Gravitywell ResearchInstitutional Investment Intelligence
India  /  Digital Public Infrastructure & Fintech

India’s Financial
Operating System.

The world’s most ambitious digital public infrastructure is maturing into a programmable financial OS. This report sizes the market, scores its attractiveness, maps where the rent accrues, weighs the citizen, and lays out the trade.

$2.1tn
Fintech TAM
· 2030E
757m/d
UPI transactions/day
· record, Jun 2026
41–84%
EBITDA at the listed
infra toll-booths
~₹10k cr
Est. annual value of
a 0.3% UPI MDR
GWR-2026-IN-001 · v1.4  ·  July 2026  ·  gravitywell.io Institutional · For Professional Investors
Gravitywell Research: India’s Financial Operating SystemContents
Inside This Report

CONTENTSMarket, Rent, Risk & the Trade

Front Matter
00Executive Summary03
—Investor Quick Read06
Part I · The Market
01Market Sizing: TAM / SAM / SOM07
02Industry Attractiveness: Five Forces08
03SWOT & PESTLE09
Part II · The Stack
04The 7-Layer Financial Stack10
05UPI at Global Scale11
Part III · Where the Rent Accrues
06The Profit-Pool Map & Its ₹ Size12
07The Listed Toll-Booths: Comps13
08Valuation & What’s Priced In14
08·iiLiquidity, Float & Fair Value15
09Account Aggregator & the Data Market16
10Value-Chain Economics & GTM17
Part IV · Capital, Funding & Exits
11Private-Market Map & Unit Economics18
12Funding, M&A, Exits & Returns19
13Scenarios (probability-weighted)20
Part V · Risk, Regulation & the Citizen
14The Regulatory Calendar21
15Credit Risk & the Unsecured Tail22
16The Citizen: Inclusion, Fraud & Exclusion23
17The Customer: Demand-Side, NPS & Pricing24
Part VI · Forces, Impact & the Trade
18International Comparables: PIX · Ant · PSD225
19Structural Forces: Banks, AI & Systemic26
20ESG, Economic Impact & Employment27
21Technology Roadmap & Disruptive Horizon28
22The Trade: Positioning & Catalysts29
Appendices
—Methodology & Data-Quality30
—Change Log & Data Vintage31
—Source Register · Glossary32
The Thesis

India built its rails as public goods, priced at zero margin. The rent migrated to regulated, listed toll-booths at 41–84% EBITDA: richly priced. The new data rent is being ring-fenced by DPDP before it forms. Own the moat; price the multiple; underwrite the risk.

Point-in-Time & Grading

Data as of 1 July 2026. FY26 = year ended Mar 2026. USD at ₹94.7/$ (RBI ref ~₹94.4, 29 Jun 2026). Every figure carries a data-quality grade A audited / B primary-estimated / C modelled (§Methodology). Not investment advice.

Gravitywell Research · Confidential02 / 33
00: Executive SummaryIndia’s Financial Operating System
Executive Summary

00The Rails Are Public. The Rent Is Private.

India has assembled the most complete stack of digital public infrastructure on earth: a $2.1tn-TAM financial OS from a 1.4-billion-person identity layer up through payments, consent-data, credit, commerce and capital markets.1 Each rail was engineered as a low-cost public utility. The consequence for an allocator is counter-intuitive: the value is not in the rails: it is in the regulated, mostly-listed choke-points the rails feed, and in who is permitted to monetise the data they generate.

$2.1tn
Fintech TAM 2030E (§01)
~84%
Peak EBITDA at the listed infra toll-booth (IEX)
42–64x
P/E at which the infra compounders trade
~0%
Margin on the zero-MDR UPI payment rail
India Fintech, At a GlanceDASHBOARD
Market

~$47bn revenue (2025) → $190bn (2030E, ~35% CAGR). 7,500+ firms · 30 unicorns · $51.1bn cumulative VC.13,41

Scale

UPI 757m txns/day · 89% account ownership · 474m AA consents · ULI ₹1.75L cr disbursed.2,10

Macro

GDP $3.9tn · real growth 7.6% · repo 5.25% · CPI 3.93% · INR 94.7 · DPI ~2% of GDP.26,42

