India has assembled the largest exit queue in its market's history: ₹4.7 trillion of intended share sales across 238 filings, opening onto a tape foreign investors sold at record pace for most of the year. The headline is an option book, not supply. What converts is what accepts the domestic bid's price. This report sizes the live window, prices the concession, and names who gets paid.
Gravitywell Research publishes independent research across five pillars — Credit Research, Capital Markets, Industry & Sector Research, Economy & Policy, and Risk & Analytics. Our work is analysis for readers who allocate capital, underwrite risk, or set policy across public and private markets — not regulated ratings or investment advice. We source to primary filings and official statistics, disclose our coverage and confidence, and mark every figure to its point in time.
Currency: ₹ = Indian rupee; ₹1 trillion = ₹1 lakh crore. Conversions at ₹95/USD (August 2026) unless the source states its own rate. Period basis: FY = April–March (FY26 = April 2025–March 2026); CY = calendar year; figures nominal. Dates: written out in running prose, abbreviated (Mon YYYY) in furniture — runheads, tables, chart labels, source lines; register and appendix as-of stamps are ISO (YYYY-MM-DD). Data as of 13 August 2026 unless dated otherwise. Estimates are ranges, not points.
Confidence tiers on figures: ● filed/official ◐ modeled/derived ○ Gravitywell Research estimate.
How to read: the Executive Summary (p03) and Investor Quick Read (p07) stand alone; the one-page tearsheet is the companion. Full sourcing in the Register (p36); underlying series in the Data Appendix (p40).
India's second-half exit window is real, and the queue behind it is the largest the market has ever assembled, but the ₹4.72 trillion headline is an option book, not supply1 — issuers convert only at the domestic bid's price, and that price runs 25–40% below the sponsors' last marks (the paused cohort's observed repricing73,74; ○, M.1) — the gap that is both why the window works and why the headline will disappoint. That band is measured to sponsors' private marks and is bracketed by two deals that refused to print; the transacted concession, measured to listed peers, runs −8% to −24% with NSE at parity. The evidence is already on the tape. Deals that launched this year cut their raises by up to 43% to clear.2 Deals that would not cut (PhonePe in March, Zepto in July) stood down instead.3 And the year that the press still calls a record is running 21% behind last year's pace in dollars, at $5.78bn priced by early August against $7.32bn a year before.2
The window's fate now rests on three prints rather than 238 filings. NSE, a pure ₹30,000 crore offer-for-sale still waiting on its regulator as of 13 August, tests one thing: will the domestic bid pay up for the market's own infrastructure?4 Jio Platforms, a ₹30,000–40,000 crore all-fresh issue with no seller in it, tests whether that bid extends to large-cap growth at ~47× earnings.5 The SBI Funds Management print of 21 July, subscribed 41.7× and up 6.85% on debut against a 17% grey-market promise, is the calibration: demand is deep, and it no longer pays the old premium.6
Those four numbers carry the argument. Conversion runs at a twelfth of the headline because supply is voluntary and repricing is not.7 The listing pop has collapsed to 3% because the marginal buyer is a domestic institution doing arithmetic, not a leveraged subscriber chasing allocation.8 The larger supply event is not the queue at all. It is the ~$31bn of lock-ups expiring June through September, before and inside the window's crush quarter.7 And the one live variable that could change the clearing price, foreign money returning, just printed its first positive month since February.9
Each seat gets a different use of the same finding. The finding: the window clears at the domestic bid's price. The clock runs out on 30 September for most of the queue, and by January for the biggest paused deal.10,3
Underwrite the concession, not the subscription multiple. A 140× institutional book no longer buys a pop: SBI Funds Management drew ~$25bn of QIB bids and still listed up only 6.85%, so the entry case has to survive on the pricing itself.6 NSE at the guided ₹4.7–6.0 lakh crore is 38–48× annualised earnings against BSE's 42.5× forward: an 11% discount at the bottom of the band on the forward basis, a premium at the top on either basis. The trade is entry at or below the midband, where the exchange carrying ~93% of cash turnover prices level with its junior peer. Above the midband you are paying up for a forced seller's paper, which is backwards.31,32 Treat anything still priced to its 2021 private mark, the paused cohort is the tell, as unpriced risk, not scarcity value.
Sell the strike, not the calendar. A 2019-vintage Category II fund at 0.5× DPI does more for its LPs taking a 30% concession now than holding for marks the clearing price no longer supports. Behind this window stands the $31bn lock-up wall, and behind that the 2027 refiling queue.25,7 Where the asset can carry it, the continuation-vehicle lane is now tax-clean and precedented at 6× to exiting LPs.28,27
For issuers, the April flexibility is the instrument. A fresh-issue resize of up to 50% without refiling turns a shelved deal into a smaller, cleared one — the Indo-MIM path, not the PhonePe path.33 For the policymaker, sequencing is the whole game. ~₹53,000 crore of PSU paper aimed at the same September–December bid as NSE and Jio risks crowding the window the state itself extended. The validity cliff already concentrates launches into one quarter.30,10
| Position | Instrument | The price | Basis |
|---|---|---|---|
| Anchor the exit window | NSE OFS | 38–48× ann. earnings | vs BSE 42.5× fwd; buy at/below midband32,107 |
| Rent the growth bid | Jio fresh issue | ~47× FY26 P/E · ~18× EV/EBITDA | at $148bn; Airtel comp ~$127bn mcap34 |
| Calibration, already priced | SBI MF (listed 21 Jul) | 38.1× FY26 P/E | peers 41.6×; +6.85% debut21 |
| The seller's trade | 2015–19 vintage exits | accept ~30% to marks | vs 0.5–0.8× DPI and the wall25 |
A view is only a view against a stated consensus. This page puts the two side by side, names the mechanism the street is mis-weighting, and states the observables that would prove consensus right and the house wrong.
| Row | Statement |
|---|---|
| The consensus | Sell-side CY26 forecasts cluster at $20–25bn — Goldman Sachs and Kotak's ~$25bn call dates from December 2025, before the record H1 outflow and the −21% tape; IIFL and JPMorgan sit above $20bn35,68,110; the press frames a record "₹4.7trn wave"1; Jefferies reads domestic inflows as fully absorbing $7–8bn/month of supply.67 No bank has publicly re-cut the December numbers through the record H1 outflow — the consensus is standing, not restated. No exchange-traded instrument prices the window itself; the sell-side number is the market-implied anchor. |
| The house delta | H2 clears $6–9bn (base, p 55%); CY26 lands $11–14bn, roughly half of consensus and 37–51% below CY25. Even the bull case stays under the $25bn call. NSE prices at parity to its peer, not at the deep discount early commentary claimed. |
| Why the mispricing | Three mechanisms. The pipeline-as-supply conflation: ₹4.72trn is a stock of permissions read as a flow forecast; even on an all-channel count, conversion runs at ≤~13%.1,7 The fiscal-calendar artefact: FY26's record masks a calendar tape down 21%.2 Stale demand reads: 2021-era subscription multiples and a 2025-vintage "$18bn H2" call still circulate as 2026 evidence.69 Each error persists because the headline series the street quotes are the fiscal-year and pipeline ones. |
| What proves consensus right | Five dated observables (the §25 box): NSDL net positive for August AND September; NSE listing inside Q4 at the top of its band with a double-digit pop; gross SIP above ₹32,000 crore through the crush; the paused cohort relaunching at its old asks; CY26 full-year proceeds above $20bn, clear of the house bull band. Any two force an edition revision toward consensus. |
On the mechanism, the house and Jefferies agree: the domestic bid absorbs the supply, and nothing here forecasts a market break. The delta is confined to price and print size — what clears, at what concession, against which peer multiple. If the observables above start falling, the desk moves toward consensus and says so in a dated edition note.
The bank forecasts need ~$19bn in five months against a $5.78bn eight-month run. Supply is voluntary; it cuts or waits rather than pay up.35,2
A forced, regulated seller priced at parity with BSE on matched bases: 38–48× annualised vs 42.5× forward. Enter at or below the midband; the top of the band is a premium to the peer.31,32
0.5× DPI vintages beat waiting: the gross lock-up wall behind this window is 5× its size as stock (~9% realised historically), and the CV lane is tax-clean from April 2026.25,28
100% fresh at ~47× earnings into the segment July flows say is thinnest — large caps. Watch the print; do not pre-pay it.5,15
The 30 Sep validity cliff pushes the approved book into one quarter alongside ~₹53,000cr of state paper. Indigestion is a Q4 event.10,30
Allocator: participate at the concession, in the anchor events, sized to the September crush. GP/LP: distribute into strength; use the CV lane; do not underwrite 2021 marks surviving 2027. Policymaker: the bid is structural but not infinite — sequence state paper around the anchor prints, not into them.
₹4.72 trillion of intended selling is not ₹4.72 trillion of supply. It is a book of options held by issuers who exercise only at their price — and most of it expires with the regulator's clock on 30 September.
The number that launched a hundred headlines is a stock of permissions, not a forecast of supply. Prime Database counts 238 companies intending to raise ₹4.72 trillion in the second half1, but only ₹2.77 trillion of that carries a SEBI approval, and the half just ended cleared 27 IPOs for ₹22,555 crore — under 5% of the stated pipeline.1,36 Bloomberg's count of what will realistically reach the market in H2 is ~$6bn (~₹60,000 crore) — and that count spans IPOs, stake sales and institutional placements, so the IPO-only conversion is lower still: at most ~13% of the headline. JPMorgan's intent framing runs to ~$40bn for the same half; nothing published reconciles the two, and we anchor Bloomberg's completed-issuance basis while carrying the spread.109 Definitions, not disagreement, drive most of that gap — the all-channel tape already ran past ₹1 lakh crore in the June quarter alone.377 The gap between the two numbers is not a delay. It is the price at which issuers decline to sell.
