Variant Perception

Figures converted from INR at historical FX rates — see data/company.json.fx_rates. Ratios, margins, and multiples are unitless and unchanged.

Where We Disagree With the Market

Sharpest variant view: the 0.74× P/B "value" anchor that Indian retail investors and aggregators are pricing off does not survive a one-page reconstruction of book value — restated tangible equity is closer to $0.29–0.33/share, which means the real price-to-book at $0.33 is roughly 1.0×, not the headline 0.74×. The market has priced the regulatory tail (SEBI forensic audit, ED chargesheet) as a binary catalyst on a 6–12 month clock — bull side underwrites a clean close into a 1.0× book mean-reversion, bear side underwrites restatement to a 0.7× tangible-book floor. Our reading of the same evidence says both sides are wrong about the time horizon and the cushion: the cushion is already gone in the no-action case, the FY26 interest-coverage rerating is at least partly a perimeter shift rather than an operational turnaround, and the regulator-clock catalysts are far more likely to drag for 18–30 months than to resolve cleanly inside Bull's mid-2026 window. The variant view does not require believing the company is fraudulent — only believing that the equity is fairly priced, not cheap.

Variant Perception Scorecard

The 68/100 variant strength reflects a real but bounded disagreement: this is not a case of the market mispricing growth or moat — those are well understood. The disagreement is about the floor. Consensus clarity is moderate (62/100) because traditional sell-side coverage is absent — pricing is anchored to retail aggregators (Screener.in, Stockopedia, Bitget) and to the binary SEBI/ED catalyst rather than to discounted-cash-flow models. Evidence strength is high (74/100) because the book-value composition, the deconsolidation mechanics, and the 57% drawdown / retail ownership concentration are all directly observable in filings and the tape. Time-to-resolution sits at 18 months because that is the realistic window in which the SEBI forensic audit, the Samalkot LCIA arbitration, the ED chargesheet, and the auditor rotation collectively land — not the 6-month window the bull case requires.

Consensus Map

Binder Error: Set operations can only apply to expressions with the same number of result columns

The most important row in this table is the last one. Reliance Power has no active major Indian sell-side coverage (Brave search of ICICI Securities, Motilal Oswal, Kotak surfaced no active target prices; published price-target consensus is from retail aggregators only). With 57.75% public retail ownership, 14.12% FII, and 3.12% DII, almost three-quarters of the float is in the hands of retail and momentum vehicles whose price discovery is dominated by Screener.in's "Pros and Cons" boxes and aggregator headlines, not by DCF or peer-multiple work. That is what makes a structural variant view possible at all — there is no institutional consensus to challenge, only a retail framing that has not been audited.

The Disagreement Ledger

Disagreement #1 — the broken P/B anchor. Consensus would say the stock trades at 0.74× book and that is a defensible floor for any India listed power asset; even NHPC and NTPC trade above 2× book, so mean-reversion is the easy trade. Our reading of the forensics and people tabs says the $0.444 stated book embeds $133M of non-cash revaluation reserve, $768M of FY22-FY25 equity created by converting promoter-affiliate debt to equity (rather than from retained operating earnings), and an unprovisioned $185M Samalkot parent-guarantee invocation now in arbitration. Adjust those out and tangible/durable equity per share falls from $0.444 to roughly $0.29-0.33 — making the real P/B 1.0-1.15×, not 0.74×. What the market would have to concede if we are right: the entire "value-with-regulatory-tail" framing is nominal — there is no value cushion in the no-action case, and the only way the stock works is on operational improvement, not on book-value reversion. Cleanest disconfirming signal: the FY26 annual report's own equity bridge, which will reconcile stated book to tangible book line-by-line — if our adjustments are wrong, the auditor's note will say so explicitly.

Disagreement #2 — the perimeter-shift dilution of the rerating ladder. Consensus reads the Q1 → Q3 FY26 interest-cover sequence (1.33× → 1.56× → 1.63×) as proof that the rerating trigger has fired. Our quarterly reconstruction shows that interest expense fell from $66M in Q1 FY25 to $41M in Q3 FY26 — a ~$20M/quarter reduction whose timing matches the September 2024 VIPL deconsolidation rather than ordinary debt amortisation alone. Underlying 9M FY26 PAT of just $17.5M ($0.006/share annualised) implies a trailing ROE of ~1%. What the market would have to concede if we are right: the bull's "underlying earnings × NTPC multiple" rerating math doesn't have a numerator — at $0.006 EPS the stock is on 78× P/E, not 16×. Cleanest disconfirming signal: the Q4 FY26 print in May 2026. Cover ≥ 1.5× ex-exceptional combined with quarterly underlying PAT > $17M falsifies the perimeter thesis; cover ≥ 1.5× combined with PAT < $11M confirms it.