Where the Margin Lives, VisualizedFIG. 0.1
EBITDA margin by layer: sourced to FY26 filings of the listed infrastructure names. The rails the public touches earn nothing; the regulated infrastructure behind them earns most.
02040 6080 IEX~84% MCX73% CDSL51% CAMS45% Infra SaaS~25% UPI apps~0% LISTED TOLL-BOOTHS SUBSIDISED RAILS
SOURCE: FY26 company filings (IEX, MCX, CDSL, CAMS); GW estimates for infra-SaaS / acquirers. Full comps §07, valuation §08.9
Gravitywell Research · Confidential03 / 33
00: Executive SummarySix Findings
Executive Summary: Continued

00 · iiSix Findings That Reframe the Trade

Finding 01 · The Rails Are Loss-Leaders

Public infrastructure, priced to mobilise: not to monetise.

UPI processes ~85% of digital payments and ~60% of the world’s real-time transactions at zero MDR: a record 757m/day (Jun 2026, 22.72bn txns / ₹28.9L cr).2,3 Adoption, not profit, was the design objective, and it succeeded.

Finding 02 · Rent Sits in Listed Toll-Booths: Priced

41–84% EBITDA, but 42–64x earnings.

IEX 84%, MCX 73%, CDSL 51%, CAMS 45%: confirmed by FY26 filings.9 But the market prices regulatory threat over growth: IEX (slowest) is cheapest (~23x); BSE/MCX (fastest) the dearest (§08).

Finding 03 · MDR Is a ~₹10k-Crore Option

The largest payment network is the least profitable, for now.

A 0.2–0.3% large-merchant MDR is worth an estimated ₹6,000–16,000 cr/yr to the industry, dropping at ~85–95% to profit. PhonePe (~46% of UPI volume, May 2026) has the most leverage. A Parliamentary panel backs tiered MDR (Mar 2026); not yet policy.31

Finding 04 · Data Is the Next Market: Now Ring-Fenced

DPDP reshapes who can monetise the consent rail.

AA leads the world in users (474m consents, 820+ FIs); ULI has disbursed ₹1.75L cr.8,10 But DPDP Rules (Nov 2025) wall the consent-manager off from monetising the data: value migrates to the analytics layer.17

Finding 05 · The Risk Lives in the Unsecured Tail

And the cautionary mirror is China’s Ant.

Household debt has climbed to 45.5% of GDP (Sep 2025) but stays manageable; unsecured loans are 53% of retail slippages and fintech books are >70% unsecured.16 Concentration in two foreign-controlled apps (~79%, below 80% for the first time) is the Ant-crackdown risk India has deferred three times (§18).

Finding 06 · Access Is Won; Usage, Trust & Export Are Not

The unpriced risks are the citizen and the hype.

89% own accounts but 16% sit inactive; bank fraud hit ₹48,021cr (FY26, +46% YoY); ~350m feature-phone users are excluded; 1-in-5 users have been defrauded.21,23,44 And the “DPI exporter” bull case still rests on ~₹330cr of cross-border UPI (FY26-to-date, §18).

“India built a financial operating system: gave the rails away, kept the toll-booths, and is now writing the rules for the data. The trade is everything bolted to the rails that the state could not give away: bought at the right multiple.” Gravitywell Research: Editorial Thesis
Gravitywell Research · Confidential04 / 33
00: Executive SummaryThe Decision-Ready Conclusion
Executive Summary: Conclusion

00 · iiiWhat To Do With This

Public-Market Allocators

Own the toll-booths, at the right multiple.

CDSL, CAMS, BSE, MCX compound at 41–84% EBITDA but trade 42–64x and are beta to market volumes, not defensives. IEX is a regulatory-discount trap. Buy the structure; underwrite the cycle and the multiple (§08).

Private Growth & VC

Back profitable infra; price the reset.

Perfios (PAT +₹104cr) and lending-led names over cash-burning rails (M2P loss ₹256cr). Wealthtech +258% funding. 2022 marks reset 30–45% at IPO: Razorpay filed, PhonePe paused (§11–12).