Even the approved bucket is softer than it looks, because the pipeline measures disagree with each other. Prime Database closed FY26 with 144 approvals worth ₹1.75 lakh crore; Uniqus counts 157 approvals for ₹2.38 lakh crore; the July headline carries 174 for ₹2.77 trillion. Different dates and different scopes explain some of the spread, and we show all three rather than pick one.1,36 What no measure disputes is the clock. SEBI's April relaxation extended observation letters expiring between April and September out to 30 September, so a material slice of the approved book exists only by forbearance.10
| Measure | Companies | Value | As of | Tier |
|---|---|---|---|---|
| Announced intent (Prime/BS) | 238 | ₹4.72 trn | Jul 2026 | ◐ |
| SEBI-approved (Prime/BS) | 174 | ₹2.77 trn | Jul 2026 | ◐ |
| SEBI-approved (Uniqus) | 157 | ₹2.38 lakh cr | Jul 2026 | ◐ |
| Approved at FY26 close (Prime) | 144 | ₹1.75 lakh cr | Mar 2026 | ◐ |
| Cleared in H1CY26 | 27 | ₹22,555 cr | Jun 2026 | ◐ |
| Realistic H2 supply, all channels (Bloomberg) | — | ~$6bn | Jul 2026 | ◐ |
Read the ₹4.72 trillion as a ceiling on ambition, not a floor on supply. The tradable number is the ~$6bn of all-channel supply that shows up priced to sell. The tradable date is 30 September, when the forbearance holding the approved book together runs out. Size positions to the smaller number; let the headline frighten someone else.
Follow the proceeds and the window changes character. Of the ₹2.37 trillion raised in IPOs from January 2025 through August 2026, ₹1.45 lakh crore (61.2%) was offer-for-sale, existing shares changing hands, and only ₹92,175 crore was fresh capital into a business.11 The pattern extends beyond listing day: one industry compilation puts post-listing block and bulk exits since 2024 at roughly ₹1.9 lakh crore, more than three times the ~₹59,000 crore that left through IPO offer-for-sale itself.12 India's primary market is functioning as private capital's cash register, and the queue should be read the same way.
The current fiscal year sharpens the same picture. Since April 2026, block deals alone exceed ₹55,000 crore (more than IPOs, qualified institutional placements and offer-for-sale programmes combined) while June's entire mainboard IPO tape was ₹2,718 crore.37 July then printed ₹52,300 crore across ~36 companies, a 20-month high — that month led by primary prints (SBI MF's ₹9,796 crore; an Adani placement near ₹15,000 crore), the supply arriving the moment price was available.38,37 Fresh-capital formation, the thing an IPO boom is supposed to finance, is the residual.
| Channel | Value | What it is |
|---|---|---|
| Block / bulk deals | >₹55,000 cr | Promoters and sponsors selling listed stakes |
| QIPs | ₹23,400 cr | Fresh capital, institutional |
| Offer-for-sale programmes | ₹18,700 cr | Mostly state PSU stake sales |
| Mainboard IPO (June) | ₹2,718 cr | The visible "boom," at its June run-rate |
| Total since 1 Apr 2026 | >₹1 lakh cr | Exit paper dominates every visible channel |
Price the window as a secondary distribution, because that is what it is. A market where 61% of IPO proceeds and the whole block tape are exits pays you for discipline on entry price, not for participation. The buyers funding capital formation get a growth story; the buyers funding an exit get whatever discount they insist on. Insist.
Every Indian issuance record has a sequel, and the queue's critics know the three by heart. The ₹52,900 crore raised in 2007 preceded the crash of 2008, with Reliance Power as the marker. India's largest-ever IPO to that date was 73× subscribed and sold out in a minute on 15 January 2008. Four weeks later it broke on debut and closed day one 17% under its ₹450 issue price.39 The 2021 wave (63 IPOs, ₹1.2 lakh crore, the best year in twenty) halved to ₹58–59,000 crore in 2022, with 19 of 52 FY22 listings below issue.40,41 The academic literature says this is not coincidence. Baker and Wurgler showed the equity share of new issuance predicts weak forward returns. Pastor and Veronesi showed waves form when valuations are high and expected returns are low.42,43
Two corrections keep this honest. First, the signal is weaker than the folklore. Jay Ritter, who has kept the issuance data longer than anyone, puts the hit rate of high-volume years predicting weak returns at "about 52% of the time". That is a coin with a lean, not a law.45 Second, the mechanism runs through price, not volume. Issuance peaks because valuations peak; supply is the thermometer, not the fever. The 2021 wave did not break the market — it repriced the new-age cohort and volume then self-corrected. The cohort's ledger below is what that repricing built. The operators that grew into their marks compounded. The one that never did still trades a quarter below its 2021 issue price, four and three-quarter years on.46
| Listing (Nov 2021 wave) | Issue | Now vs issue | The read |
|---|---|---|---|
| Eternal (Zomato) | ₹76 | ≈ +298% | Grew into the mark; the cohort's win |
| Nykaa (bonus-adjusted) | ₹187.5 | ≈ +78% | Compounded after a deep 2022 drawdown |
| PB Fintech | ₹980 | ≈ +63% | Repriced, then earned it back |
| Paytm (One97) | ₹2,150 | ≈ −25% | Max drawdown ~85% (May 2024); never recovered issue |
Price marks are indicative retail-platform quotes ○, direction confirmed across sources; Paytm has a conflicting ₹1,343 print. Nykaa measured against its bonus-adjusted issue price — the raw ₹1,125 comparison overstates the gain.
Use 2021 as the base rate for price, not apocalypse. A record queue into a choosier bid means the marginal deal reprices or withdraws; it does not mean the index breaks. The discipline is issuer-level: own the Eternals of this cohort at the concession, and let the Paytms find someone with more faith than arithmetic.
The awkward context for a "record Indian year" is that the rest of the world is actually having one. Global IPO proceeds tripled to $193.6bn in the first half, up 210% on the year, while India's fell 21%.47,2 Hong Kong alone raised HK$328bn (~$42bn) by end-July, up 154%, reclaiming the #2 slot India briefly contested. India stands third by proceeds, as it did across 2025, when its #1 ranking was by deal count only.48,49,47 The queue is not waiting for a global thaw. The thaw came, and it went elsewhere.
Nor is the macro coming to help. The Fed under Chair Warsh held at 3.50–3.75% in July with three dissents cast for a hike and a median end-2026 dot of 3.8% (a committee priced for tightening, not the cutting cycle the 2025 playbooks assumed) and the dollar is up on the year.50,51 The flows tell the harder truth. Emerging-market equity funds took a record $25bn in a single July week, and India-dedicated funds stayed in outflow. The money is choosing the AI-levered markets, and India's MSCI EM weight has slid to ~12% from a 2024 peak near 20%.52,53 The one genuine de-escalation is trade. February's deal cut the US tariff on Indian goods from 50% to 18%. But the Supreme Court then struck the statute it rode on, a fresh 10% measure landed in July, and the surviving legal basis is an open question we carry, not an answer we assume.54,55,56
| Variable | Level | Direction for the window |
|---|---|---|
| Fed funds / end-26 median dot | 3.50–3.75% / 3.8% | Against — hike bias, no easing bid |
| Dollar index (DXY) | 99.87 · +1.65% y/y | Against — EM headwind |
| Brent | $87.92/bbl | Against — import bill, post-Hormuz |
| US tariff on India | 18% + 10% (Jul) | Mixed — resolved in rate, contested in law |
| EM fund flows | record +$25bn/wk | Available — but skipping India so far |
| India weight in MSCI EM | ~12% vs ~20% peak | Against — passive bid shrank |
Underwrite the window on domestic arithmetic alone. The global bid exists; it re-entered EM at record pace in July. But it is conditional on India rejoining the AI-capex story or the tariff question resolving, and neither is on the H2 calendar. If foreign money returns anyway, treat it as upside to the clearing price, never as the base case.
The buyer facing the queue has changed hands. A record foreign exit met a record domestic bid, and the ownership tables crossed. The institution that now sets the price runs on a strong floor and a thinning margin.
The selling that framed this year was without precedent. Foreign portfolio investors sold ₹2.74 lakh crore of Indian equity in the first half, the largest six-month outflow since the market opened to foreign money in 1993 and a full-year pace that would eclipse 2022's record. Their share of NSE-listed value fell to 15.88%, a 14-year low.57,4,14 March alone took out ₹1.18 lakh crore, the worst single month on record, as the Iran conflict closed Hormuz and the Nifty drew down 11%.9,58 Then July printed +₹20,199 crore, the first positive month since February, and early August followed at +₹12,921 crore.9
The named reasons for the exit were four: repeated earnings downgrades, rupee weakness that reached a record 96.84 in May, a stubborn valuation premium at ~21× forward earnings, and, most painfully, no AI-capex complex to own while that trade absorbed every EM dollar.4,59,60 The July turn's composition is the detail that matters for the queue: ₹13,467 crore of the month's buying came through the primary market against ₹6,731 crore on the exchanges.9 Foreign money is not chasing the tape back. It is buying new paper at issue — at the window's price.
Treat the turn as real, conditional and priced. Two-thirds of July's foreign money entered through IPO books, which says the marginal foreign buyer will fund the exit window at the concession even while it stays underweight the index. Watch the August and September NSDL prints: a second and third positive month roughly doubles the window's absorption capacity; a reversal returns pricing power to the domestic bid alone.