Disagreement #3 — the multi-year overhang. Consensus is binary: bull anchors to a clean SEBI close inside 2026, bear anchors to a forced restatement on the same clock. Both sides believe the regulator-controlled clocks resolve cleanly inside 12 months. Indian regulatory precedent — RHFL took 18 months for the forensic audit alone, the underlying SEBI order took 5 years; Reliance Capital's IBC resolution ran 3 years — says the realistic resolution window is 18-30 months with continuous quarterly disclosure friction. What the market would have to concede if we are right: sizing the position as a binary catalyst trade (long for May 2026 print, short for SEBI restatement) is the wrong implementation; the dominant scenario is range-bound choppy trading at 0.7-1.1× stated book for two years. Cleanest disconfirming signal: an explicit SEBI scope letter or interim findings memo released inside the next 90 days. Three quarters of silence from SEBI confirms the overhang thesis in real time.

Disagreement #4 — Sasan as a wasting 13-year option. Consensus treats the bull's $4.5-6.7B Sasan asset value as a perpetual fortress that backstops the entire equity at multiples of market cap. The 25-year PPA actually expires around 2039, leaving roughly 13 years of contracted cash flow on a $0.013/kWh tariff against an Indian solar+BESS substitution path priced at $0.027/kWh. Realistic 13-year NPV is closer to $1.7-2.8B. What the market would have to concede if we are right: Sasan still covers the equity, but with 1.5-2× cushion, not 4-5× — meaning bear's contingent-liability stack ($185M Samalkot + ~$234M RSTEPL impairment risk + ~$385M restated VIPL gain reversal) eats more of the cushion than bulls assume. Cleanest disconfirming signal: any CERC tariff order extending Sasan or pricing its late-life true-up; or competitive bid data showing greenfield UMPP-style coal still wins on landed cost vs hybrid renewable+BESS at FY26 tender clearings.

Evidence That Changes the Odds

How This Gets Resolved

The first three rows are the high-leverage signals. The Q4 FY26 print on its own does not resolve the variant view — it has to be paired with the underlying quarterly PAT trajectory to distinguish "perimeter shift made cover look better" from "real operational rerating happened." The auditor rotation in August/September is a single-event signal that is hard to fudge — either a Big-4 firm signs off cleanly on the opening balance, or it doesn't. The FY26 AR equity bridge will publicly settle the most important factual question — whether tangible book ex-revaluation-reserve is closer to $0.39 (consensus) or $0.29-0.33 (variant).

What Would Make Us Wrong

The cleanest path to being wrong on the broken-P/B disagreement is the SEBI forensic audit closing without restatement and the FY26 statutory audit (signed by an incoming Big-4 firm with no opening-balance qualification) explicitly endorsing the related-party preferential issues, the December 2024 land revaluation, and the VIPL deconsolidation accounting. If both regulators stand behind the stated book, then the $0.444 figure becomes harder to dismiss as nominal, and the 0.74× P/B reverts to a defensible value floor regardless of how the underlying transactions look to a forensic eye. The Samalkot LCIA arbitration ruling that no debt was due removes the largest single concrete adjustment; full conversion of the residual 33.4 cr warrants by RInfra and Authum at $0.39 would put ~$122M of fresh cash on the balance sheet and visibly contradict the "smart money is walking away" reading of the April 2026 forfeiture.

The cleanest path to being wrong on the perimeter-shift disagreement is the Q4 FY26 print delivering both ≥ 1.5× interest cover ex-exceptional and underlying quarterly PAT > $17M — implying ~$67M of annualised underlying PAT, ~$0.016 EPS, and a forward P/E of 20× rather than 78×. If that prints in mid-May 2026, the rerating ladder Bulls cite is genuine and the perimeter-shift attribution we have used overstates VIPL's drag. We would also be wrong on this disagreement if the new auditor explicitly attests that the consolidated income statement is comparable on a like-for-like basis (i.e. that VIPL's removal does not mechanically explain the $20M/quarter interest-line reduction).

The cleanest path to being wrong on the multi-year-overhang view is straightforward: SEBI publishes interim findings inside 90 days that are either definitively adverse (validating bear) or definitively clean (validating bull). If the regulatory clock is faster than RHFL/RCap precedent, the binary catalyst framing is right and the choppy-range reading is wrong. The same applies to a settled Samalkot arbitration inside 12 months or an ED chargesheet that is either narrowed to former-CFO-only liability or expanded to current management quickly enough for the market to re-price decisively.

We are also explicitly not claiming the stock is short-able into the next 6 months — disagreement #1 says the value cushion is illusory, not that the equity is impaired. The risk we are most worried about being wrong on is the implementation variant: a tape with 109% realised vol and 75% retail ownership can squeeze 30% higher on no news, regardless of whether tangible book is $0.29 or $0.44.

The first thing to watch is the Q4 FY26 results filing in mid-May 2026 — specifically the underlying quarterly PAT line ex-exceptional, paired with the consolidated income-statement bridge to interest expense.