merchant realisation fell from ₹6.92 (FY24) to ₹5.93 (FY25) to ₹5.44 (9M FY26), and the ~₹6 H2 guidance missed by a wide margin in Q3.

The self-funding narrative deserves its own line. “No additional borrowing” (May 2025) became “interim bridge funding” (August) became “domestic capital markets and banks” (October) became “a ₹60,000 crore gap” (January) — and ₹15,000 crore of new debt on the FY26 balance sheet. Leverage remains modest against ~₹22,000 crore of annual EBITDA and the rating held at AA through all of it, so this is evolution rather than distress — but the gap between the May rhetoric and the March books is the single clearest pattern in the period. The structural bet underneath is sound in design: the new 24 GW earns from capacity charges with fuel passed through, making future EBITDA a function of uptime rather than coal prices or merchant whims. Whether the states’ demand projections justify 42 GW is the question one regulator has already asked — Rajasthan’s 3,200 MW tender was sent back over exactly that.

The Four Checks

1. Quality and moat. A scale business in a commodity product, with the moat residing in contracts rather than the electricity itself. Power is undifferentiated; what Adani Power has is 18.15 GW of operating coal capacity (7.3% of India’s coal-based generation), owned land and 100% advance-ordered equipment for the next 24 GW, and — most importantly — long-term PPAs covering about 90% of capacity, with every new contract paying fixed capacity charges and passing fuel through entirely. That converts new plants into availability-paid annuities, which is a real edge while the contracts run. But the counterparties are state discoms, the asset is thermal in a decarbonising grid, and the period itself showed how exposed the uncontracted sliver is — merchant realisation fell from ₹6.92 to ₹4.37/unit in two years. Call it a contracted-infrastructure position with execution skill on top, not a structural fortress.

2. Returns on incremental capital and runway. The runway is enormous and the returns are visibly stepping down. ROCE has gone 32% (FY24) to 23% (FY25) to 17.2% (FY26 snapshot) as roughly ₹35,000 crore sits in capital work-in-progress earning nothing — capacity that won’t generate until FY27–FY31. The plan is ₹2 lakh crore of capex to reach 42 GW by FY32, against ~₹22,000 crore of annual EBITDA, so there is no shortage of places to put money. The question is the rate: the new capacity-charge PPAs trade away merchant upside for contracted stability, which points to solid-but-moderate project returns rather than the 30%-plus the harvest years showed. Mid-to-high-teens returns on a very long runway is the honest read — with the caveat one regulator has already raised, since Rajasthan returned a 3,200 MW tender questioning whether the demand projections justify the capacity.

3. Capital allocation for the stage. Mostly right for a build phase, with credibility quibbles. The company is reinvesting everything — capex went ₹2,602 crore (FY24) to ₹11,485 crore (FY25) to ₹23,350 crore (FY26) — which is the correct posture while contracted projects are available, and the distressed-plant acquisitions (Butibori generating within two months of purchase) have been genuinely value-accretive. Extinguishing the promoter perpetuals (₹2,579 crore principal plus ₹1,146 crore of distributions) cleaned up a related-party layer. Against that: not a rupee of dividend in eleven years of the table, defensible now but a pattern that predates the build; the “no additional borrowing” pledge of May 2025 became ₹15,000 crore of new debt and a stated ₹60,000 crore funding gap within nine months; screener flags possible interest capitalisation; and the FY26 report’s new appetite for “nuclear and international projects” hints at scope creep. Rational in deed, loose in word.

4. Price. Demanding. As of the June 2026 snapshot, the stock trades at ₹225 — a ₹4.33 lakh crore market cap, 33.9 times earnings and roughly 6.7 times the ₹33.7 book value — for a business whose revenue just declined 3.5%, whose ROCE has halved in two years, whose free cash flow turned negative (−₹2,817 crore), and which pays nothing out. The multiple is a bet that the 42 GW build lands on time and the capacity-charge annuities re-rate the earnings base from FY28 onward; nothing in the FY26 numbers themselves supports 34 times. Thermal utilities with 17% ROCE do not usually command this price — the next six gigawatts are already paid for in the quote.

Sources

• Concall transcripts read: Q4 FY25 (May 1, 2025), Q1 FY26 (August 1, 2025), Q2 FY26 (October 30, 2025), Q3 FY26 (January 29, 2026). The Jun 2026 (Q4 FY26) link on screener served HTML instead of a PDF; several other listed quarters were PPT-only. So Q4 FY26 commentary is absent — its numbers come from the financial tables only.
• Annual reports read: FY24, FY25, FY26 (high-signal sections; chairman’s letters were not in the extracts).
• Financial tables: screener.in snapshot fetched 2026-06-05 (logged-out session).
• Research dumps: vault/Sources/Earnings/Adani Power Ltd/ (not published).