The domestic bid did a record ₹8.5 lakh crore in FY26, and ₹6.4 lakh crore of it was mutual funds.13 That composition is the whole story of who can absorb the queue. The systematic investment plan is the price-insensitive core — ₹31,961 crore in July, a four-month high, roughly ₹3.8 lakh crore a year that arrives whatever the tape does.15 Around it sit institutions that look enormous on holdings and small at the margin: LIC holds ₹15.11 lakh crore of equity but bought a net ₹8,137 crore in the June quarter; the EPFO runs 10.57% of a ~₹31 trillion corpus in equity but only through index ETFs — it cannot bid for an IPO share at any price.61,14,62
| Institution | Equity position | The binding rule | Can it buy the queue? |
|---|---|---|---|
| Mutual funds | 11.58% of mkt · ATH | Scheme mandates; cash ₹1.84 L cr (CY26 low) | Yes — 33% of anchor books reserved |
| LIC + insurers | ₹15.11 L cr held | IRDAI: AIF/unlisted combined 3–5% caps | Selectively; 7% anchor reservation |
| EPFO | 10.57% of ₹31trn | 5–15% of incremental, index ETFs only | No — index inclusion later, never at issue |
| NPS | growing | 75% cap common; 100% option from Oct 2025 | Marginal today, compounding |
The marginal rupee that prices an IPO is the mutual fund's discretionary rupee, and only that. The SIP floor guarantees the bid exists; the ceilings guarantee it is finite and choosy. That is the mechanism behind every 40% deal cut in this report: the queue is negotiating with one buyer, and that buyer holds the arithmetic.
In March 2025, for the first time since the data begins in 2009, domestic institutions owned more of the NSE-listed market than foreign ones (17.62% against 17.22%), and the gap has widened every quarter since, to 19.15% against 15.88% in June 2026.63,14 Mutual funds alone hold an all-time-high 11.58%, their twelfth consecutive quarterly rise; the FII-to-mutual-fund gap, 17.14 points in March 2015, is down to 4.30.14 Ownership is pricing power. The institution that holds the marginal rupee writes the marginal price, and for the first time in the modern market's history that institution answers to Indian savers.
The primary market flipped faster than the secondary. The foreign share of IPO anchor books fell from ~63% in 2021 to ~51% in 2025, while the mutual-fund share rose from under 30% to ~36%. In the wider institutional segment the mutual-fund share more than doubled, to ~33%.64 Regulation then locked the shift in: since November 2025, 40% of every anchor book is reserved for domestic institutions — 33% for mutual funds, 7% for insurers and pension funds.65 Domestic participation in 2025's billion-dollar-plus IPOs ran to ~75%.64
| Share of the book | 2021 | 2025 | Direction |
|---|---|---|---|
| Anchor book — FPI | ~63% | ~51% | Foreign anchor fading |
| Anchor book — domestic MF | <30% | ~36% | + 40% reservation rule, Nov 2025 |
| QIB segment — MF | <15% | ~33% | The doubling that moved pricing |
| QIB segment — FPI | ~63% | ~52% | Still present, no longer decisive |
Retire the 2021 pricing grammar. Grey-market premiums and subscription multiples described a market where leveraged retail and foreign anchors set the clearing price; both have left the mechanism. The number that now decides a deal is the domestic anchor book's entry multiple, which is why SBI Funds Management priced below its peer average and why NSE will have to. The concession is not a sentiment reading. It is the new price-setter's arithmetic.
Underneath the record annual numbers, the discretionary margin of the domestic bid has been thinning all year. Equity fund net inflows fell 40% in May to ₹22,898 crore, a 12-month low, and July's ₹24,697 crore was 15% below June.16,15 The SIP stoppage ratio crossed 100% in March and April — more accounts closed than opened, so the rising rupee headline rides on existing investors stepping up, not new ones arriving.17 Scheme cash is down to ₹1.84 lakh crore, the lowest of the calendar year: the dry powder is being spent, not rebuilt.18
Composition is the sharper problem. July's equity money went where the queue is not: smallcap funds took ₹7,767 crore and midcap ₹6,192 crore, while largecap funds saw a net outflow of ₹1,322 crore.15 The two largest live deals (NSE at ~₹30,000 crore and Jio at ₹30,000–40,000 crore) are precisely large-cap paper, arriving in the one bucket the flow data says is shrinking. If you allocate to India primaries, this is the mismatch to underwrite: the bid exists, and it is standing in a different aisle from the supply.
| Bucket | Jul net flow | The queue's paper | Match? |
|---|---|---|---|
| Smallcap funds | +₹7,767 cr | tail of 238 | Oversupplied with money |
| Midcap funds | +₹6,192 cr | mid-tier issues | Adequate |
| Largecap funds | −₹1,322 cr | NSE, Jio: ₹60–70,000 cr | The mismatch |
The mega prints cannot rely on their natural bucket. They clear through one of three doors: the anchor reservation compels the allocation, the FPI turn holds and fills the gap, or the price concedes until multi-cap money crosses the aisle. Watch which door opens for NSE — it sets the template for everything behind it.
Supply is meeting the bid through price, not volume. This Part puts the bull case on the stand (three objections, cross-examined against the tape) then shows what the market has already paid, and the supply the headlines never count.
A thesis that supply must concede to clear deserves to face its strongest objections in their strongest form. The bull case against this report makes three, and Part II's evidence answers two of them. We take each in turn, on the record.
The record is real: ₹8.5 lakh crore in FY26, a third straight ₹5 trillion calendar year running.13,66 But the same sell side that cites the record has already netted it. Jefferies sizes domestic inflows at $7–8bn a month against $7–8bn a month of fresh equity supply, and reads the balance as "keeps the upside in check". A cap, not a cushion.67 A bid fully employed absorbing supply is not a bid that re-rates it. Sustained: the record proves the floor, not spare capacity.
Goldman and Kotak published ~$25bn CY26 forecasts in December 2025; IIFL and JPMorgan sit above $20bn, and none has publicly re-cut since.35,68,110 The tape has already voted: $5.78bn priced through 3 August means the forecasts need ~$19bn in five months — more than three times the year's monthly run-rate, sustained through the September validity cliff. And the "$18bn second-half surge" quote still circulating in support is dated 10 July 2025, about H2 2025.69 Sustained: the forecasts price the stated intent, not the exercise of it.
This one we cannot dismiss, and do not. July printed +₹20,199 crore and the first week of August +₹12,921 crore, with two-thirds of July's money entering through primary books.9 If the August and September prints hold positive, absorption capacity roughly doubles and the concession this report is built on narrows. The objection stands unresolved — it is the live variable of the second half, it appears again in the risk register, and the honest position is that we do not know.
| Objection | Its number | The tape's number | Verdict |
|---|---|---|---|
| Record DII flows absorb anything | ₹8.5 L cr FY26 | supply $7–8bn/mo = inflows | Fails — fully employed, caps upside |
| $25bn sell-side year | $20–25bn CY26E (Dec 25) | $5.78bn in 8 months | Fails — needs 3× run-rate in 5 months |
| The FPI turn | +₹20,199 cr Jul | 1 month vs 4 negative | Open — the window's live variable |
The bull case survives only where it is humble: not "flows conquer supply" but "foreign money might return to a market it is underweight." Position for the version that survives — take the concession where it is offered, and let the FPI prints, not the forecasts, tell you when to pay up.
The market answered the absorption question before anyone framed it, and it answered in price. The median listing gain fell to 3.0% in FY26 from 21.2% the year before; the average fell to ~8% from 28–30%, the base year disputed across compilers but the collapse not.8,44 Post-listing, the average FY26 IPO returned −7% to year-end.44 More than half of the 155 large IPOs of 2023–25 trade below issue, and only about a third of post-listing buyers beat the benchmark.19 A record issuance year with a vanished pop is a market clearing exactly as theory says it should: the price concedes until the paper moves.
Demand did not disappear; its quality changed. Average oversubscription fell from 71× to 39× and retail participation collapsed 65.6%, to 8.1× from 23.56× on the compiler's 18-issue sample, with roughly half of 2026's issues failing to fill the retail quota.70,20 The multiples that remain are made of different money. RBI's ₹1 crore cap on NBFC IPO funding, in force since April 2022, dismantled the seven-day leverage machine that manufactured 2021's books. Today's 39× is unlevered institutional demand and cannot be compared to 2021's headline prints.71
| Measure | 2025 | 2026 | The read |
|---|---|---|---|
| Average oversubscription | 71× | 39× | Real money, no leverage — deeper than it looks |
| Retail subscription (18-issue sample) | 23.56× | 8.1× | The 2021 marginal buyer withdrew |
| Large 2023–25 IPOs below issue | >51% of 155 | The cohort's buyers are underwater | |
| Post-listing buyers beating benchmark | ~32% | Participation ≠ alpha | |
Stop reading subscription multiples as enthusiasm; read them as an unlevered institutional book that pays fair value and no more. The repricing the bears expect from a supply shock has largely happened — deal by deal, at the point of issue. What remains is for the sellers still marked to 2021 to acknowledge it, which is Part IV's subject.
What a clearing price looks like in practice: Indo-MIM came to market asking ~$700m, cut 43% to ~$396m, and closed 72 times subscribed. Juniper Green Energy cut 40% and cleared. Manipal Health trimmed its billion-dollar ask to $960m and completed India's largest healthcare IPO.2,72 The deals that would not cut tell the same story from the other side: PhonePe's target valuation was marked from $12–15bn to $9.5–10.5bn before the deal was shelved outright, and Zepto's path ran $7bn to $4–5bn before its July pause.73,74 As DAM Capital's Dharmesh Mehta put it, companies are choosing to raise less rather than dilute at lower valuations.2
The completed prints put a number on the concession. SBI Funds Management priced at 38.12× FY26 earnings against a listed-peer average of 41.64× (ICICI Prudential AMC at 49.9×, Nippon Life at 48.2×), an 8% discount to the average and a 24% discount to the nearest comparable, from a seller with every negotiating advantage.21 NSE's guided band implies 38–48× annualised first-quarter earnings against BSE's 42.5× forward, and 46–58× trailing against BSE's ~54× — parity at the midband, not the deep discount early commentary claimed. Data providers print BSE anywhere from ~50× to ~65× trailing and ~42.5× forward; we anchor the cited screens.32,10731,32 The pattern is uniform: new paper prices inside its listed comparable, or it does not price.
| Deal | Priced at | Listed peer | Concession |
|---|---|---|---|
| SBI Funds Mgmt (done, 21 Jul) | 38.1× FY26 P/E | peers avg 41.6× | −8% to avg, −24% to ICICI Pru |
| NSE (guided band) | 38–48× ann. earnings | BSE 42.5× fwd · ~54× trl | −15% to +13% across bases — parity band |
| PhonePe (revised, then shelved) | 11.4–12.6× P/S | Paytm ~11.5× | parity — and still no deal |
The concession is not distress; it is the price of certainty in a buyer's market. SBI MF's 8–24% discount to its peers is the cleared evidence, and NSE's parity band is the frontier case: the senior franchise at the junior's rating, but only below the midband. The trade is taking the other side of the seller's urgency, at a price the peer multiple can defend.
The supply that will actually test the bid is not in the pipeline tables. Nuvama counts roughly $31bn of shares across 75-plus companies coming out of lock-in between June and September — about five times the ~$6bn of new issuance expected in the entire half and, on our arithmetic, ~0.6% of market capitalisation (○, M.1).7 The wall is rolling, not a one-quarter event: a January–April tranche ran $45bn across 96 companies105, and a May–August cutoff prints $34bn across ~73106; we anchor the June–September window on the $31bn Bloomberg print. None of it appears in Prime Database's fundraising series, because a lock-up expiry is not a fundraise — and not all of it will sell, since a sizeable share sits with promoters and group holders. All of it, though, becomes free to sell, and much of it sits with holders whose fund lives are expiring. Discipline cuts both ways, so apply the queue's own conversion logic here. H1's total open-market PE/VC exits ran $4.1bn against the $45bn January–April expiry tranche — a realisation rate near 9% (○, M.1). At that rate the June–September $31bn implies ~$3bn of actual selling: on the order of the new-issue print, not five times it. The 5× compares stocks. The realised flows are comparable — which is why the wall is a risk-register trigger with a calendar, not a crash forecast.
The first half's named selldowns preview the wall's mechanics. Kedaara Capital and Partners Group sold 14% of Vishal Mega Mart in one $835m block. Lenskart's November 2025 lock-up released three separate holders for $914m — 22% of all open-market exit value in the half, from one company's register. Groww's expiry let out Peak XV, Ribbit and Y Combinator in sequence.23,75 Each of these companies listed into the very window this report covers; their sponsors' second act arrived within months of the first. The 2025–26 listing cohort's own walls now stand behind the 2026 queue, and the anchor-investor 90-day cliffs land in Q4.
| Company (listed) | Sellers released | Value | Share of H1 open-market exits |
|---|---|---|---|
| Lenskart (Nov 25) | Three holders across three blocks | $914m | 22% — from one register |
| Vishal Mega Mart (Dec 24) | Kedaara, Partners Group — 14% stake | $835m | ~20% |
| Groww (Nov 25) | Peak XV, Ribbit, Y Combinator | ~$500m | sequenced exits |
| H1 open-market PE/VC exits, total | $4.1bn | 44% of all exit value | |
Underwrite the wall, not just the window. For the allocator, the concession you demand at issue must cover the second wave of paper already scheduled behind it — the 90-day anchor cliffs and the six-month sponsor releases. For the index investor, the pressure point is Q4 2026 through Q1 2027, when this year's listings meet their own expiries. The queue is the visible supply; the wall is the supply with a calendar.
For private capital, the window is the clearing price for a decade of unexited positions. The exit ledger is worse than its headline, the overhang is real once the folk number is retired, and the parallel lane has already repriced.
The exit recovery that justified the queue is one deal wide. EY's ledger shows CY2025 exits of $32.9bn, up 17% — but Temasek's sale of a 35% stake in Schneider Electric India accounts for $6.4bn of it, and stripping that single trade leaves ~$26.5bn, down roughly 6% on 2024.22,24 The first half of 2026 removed the ambiguity: $9.4bn across 95 exits, down 29% year on year and down 53% on the second half of 2025; the IPO route specifically fell 47% to $801m.23 PE and VC funds cashed a four-year-high ₹20,643 crore out of IPOs in 2025, yet their share of primary issuance hit a decade low — the sellers are using the window, and the window is still not big enough.76
The record's fragility shows in the sources themselves. EY books the Schneider trade as a strategic exit and reports CY25 open-market exits down 30%. Bain counts CY25 at ~$34bn, puts public-market exits at 41% of value, and books the same trade as a buyback. The two houses' route splits cannot be reconciled, the gap traces to that one classification, and we show both rather than adjudicate a difference the underlying data does not let us settle.22,24 Either way the direction in 2026 is shared: every route except open-market block sales shrank in the first half.23
| Route | EY ($bn / trend) | Bain (share / trend) | Agreement? |
|---|---|---|---|
| Strategic | 15.9 · +211% | 21% · +41% | Direction yes, scale no — Schneider sits here for EY |
| Open market | 9.0 · −30% | 41% (public mkt) | Different definitions |
| Secondary (sponsor-to-sponsor) | 3.9 · −41% | 17% · −28% | Both negative |
| IPO | 3.9 · +18% | in public mkt | — |
| Buyback | 0.1 | 21% · +1,244% | No — the Schneider booking; unresolved |
Read the exit "recovery" as unproven and the sellers' need as proven. A ledger whose direction flips on one deal's classification is not a trend; a first half down 29% with the IPO route down 47% is. The queue exists because the other exits are not arriving — which is exactly why the sellers will, eventually, take the window's price.
A "$100bn exit overhang" circulates in this market's commentary, and we went looking for its source: the trail ends at an uncited opinion column, in no Bain, EY, Preqin or SEBI dataset. We retire it. What the filed data supports is narrower and worse. SEBI's Category II AIF pool, where Indian PE and private credit live, holds ₹12.74 lakh crore of commitments against ₹4.44 lakh crore actually raised and ₹4.13 lakh crore invested77; and CRISIL's benchmark ladder shows what has come back: Indian 2019, 2020 and 2021 vintages all sit at 0.5× DPI, 2022 and 2023 at 0.1×, and even 2015–2017, funds now ten years old, have returned only 0.7–0.8× of paid-in capital.25
The global backdrop offers the sellers no rescue. Worldwide, PE distributions ran at 14% of net asset value in 2025 (a fourth straight year under 15%, territory last seen in 2008–09) with ~32,000 unsold portfolio companies worth $3.8 trillion in the backlog.78 Yet LPs have not given up on India. In McKinsey and IVCA's 2026 survey of 50-plus global LPs, 31% rank India first in Asia-Pacific and 76% place it top-three. More than half plan to increase India-dedicated allocations.79 That is the sellers' paradox in one page: the capital wants in, the capital already in cannot get out, and both facts price the same window.
| Measure | Value | Source basis |
|---|---|---|
| Cat II AIF commitments | ₹12.74 lakh cr | SEBI quarterly AIF data via press ◐ |
| — of which raised / invested | ₹4.44 / ₹4.13 lakh cr | SEBI via press ◐ — the deployed pool seeking exits |
| DPI, 2019–21 vintages | 0.5× | CRISIL benchmarks ◐ |
| Global unsold backlog | 32,000 cos · $3.8trn | Bain GPER 2026 ◐ |
| "$100bn India overhang" | retired | Untraceable to any dataset — do not cite |
Size the seller pressure from the DPI ladder, not a folk number. A 2019 fund at 0.5× DPI entering year seven has one mandate: convert marks to cash. That mandate, multiplied across every vintage since 2018, is the standing order flow behind the queue, the block tape and the wall, and it does not go away if the window disappoints. It just reprices.
While the IPO queue waits for its price, the other exit market has been printing at records. Global secondary volume reached $240bn in 2025, up 48%. The first half of 2026 set a record $121bn. GP-led deals, the continuation-vehicle lane, took $65bn of that — a 54% share, the highest ever recorded.80,26 The secondary buyer does not wait for a listing window. The discount to net asset value is the window, standing open at whatever width the seller will accept. Private capital's exit problem clears in this lane deal by deal, quietly, while the headlines watch the exchanges.
India's lane is open and precedented. ChrysCapital's ~$700m single-asset continuation vehicle on its NSE stake closed in April 2024. It paid exiting Fund VI LPs a 6× multiple. The position it rolled is marked roughly 3× higher within 18 months as the IPO approaches.27,81 Multiples closed India's largest multi-asset CV at $430m. Lightbox and TR Capital ran $100–150m vehicles. 360 ONE is raising a ₹4,000 crore dedicated secondaries fund, with Neo Asset a first close behind it.75,25 Two rule changes cleared the lane's plumbing this spring. Category I/II AIF gains are capital gains by statute from 1 April 2026, which retires the recharacterisation risk on every secondary transfer. And SEBI's March board let AIFs retain liquidation proceeds past fund life rather than fire-sale.28,82 One caution belongs on the record. There is no Competition Commission safe harbour for continuation vehicles; the Green Channel is a self-assessed declaration, not an exemption. Structure accordingly.83
| Vehicle | Size | What it proved |
|---|---|---|
| ChrysCapital NSE CV (Apr 24) | ~$700m | 6× to exiting LPs; the asset re-marked ~3× since |
| Multiples multi-asset CV (May 25) | $430m | India's largest; Hamilton Lane, HarbourVest bought |
| TR Capital / Lightbox CVs | $100–150m | The mid-market lane functions |
| 360 ONE secondaries fund | ₹4,000 cr target | Domestic capital entering the discount business |
Every GP holding a queue position has a live second bid. Price the IPO against it. The CV lane pays today, at a stated discount, with no validity cliff and no lock-up wall. The window's concession and the secondary's discount are converging; in the SBI MF and NSE pricing they already have. At that point the only reason to wait for the listing is vanity about the mark.
If you hold a 2019-vintage position, and by CRISIL's ladder your fund has returned half its paid-in capital, the window forces one decision, and it is arithmetic before it is judgment. Accept the clearing price now, roughly 30% under the carrying mark. Or hold for the mark, behind a $31bn lock-up wall, a refiling queue that runs to late 2027, and a DPI clock your LPs read quarterly.25,7 The arithmetic below is ours (a Gravitywell Research construction, assumptions in the open) and it is deliberately simple.
Read the curve as the seller's hurdle. Selling now realises 70 of every 100 of mark. Waiting eighteen months breaks even only if the exit then realises ~86 — that is, if the mark holds entirely and the concession narrows by more than half. And that is before counting the wall's pressure on prices, the 2027 refiling congestion, and the chance the fund crosses its term. The completed evidence sits on the wrong side of the line: PhonePe's ask fell ~25% during sixteen months of waiting; Zepto's marks fell ~35% over twelve.73,74 Against that, a sale priced today at the SBI MF-calibrated concession funds distributions your 0.5× DPI vintage can report this year — and the CV lane (§15) will warehouse whatever genuinely deserves more time.
| Path | Realises | When | PV @14%/yr | Carries |
|---|---|---|---|---|
| Take the window's price | ₹70 | now | ₹70 | DPI now; done |
| Wait, marks hold, concession halves | ₹85 | +18 mo | ₹69.9 | wall risk, term risk |
| Wait, marks fade as PhonePe's did | ₹55–65 | +18 mo | ₹45–53 | the observed path |
| Roll into a continuation vehicle | ₹75–80 typ. | now | ₹75–80 | for assets that earn the time |
All rows ○ Gravitywell Research construction; clearing price and CV pricing calibrated to §11 and §15 observed transactions; 14%/yr is an illustrative LP opportunity cost, not a measured rate.
Sell the strike, not the calendar. The only holdings that justify waiting are the ones a continuation vehicle would fund at a narrower discount than the window's — that is the market's own test of whether the asset deserves time. Everything else goes now, at the concession, while the window is paying cash for certainty.
The window's fate is three prints, not 238 filings — and the two biggest are structural opposites. One is the purest exit in the book. In the other, nobody is cashing out at all.
Everything this report argues converges on one deal. NSE's ₹30,000 crore offer is 100% offer-for-sale: roughly 6% of the company, ~20 sellers led by SBI, zero fresh capital. A float that small is legal only because March's SCRR amendment rewrote the minimum-offer slabs for mega-issuers.84,29 It would be India's largest IPO, past Hyundai's ₹27,870 crore, and it arrives priced to a band the peer can test: ₹4.7–6.0 lakh crore is 38–48× annualised first-quarter earnings against BSE's 42.5× forward, and 46–58× trailing against its ~54× — an 11–15% discount at the bottom of the band, an 8–13% premium at the top, depending on basis.84,32 The ten-year obstacle is paid off: the co-location saga settled for ₹1,491.21 crore, with the final ₹714.74 crore wired on 31 July, and the roadshow has met 30-plus global institutions.85 What it does not have, as of 13 August, is SEBI's final approval. The anchor event of India's exit window is itself contingent.86
The franchise needs no adjectives: ~93% of cash equities turnover, ~99.8% of equity futures, ~74.7% of options premium.87 Our Financial Operating System report (GWR-2026-IN-001, July 2026) built its core long on exactly this class of regulated market infrastructure, and NSE is that thesis at national scale. But hold the two facts this page cannot net. The franchise wins on share while losing on its regulatory lever: options are 60.22% of operating revenue, SEBI has been tightening derivatives mechanics all year, and FY26 EBITDA fell 12% under those measures.87,88 A monopoly with ~70% of operating revenue inside exchange-traded derivatives (options 60.22%, futures ~9%) is both the safest and the most exposed asset in the queue, and which dominates depends on the reader's horizon, so we price the ambiguity rather than resolve it: it is a large part of why the band sits at parity with the junior peer rather than at the premium the franchise's market share would otherwise command.
| Date | Event | What it removed |
|---|---|---|
| 2016–17 | First DRHP era; co-location case opens | — the decade's blocker begins |
| 30 Jan 2026 | SEBI no-objection certificate | Listing delinked from pending cases |
| 13 Mar 2026 | SCRR Amendment Rules | ~6% float made legal for mega-issuers |
| 17 Jun 2026 | DRHP filed: ₹30,000 cr, 100% OFS | Structure locked; ~20 sellers, SBI largest |
| 31 Jul 2026 | ₹1,491.21 cr settlement fully paid | The co-location liability |
| 13 Aug 2026 | SEBI final approval | Still pending — the contingency |
Anchor the window trade here, at or below the midband, sized to the contingency. At that entry the OFS overhang and the options-revenue exposure come priced in at the peer's own rating; above the midband they come at a premium. A September approval makes this the print that re-opens the market, and a slip past the Navratri–Diwali window is the single event most likely to break the H2 numbers in this report. Both legs of that bet are stated in the register.
Jio Platforms is the queue's structural opposite, and the contrast is deliberate. The DRHP filed on 19 June is a 100% fresh issue of up to 27 crore shares — ₹30,000–40,000 crore into the company, zero offer-for-sale. Reliance dropped the selling-shareholder route in March.5 Meta, Google, PIF, KKR, Vista and Silver Lake, the 2020 vintage that paid in at a ~$58–65bn valuation, stay put. Announcing at the AGM that nobody is cashing out is Mukesh Ambani pricing the signal itself.5,89 The deal tests something different from NSE. Will the domestic bid fund large-cap growth at ~47× earnings, in the size bucket July's flows were leaving?34,15
The operating case is strong and priced accordingly: FY26 revenue of ₹1,46,885 crore up 15%, EBITDA margin at 51.9%, profit up 15% to ₹30,049 crore.5 The honest complication is disclosure. The DRHP does not break out telecom from the digital businesses the multiple is being asked to pay for; effectively all reported revenue is Reliance Jio Infocomm. The buyer of the "platform" story is underwriting a segment note that does not yet exist.5 SEBI sought clarifications on 25 June; the observation letter is pending and the listing is guided at August–December. A valuation above Airtel's entire market value, for a business whose non-telecom leg is undisclosed, sold into the bucket with net outflows — that is a genuinely open print.90
| NSE | Jio Platforms | |
|---|---|---|
| Structure | 100% OFS — pure exit | 100% fresh — no seller |
| Who gets the money | ~20 selling shareholders | The company |
| What the print tests | The exit bid: will domestic money buy sellers out at a discount? | The growth bid: will it fund expansion at ~47×? |
| Valuation vs comparable | 38–48× ann. vs BSE 42.5× fwd — parity band | ~$123–154bn vs Airtel ~$127bn — premium in the top half |
| Approval status (13 Aug) | Pending — settlement paid | Pending — clarifications 25 Jun |
Jio is the window's live experiment, and the watch-items are dated: the observation letter first, then the final price against the $120–140bn consensus. A strong clear at the top of the band is the best evidence available that the FPI turn has depth — the signal to re-rate the whole H2 capacity. A struggle at the bottom says even the growth bid now demands the window's terms. The print informs either way; the ask does not need to be paid to learn from it.
One mega-deal has already crossed the window, and it calibrates everything behind it. SBI Funds Management (₹9,813 crore, listed 21 July) was subscribed 41.73× with the institutional book at 140.11×, total bids of ~₹3 trillion, India's fourth-most-bid IPO ever.6 It priced at 38.12× FY26 earnings against a 41.64× peer average, listed up 6.85% at ₹613.30, touched ₹625, and closed its first day at ₹610.21,91 The grey market had promised 17%.6 Every number in that sequence teaches the same lesson: demand is enormous, and it is done paying for it.
The rest of the live mega-book reads off the same ruler. Manipal Health completed India's largest healthcare IPO at ₹9,275 crore, after trimming the ask, with fresh capital dominating the structure.72 OYO's parent PRISM holds approval for a ₹6,650 crore, 100%-fresh issue with no offer-for-sale leg, a structure choice that reads as sponsors declining today's strike.92 Below the megas, the first half's biggest prints (Clean Max Enviro at ₹3,080 crore, Fractal Analytics at ₹2,834 crore) cleared without drama at disciplined pricing.36 The market is open. It is simply quoting one price, and it is not 2021's.
| Deal | Size | Structure | Status · 13 Aug | The read |
|---|---|---|---|---|
| SBI Funds Mgmt | ₹9,813 cr | OFS-led | Listed +6.85% | The calibration: cleared at a peer discount |
| NSE | ₹30,000 cr | 100% OFS | Approval pending | The anchor; §17 |
| Jio Platforms | ₹30–40,000 cr | 100% fresh | Observations pending | The growth test; §18 |
| Manipal Health | ₹9,275 cr | fresh-led | Completed | Cut the ask, cleared the print |
| OYO (PRISM) | ₹6,650 cr | 100% fresh | Approved, unlaunched | Sponsors declining the strike |
Use SBI MF as the window's pricing ruler: a dominant franchise, maximal demand, and still a single-digit pop off a below-peer multiple. Model every remaining mega at that discipline, NSE included, and treat any bank case built on double-digit listing pops as a 2025 artifact. The ruler also cuts the other way: deals priced to it are clearing, so execution risk in the window is pricing risk, nothing darker.
The deals that stood down are the window's most informative prints, because a pause is a price signal with a date on it. PhonePe took SEBI approval on 20 January, watched its target valuation get marked from $12–15bn to $9.5–10.5bn, parity with listed Paytm on a price-to-sales basis, and postponed on 16 March rather than accept it.3,73,93 Zepto's marks ran $7bn in October to $4–5bn by its 30 July pause, with roadshow bids implying as little as ~$2.3bn (a 68% valuation cut) along the way.74,94 Both companies looked at the clearing price this report keeps measuring, and both said no. That no is the strike, observed.
The clocks are the pressure the pauses cannot escape. PhonePe's SEBI approval has a twelve-month shelf life, launch by roughly 20 January 2027 or refile, and its regulatory overhang compounds: NPCI's proposed 30% UPI market-share cap, deferred to December 2026, sits over a business holding ~46–48% share.3,73 Zepto's runway is its own clock: ₹5,681 crore of cash against a ₹5,041 crore adjusted-EBITDA burn is thirteen months at the FY26 rate, in a market where Blinkit already prints positive EBITDA.95,96 Behind these two stand thirty-plus more approved-and-paused issuers (Indira IVF, Ajay Poly, Asha Resins among them) each holding an option expiring against the same 30 September validity cliff.24
| Issuer | Last ask → mark | Why it paused | The clock runs out |
|---|---|---|---|
| PhonePe | $12–15bn → $9.5–10.5bn | Parity with Paytm ≠ the premium story | ~20 Jan 2027 |
| Zepto | $7bn → $4–5bn | 68% issue cut; anchor pushback | Nov 2027 (refile) |
| OYO (PRISM) | ₹6,650 cr approved | Sponsors declining the strike | approval shelf life |
| 30+ approved others | — | Waiting on price | 30 Sep 2026 cliff |
Read every pause as a quote. The sellers' strike sits 20–40% above the market's mark across the cohort. The clocks (January for PhonePe, November 2027 for Zepto, September for the long tail) all force the choice this report priced in §16. Converge on the bid, or refile into a 2027 queue that already holds the wall's paper. For the allocator, the paused book is next year's supply at this year's revealed strike; underwrite 2027 accordingly.
The state built this window and competes inside it; the tax code just rewrote the seller's arithmetic. What remains is to price the three cases, name what the thesis hinges on, and give each seat its call.
Read the year's rulebook as a single policy: keep the exit window open. The March SCRR amendment replaced four minimum-offer slabs with six. An issuer above ₹5 lakh crore of capital can now list ~6% and take years to reach full public float. That is the rule that makes NSE listable, and it extends retroactively to companies already listed.29 April brought the twin forbearances. Penal action on minimum-public-shareholding breaches froze to 30 September; observation letters expiring mid-year extended to the same date; fresh-issue resizing widened to ±50% without refiling. Pointedly, none of that applies to the offer-for-sale leg — the state will flex how much capital a company raises, not how much a sponsor sells.10,33 November's anchor reform had already reserved 40% of every anchor book for domestic institutions.65
| Date | Rule | What it does for the window |
|---|---|---|
| 30 Nov 2025 | Anchor book: 40% domestic reservation | Locks the domestic price-setter into every book |
| 13 Mar 2026 | SCRR: six-tier minimum-offer slabs | Makes ~6% float mega-listings legal (NSE, Jio) |
| 7 Apr 2026 | MPS penal freeze + validity extension | Buys the queue six months — the 30 Sep cliff |
| 15 Apr 2026 | ±50% fresh-issue resize, no refiling | The Indo-MIM path: cut to clear (OFS excluded) |
| 1 Aug 2026 | Open-market buybacks reinstated | A competing use of the same corporate cash |
| FY27 budget | Disinvestment target ₹80,000 cr (+first raise in 6 yrs) | ~₹53,000 cr still to land on the same bid |
The competing supply is the state's own. The FY27 disinvestment and monetisation target of ₹80,000 crore was raised for the first time in six years. It was about a third done by late July, leaving ~₹53,000 crore of PSU paper for the same second half. CY26's offer-for-sale tape is already two-thirds government stake sales.30,37 If you sequence the disinvestment calendar, the arithmetic in this report is your planning constraint. The validity cliff you created concentrates the private queue into September–December. The anchor reservation routes it all to the same domestic institutions. Every DIPAM tranche placed inside that corridor bids against NSE and Jio for the marginal large-cap rupee. Sell the state's paper in January, or watch it cheapen everyone's — including the state's.
The rules have converted a diffuse pipeline into a scheduled event: one quarter, one buyer, everything priced off two prints. That is manageable if the sequencing holds and self-inflicted if it does not. For the investor, the practical instruction is timing. The richest concessions will be quoted in the crush weeks either side of the cliff, when the most sellers meet the same finite bid at once.
The quietest rule change of the year may move the most money. From assessment year 2026-27, securities held by Category I and II alternative investment funds are capital assets by statute. Gains on transfer are capital gains, not business income. Carried interest follows the same character.28 For a decade, the recharacterisation question (capital gains at 12.5%, or business income at the fund's marginal rate) sat as litigation risk on every domestic PE exit and every secondary transfer. It resolved on 1 April, three months before the exit cohort this report prices reached its window. That timing is §21's policy again, executed through the Finance Act.
| Seller | Route | Rate on gains | What changed |
|---|---|---|---|
| Cat I/II AIF (domestic PE/VC) | IPO OFS / secondary sale | LTCG 12.5% | Capital-gains character now statutory28 |
| — carried interest | distribution waterfall | follows fund character | Recharacterisation risk retired28 |
| Any holder, listed equity | market sale ≥12 mo | 12.5% above ₹1.25 lakh | Unchanged by Budget 202697 |
| Any holder, unlisted shares | pre-IPO secondary | 12.5% at 24-mo holding | No indexation; period unchanged97 |
| FPI, government securities | interest + capital gains | exempt | Full exemption from 1 Apr 202698 |
| Derivatives traders (context) | STT on options/futures | 0.15% / 0.05% | Raised — the one tightening97 |
The buy side's plumbing eased in parallel. SWAGAT-FI went live on 1 June, giving trusted foreign investors ten-year registration and ten-year KYC cycles. IRDAI lifted insurer limits on AIF and unlisted exposure in July.99,100,101 Two cautions balance the ledger. The February budget raised the securities transaction tax on derivatives. That lifts the cost of the options complex that alone generates 60% of NSE's revenue — the window's anchor asset is partly taxed by the window's tax code. On the international layer, India has notified no Pillar Two minimum-tax rules. Cross-border fund structures planning around GIFT City's holidays are planning against legislation that does not yet exist. We treat that as structuring risk, not settled law.102
For a domestic GP, the after-tax case for selling into this window improved by the width of a lawsuit: 12.5% certain beats a recharacterisation dispute at 30-plus percent uncertain, and it compounds the §16 arithmetic toward selling now. For structures routed offshore, the unfinished Pillar Two picture argues for keeping optionality. Read the next Finance Bill before assuming the holiday math survives it.
The cases below price the H2 window in dollars of completed issuance, from anchors this report has already established. Those anchors: $5.78bn cleared to 3 August; a ~$6bn all-channel clearance estimate for July–December; two mega-deals worth ~$7bn combined if both price; and a validity cliff that forces the long tail to launch or lapse by 30 September. The windows are defined once: H2 bands run July–December, and the CY26 rows add them to a Jan–Jun first half of ~$4.7bn (○, M.1), netting the ~$1.1bn of July prints (led by SBI MF's ₹9,813 crore) that already sit on the tape.2,7 Probabilities are the desk's assignment, not a model output, and the assumptions sit in the open in the table.
| Case | H2 / CY26 | Requires | CY26 vs CY25 |
|---|---|---|---|
| Bear | $4–6bn / $9–11bn | NSE approval slips past the Navratri–Diwali window; Jio's review extends; the 30 Sep cliff lapses the tail; the FPI turn fails on a Fed hike | −51 to −60% |
| Base | $6–9bn / $11–14bn | NSE prices in Q4 at or below the ₹5.0–5.25 lakh cr talk; Jio at the $120–140bn consensus; OYO clears; the mid-market keeps cutting 30–45%. PhonePe and Zepto stay 2027 events | −37 to −51% |
| Bull | $12–15bn / $17–20bn | Both megas top-of-band inside Q4 (~$7bn); the FPI turn extends past its two-month life with the primary-book bias intact; PhonePe relaunches before its January expiry | −11 to −24% |
The twelve-month view runs on the wall rather than the window. In the bear mechanism, the H2 cohort's 90-day anchor cliffs land in Q1 2027 and its twelve-month lock-ups in H2 2027 (the machinery that put 19 of 52 FY22 listings below issue as the 2021 cohort's lock-ups rolled off40) with Jio's untouched 2020 cap table (Meta, Google, KKR, PIF, Silver Lake, Mubadala) as the open 2027 supply question.5 The base-rate caution for index readers: in 2022 the repricing landed on the primary market and its own cohort (proceeds halved; 19 of 52 FY22 listings broke issue) while the index escaped comparatively lightly.40
Underwrite the base case and structure for the spread. A $6–9bn half is a functioning window at the concession. The year still lands 37–51% below 2025 — which is why every "record year" narrative you will read in September is already wrong on the calendar arithmetic. The bull case is a real possibility with a named driver. Note carefully: it is still below the consensus it would be reported as vindicating.
A quantified thesis owes the reader its quantified failure modes. Four variables carry this one, and their swings (desk constructions, calibrated to the sourced anchors) differ by an order of magnitude, which is itself the finding: watch the two big ones, monitor the two small ones.
| Variable | The mechanism | Swing | Watch (dated) |
|---|---|---|---|
| The FPI turn | Two-thirds of July's foreign buying was primary allotment; sustained reversal roughly doubles book depth for the megas | ±$4–6bn | NSDL monthly net, Aug & Sep prints |
| NSE's clock | ₹30,000 cr in one print, plus the reopening effect on the tail behind it | ±$3.5bn | SEBI observation letter; RHP + band |
| The SIP floor | Every ₹1,000 cr/month of gross SIP ≈ $1.3bn/yr of standing bid; stoppage >100% says the floor is thinner than its headline | −$2.5bn | AMFI gross SIP <₹31,000 cr for 2 months |
| State sequencing | ~₹53,000 cr of PSU paper inside the corridor competes; deferred to Q4FY27 it releases the bid | ±$2bn | DIPAM OFS announcements, Q3FY27 |
Two hinges we monitored and set aside deserve their line. A tariff re-basing under a surviving statute would reopen the foreign exit — it is in the risk register with its trigger, but it is an event, not a dial. And a genuine banking-liquidity turn (the RBI easing into a slowing tape) would lift the whole bid — we found no evidence of it in the data through August, and assuming it would be hope, not analysis.
The monitoring stack, in order: the NSDL monthly print, the NSE observation letter, AMFI's gross SIP line, the DIPAM calendar. The first two decide the scenario by October; the second two decide whether the base case's floor holds through it. Everything else in this report is positioning; these four numbers are the verdict.
One market, five seats, five different uses of the same clearing price. Each call below carries its number and its exit condition; the register behind them is the same evidence the rest of this report walked through.
| Seat | The call | The price | Changes if |
|---|---|---|---|
| Allocator, public | Anchor in NSE at the band; treat SBI MF as the pricing ruler; skip premium-to-peer prints | ≤~43× ann. (midband); ≤peer avg elsewhere | NSE slips past Q4 → hold cash for the 2027 repricing |
| Allocator, timing | Bid the crush weeks around 30 Sep, when the most sellers meet the finite bid | the widest concessions of H2 | FPI turn extends → concessions narrow, move early |
| GP | Distribute 2015–19 vintages into the window; CV-lane anything that genuinely earns time | accept ~30% to marks (§16 curve) | Marks recover two quarters running → re-run the curve |
| LP | Press for DPI over IRR; benchmark every hold against the secondary discount | 0.5× DPI is the status quo you are funding | Distributions beat the ladder → relax the pressure |
| Issuer | Resize down and clear (the Indo-MIM path); do not park behind the cliff | −30–45% quantum, per the cleared cohort | Bull case confirms → the resize flexibility works upward too |
| Policymaker | Sequence DIPAM tranches out of the Sep–Dec corridor; hold the cliff date firm | ~₹53,000 cr rescheduled ≈ +$2bn of private capacity | Private queue lapses en masse → the corridor was too narrow |
The board's common thread is symmetry of payment. The allocator is paid in concessions for providing certainty. The GP is paid in DPI for surrendering the 2021 mark. The issuer is paid in execution for cutting the ask, and the state is paid in a functioning exit market for sequencing its own selling. The seat that refuses (the sponsor holding for the old mark, the fund waiting for the old pop) pays everyone else's bill in time, and time is what the cliffs, the wall and the fund clocks are all denominated in.
Five observables, all dated, any two of which would force an edition revision: (1) NSDL prints positive for August AND September; (2) NSE lists inside Q4 at the top of its band with a double-digit pop; (3) AMFI gross SIP holds above ₹32,000 crore through the crush; (4) the paused cohort relaunches at its old asks rather than the marked ones; (5) CY26 full-year proceeds exceed $20bn — above even the house bull band. Each would say the bid is deeper, or the concession narrower, than this report prices.
The exit window is open, priced, and on a clock. Take it as it is (at the concession, in the anchors, sized to the crush) or state plainly which of the five observables above you are betting against. That is the whole trade.
Each risk carries its mechanism and a checkable trigger. Severity is the desk's assignment. The register spans the taxonomy (commercial, operational, financial, regulatory and tax, geopolitical, technological) not only the market risks.
The thesis, held to the end: the exit window is real, the queue's headline is an option book, and the clearing price sits 25–40% below the sellers' last marks73,74 (○, M.1), tested deal by deal on the tape — the transacted, peer-basis concession runs −8% to −24%. The positions that follow are stated as positions — with the price, and with what retires them.
NSE at or below the midband is the window's defining trade: a ~93%-share franchise offered at 38–48× annualised earnings, level with BSE's 42.5× forward, the offer-for-sale overhang and the options-revenue exposure carried at the peer's own rating.87,32 The position retires if the approval slips into 2027 — then the same discipline applies to whatever reprices in its absence.
For the 2015–19 vintages at 0.5–0.8× DPI, distribution now at the clearing price dominates holding for marks the price no longer supports; the §16 indifference curve is the desk's standing answer to "wait for better." The continuation-vehicle lane (tax-clean since April, precedented at 6× to exiting LPs) is the instrument for the assets that genuinely earn more time.25,28,27
SBI Funds Management set the window's terms: maximal demand, below-peer pricing, single-digit pop.21,6 Any deal asking a premium to its listed shadow, Jio included, is a test of the growth bid to be watched, not pre-paid; a strong Jio print is the signal to re-rate H2 capacity upward, and the only one we would accept.
The H2 cohort's anchor cliffs land in Q1 2027, its twelve-month lock-ups in H2 2027, and the paused cohort refiles into the same calendar at strikes the market has already marked down.7,74 Reserve capacity for that supply; it will be cheaper than this year's, for the same assets.
A last word on the asymmetry, because it is the finding a skim would miss. Every actor in this market is behaving as if the clearing price is what this report says it is: the regulator extending approvals, the issuer cutting quantum, the GP rolling assets into continuation vehicles, the state raising its own selling target. The only prices still set above it are the marks in private books. One side of that gap moves every time a deal prints, and it has not once, in eighteen months of prints, been the market's side.
Coverage. In scope: India's H2CY26 primary-market pipeline and its conversion; FPI/DII flows and ownership; PE/VC exits, DPI and the secondary lane; the five largest live or recently completed issuers; the 2025–26 regulatory and tax changes bearing on issuance. Out of scope: SME-board issuance (except where a source's total includes it, noted inline), bond and hybrid supply, and company-level equity research on any issuer — pricing references are market-level. Data is thin, and flagged as such, on: LIC and NPS annual deployment (only quarterly or stock figures are published), per-episode supply-as-%-of-market-cap history, and unlisted-market price marks (indicative dealer quotes).
Source policy. Every figure carries a numbered register entry. Tier ● is reserved for regulators, exchanges, depositories and official statistics read directly (SEBI, NSDL, AMFI, RBI, MoF/PIB, Federal Reserve, BLS, EIA) and peer-reviewed academic papers read directly; consultancy datasets (EY, Bain, CRISIL, Prime Database compilations) and all press-mediated figures are ◐ even where the press quotes a primary; desk constructions are ○ and listed below. Where sources disagree we show the spread and name the anchor; four conflicts remain unresolved and are carried as such (Temasek routing; the pipeline's size; the tariff's legal basis; the H2 supply figure — Bloomberg's ~$6bn completed-issuance basis against JPMorgan's ~$40bn intent framing, a definitional 7× spread).
| Construction | Method | Sensitivity |
|---|---|---|
| FX basis ₹95/USD | Back-solved from the NSE $55.1bn ↔ ₹5.25 lakh cr dual print; Aug 2026 spot ~95.5 | ±2% across 2026 prints |
| NSE 38–48× / 46–58× | Banker band ₹4.7–6.0 lakh cr ÷ annualised Q1FY27 PAT ₹12,480 cr / FY26 PAT ₹10,302 cr | Q1 annualisation assumes no seasonality |
| §16 indifference curve | PV of realisation paths at an illustrative 14%/yr LP opportunity cost from a 70%-of-mark clearing price | ±2pp on the rate moves the 18-mo hurdle ±3pp |
| §24 sensitivity swings | Scaled from cited anchors: NSDL monthly range, deal sizes, SIP arithmetic, DIPAM residual | Order-of-magnitude guides, not point estimates |
| Wall realisation ~9% | H1 open-market PE/VC exits $4.1bn ÷ the Jan–Apr $45bn expiry tranche; applied to the Jun–Sep $31bn ⇒ ~$3bn of probable selling | Tranche mix shifts the rate; promoter share unsold |
| CY26 bridge (H1 ≈$4.7bn) | Tape $5.78bn to 3 Aug minus ~$1.1bn of July prints (SBI MF ₹9,813 cr; Manipal ₹9,275 cr booked 24 Jul but listed 5 Aug, so counted in H2 on the listing-date basis) = Jan–Jun ≈$4.7bn; CY26 rows = H1 + the Jul–Dec H2 band, no overlap | ±$1bn on the listing-date basis (Manipal priced late Jul, listed 5 Aug) |
| ≤~13% conversion ceiling | All-channel ~$6bn (IPOs + stake sales + placements) ÷ IPO-only ₹4.72trn — a scope-generous ceiling; the IPO-only rate is lower. JPMorgan's ~$40bn intent framing carried as a conflict | Definitional; see Methodology conflicts |
| Headline "25–40% concession" | Observed repricings of the paused cohort, mid-to-mid: PhonePe ask ~$13.5bn → ~$10bn mark (−26%); Zepto $7bn → ~$4.5bn (−36%); §16 clearing-price construction (−30%). The 25–40% band brackets these with the endpoint spreads disclosed in §20. Denominator = the sponsor's last ask or carrying mark. n=2, both non-transactions; SBI MF (−8/−24% to peers) is the nearest completed calibration, on the other denominator | Bounds move with the two named paths |
| Scenario probabilities | Desk assignment on the §23 assumption sets | Not model outputs |
| Wall at ~0.6% of market cap | $31bn ÷ $5,095bn India market capitalisation (Jul 2026) | Sizing only; not a net-supply threshold |
| BSE forward ~42.5× | Third-party consensus screen (register 107); not published on the Screener page; consistent with a desk annualisation of BSE Q1FY27 earnings (~42×) | ±1× across providers |
| 2024 secondary volume ~$162bn | Derived from the 2025 level and its published +48% growth | Rounding only |
Known non-comparabilities, held apart in the text. 2021 vs 2026 subscription multiples (the RBI ₹1 crore NBFC funding cap changed what a multiple measures); DII ownership at 19.15% (PRIME all-DII) vs 17% (SEBI universe) — different denominators, never blended; mainboard-only vs all-scope fundraising series (Prime ₹22,555 cr vs ₹56,322 cr Jan–May); our supply/market-cap ratio (~1.6–1.9% ○) vs Jefferies' 1% threshold, which is defined on net supply — we do not publish the comparison as a breach.
Point in time. Data as of 13 August 2026 unless dated otherwise; the NSDL August figure is a partial month (to 7 August) and marked so. Nothing restates a prior edition (first edition). Errata will be dated and appended; the permalink serves the current version. This report builds no composite index or rating; the OECD/JRC and IOSCO frameworks the desk applies to its index family are therefore not invoked here. Confidence is Medium by the house scale: the thesis rests on reconciled multi-source figures and filed regulatory facts, but key market figures are press-mediated (◐), four register-level conflicts are carried, two bearing directly on load-bearing numbers, and the central events remain regulator-contingent.
Every in-text superscript resolves here, numbered in order of first citation. ● primary (regulator/exchange/official statistic or peer-reviewed academic paper, read directly) · ◐ secondary or press-mediated. Figures marked to the dates shown.
50 numbered exhibits. Underlying series for the data-bearing exhibits: Data Appendix (p40–41) and the machine-readable pack india-exit-window-data.csv.
The series behind the data-bearing exhibits, so the analysis is reproducible. Tier: ● filed/official · ◐ modeled/derived · ○ estimate. Source # per the register.
| Month | ₹ crore | Tier | Src |
|---|---|---|---|
| Jan | −35,962 | ● | 57 |
| Feb | +22,615 | ● | 57 |
| Mar | −1,17,775 | ● | 57 |
| Apr | −60,847 | ● | 57 |
| May | −32,963 | ● | 57 |
| Jun | −49,340 | ● | 57 |
| Jul | +20,199 | ◐ | 9 |
| Aug 1–7 | +12,921 | ◐ | 9 |
| H1 total | −2,74,272 | ◐ | 57 |
| Stage | Value | Tier | Src |
|---|---|---|---|
| Announced, 238 cos | ₹4.72 trn | ◐ | 1 |
| SEBI-approved, 174 | ₹2.77 trn | ◐ | 1 |
| Approved (Uniqus, 157) | ₹2.38 L cr | ◐ | 36 |
| Approved (Prime, Mar, 144) | ₹1.75 L cr | ◐ | 1 |
| H1 cleared, 27 IPOs | ₹22,555 cr | ◐ | 36 |
| Realistic H2 | ~$6bn | ◐ | 7 |
| Lock-in expiries Jun–Sep 26 | ~$31bn | ◐ | 7 |
| FY | Median % | Average % | Tier · Src |
|---|---|---|---|
| FY23 | 7.9 | — | ◐ · 8 |
| FY24 | 15.4 | — | ◐ · 8 |
| FY25 | 21.2 | 28–30 | ◐ · 8, 44 |
| FY26 | 3.0 | ~8 | ◐ · 8, 44 |
| Month 2026 | ₹ crore | Tier | Src |
|---|---|---|---|
| Apr | 38,426 | ◐ | 16 |
| May | 22,898 | ◐ | 16 |
| Jun | 28,973 | ◐ | 15 |
| Jul | 24,697 | ◐ | 15 |
| Jul — smallcap | +7,767 | ◐ | 15 |
| Jul — midcap | +6,192 | ◐ | 15 |
| Jul — largecap | −1,322 | ◐ | 15 |
| Gross SIP, Jul | 31,961 | ◐ | 15 |
| Date | FPI | DII | MF | Src |
|---|---|---|---|---|
| Mar 2015 | ~23 | ~13 | — | 14 ◐ |
| Mar 2025 | 17.22 | 17.62 | — | 63 ◐ |
| Mar 2026 | 16.12 | 19.24 | — | 14 ◐ |
| Jun 2026 | 15.88 | 19.15 | 11.58 | 14 ◐ |
| Vintage | DPI × | Vintage | DPI × |
|---|---|---|---|
| 2014 | 1.2 | 2019 | 0.5 |
| 2015 | 0.8 | 2020 | 0.5 |
| 2016 | 0.8 | 2021 | 0.5 |
| 2017 | 0.7 | 2022 | 0.1 |
| 2018 | 0.6 | 2023 | 0.1 |
A.6 tier ◐ throughout · src 25 (CRISIL benchmarks, Sep 2024 data). Machine-readable copy of all appendix series: india-exit-window-data.csv. The maintained panel behind them is minted at data/india-exit-window-panel/ (long-format CSV + README: definitions, tiers, vintage policy, update cadence) — the dataset outlives this report.
| Year | US$ bn | Count | Tier · Src |
|---|---|---|---|
| 2021 | 39.6 | 279 | ◐ · 22 |
| 2022 | 18.3 | — | ◐ · 22 |
| 2023 | 24.9 | 305 | ◐ · 22 |
| 2024 | 28.2 | 285 | ◐ · 22 |
| 2025 | 32.9 | 257 | ◐ · 22 |
| 2025 ex-Schneider | ~26.5 | 256 | ◐ · 24 |
| H1 2026 | 9.4 | 95 | ◐ · 23 |
| PE-backed IPO exits, H1 26 | 0.80 | 12 | ◐ · 23 |
| Period | US$ bn | GP-led | Tier · Src |
|---|---|---|---|
| 2024 (derived) | ~162 | — | ○ · 80 |
| 2025 | 240 | 115 (48%) | ◐ · 80 |
| H1 2026 | 121 | 65 (54%) | ◐ · 26 |
| Anchor | Value | Tier · Src |
|---|---|---|
| CY26 YTD proceeds (3 Aug) | $5.78bn | ◐ · 2 |
| CY25 full year | $22.36bn | ◐ · 2 |
| Sell-side CY26E (Dec 25 calls) | $20–25bn | ◐ · 35 |
| NSE + Jio if both price | ~$7bn | ◐ · 84, 5 |
| Metric | Value | Tier · Src |
|---|---|---|
| NSE issue / structure | ₹30,000 cr · 100% OFS | ◐ · 84 |
| NSE band (banker talk) | ₹4.7–6.0 L cr | ◐ · 104 |
| NSE FY26 PAT / Q1 ann. | ₹10,302 / 12,480 cr | ◐ · 103, 31 |
| BSE P/E, trailing / forward | ~54× / 42.5× | ◐ · 32, 107 |
| Jio raise / structure | ₹30–40,000 cr · fresh | ◐ · 5 |
| Jio scenarios (GS) | $98 / 123 / 154bn | ◐ · 89 |
| Jio FY26 PAT | ₹30,049 cr | ◐ · 5 |
| SBI MF price / peers | 38.12× / 41.64× | ◐ · 21 |
| SBI MF subscription | 41.73× (QIB 140×) | ◐ · 6 |
| SBI MF listing | +6.85% vs GMP 17% | ◐ · 6, 91 |
| PhonePe valuation path | $12–15bn → $9.5–10.5bn | ◐ · 73 |
| Zepto valuation path | $7bn → $4–5bn | ◐ · 74, 94 |
| Name | Issue ₹ | vs issue | Tier · Src |
|---|---|---|---|
| Eternal (Zomato) | 76 | ≈ +298% | ○ · 46 |
| Nykaa (adj.) | 187.5 | ≈ +78% | ○ · 46 |
| PB Fintech | 980 | ≈ +63% | ○ · 46 |
| Paytm | 2,150 | ≈ −25% | ○ · 46 |
A.11 marks are indicative platform quotes (snippet-grade); Paytm carries a conflicting ₹1,343 print. Direction robust, levels approximate — flagged in the methodology.
The Gravitywell Research Capital Markets desk responsible for this report certifies that the views expressed accurately reflect its independent judgement about the subjects and securities discussed, and that no part of its compensation was, is, or will be directly or indirectly tied to the specific recommendations or views expressed herein.
As of the publication date, Gravitywell Research and its analysts do not hold positions in the securities or assets discussed. Gravitywell Research has no advisory, banking, or commercial relationship with the entities named. This report was not commissioned or reviewed by any issuer named.
Conviction — High · Medium · Low, set by coverage depth, source tier and how stress-tested the thesis is (matches the site's confidence field). Stance — Constructive · Neutral · Cautious, the direction of the house view on the opportunity. This report: Medium conviction · Constructive — constructive on the window at its clearing price, with the conviction cap reflecting press-tier market figures and two regulator-contingent anchor events.
Prepared for readers who allocate capital, underwrite risk, or set policy across public and private markets. Not for general retail distribution, nor for readers who lack the expertise to assess the assumptions. Intended recipients may not redistribute without attribution. Availability of this research in some jurisdictions may be restricted; recipients are responsible for their local rules. Independent research; not regulated ratings, not investment advice, not an offer or solicitation.
Material errors are corrected in a dated erratum appended to this report and noted in the next edition; the permalink serves the current version. Permalink: gravitywellresearch.xyz/research/india-exit-window. Desk contact: [email protected]. Version: GWR-2026-IN-003 · v1.1 · as of 14 Aug 2026. Cover art is produced by the Gravitywell art desk and credited on the cover.108 v1.1 (14 Aug 2026) adds the Consensus & the Variant page (p06) and the full-bleed cover, applies a readability pass, and restates the CY26 scenario aggregates on a July-netted bridge (base $12–15bn → $11–14bn, with bear/bull and the y/y percentages moving accordingly); the H2 bands and all underlying source figures are unchanged. The report's falsifiable calls are logged to the desk's public calls register and will be scored at maturity.
First edition on this topic — no prior Gravitywell Research call on the Indian primary market to score. The desk's prior published research is cited in the body where it bears on the argument (GWR-2026-IN-001, §17); its calls will be scored when their horizons mature.