Policymakers & Regulators

Price the rail; bank the impact.

A tiered MDR restores an investable payments economy without taxing the citizen. The sector is 250k→400k jobs and ~2% of GDP; harden the single point of failure and measure inclusion outcomes (§19–20).

The One-Line Conclusion

India’s financial OS is a finished public good with an unfinished profit model and an unwritten risk model. Capital compounds at the listed choke-points, at a price; the new data rent is walled off as it forms; the citizen carries a fraud-and-exclusion tail; and the export bull case is, so far, diplomacy. Own the moat the state built; pay the right multiple; underwrite the rail, and the risk: it has not yet priced.

What v1.4 Adds for the Professional Readerv1.4
  • Data refreshed to 1 Jul 2026: Jun-2026 UPI (22.72bn / ₹28.9L cr), the Jun-2026 RBI Financial Stability Report (household debt 45.5%, GNPA 1.8%), and RBI Annual Report FY26 bank-fraud (₹48,021cr).
  • §08·ii Liquidity & fair value (new): free float, FII headroom, F&O and a fair-value multiple band for the seven listed toll-booths.
  • §08 MDR sensitivity (expanded): full large-merchant × MDR grid; leverage ranking quantified.
  • §22 Monitoring dashboard (new): dated catalysts and tripwire KPIs to track the thesis.
  • Corrections: microfinance stress restated to MFIN PAR (not a 16% GNPA); demat growth reframed to the FY26 slowdown; IEX coupling timeline updated (APTEL dismissal, Feb 2026).
  • Methodology: confidence rating, change log, AI-assisted-analysis disclosure and a per-section data-vintage table.
Gravitywell Research · Confidential05 / 33
Investor Quick ReadFor Time-Constrained Readers
The 90-Second Version

QUICK READFive Calls, Five Risks

The calls favour the regulated and infrastructural over the subsidised and consumer-facing, at a disciplined entry price. The risks are policy, credit-tail, concentration and trust, not adoption.

Five Conviction CallsCONVICTION
1 · Own the listed toll-boothsHigh

CDSL, CAMS, BSE, MCX: 41–84% EBITDA. But market-volume beta at 42–64x; buy on cycle pullbacks. → §07,08

2 · Back profitable picks-and-shovelsHigh

Perfios (PAT +₹104cr), wealthtech (+258% funding) over cash-burning rails and MDR-capped apps. → §11,12

3 · Hold MDR reform as a ~₹10k-cr optionMedium

PhonePe/Razorpay/Paytm carry the leverage. Asymmetric, binary, politically gated: size it small. → §08,14

4 · Underwrite AA/ULI net of DPDPMedium

ULI past ₹1.75L cr; the consent-manager wall caps who monetises: own the analytics layer. → §09,14

5 · Avoid the unsecured tail & the de-rating moatHigh

Fintech books >70% unsecured; IEX moat under market-coupling attack. → §15,08

Five Material RisksSEVERITY
■
Zero-MDR persists
No reinstatement keeps the payments layer structurally unprofitable.
High
Structural
■
Unsecured / thin-file NPA
53% of retail slippages; fintech 36% growth, >70% unsecured.
High
Active
■
Concentration → Ant risk
~79% UPI in two foreign-owned apps; 30% cap deferred 3×.
High
Deferred
▪
Systemic single point
NPCI runs the whole rail; 4 major UPI outages in 14 months.
Medium
Recurring
◦
Trust & exclusion
Fraud ₹48,021cr (FY26); 1-in-5 defrauded; ~350m feature-phone excluded.
Monitored
Reputational
■High▪Medium◦Monitored
The Bottom Line

Treat India’s financial OS as a utility with a few richly-priced, cyclical private toll-booths attached. The compounders are regulated and infrastructural: buy them on the cycle, not the story; the optionality is a repriced rail and a ring-fenced data market; the risk is a credit tail, a concentration mirror to China’s Ant, and a trust deficit the aggregates hide.

Gravitywell Research · Confidential06 / 33
Part I · The Market
Sections 01: 03

The Size
& the Shape

Before the trade, the market (how big the opportunity is, how attractive its structure, and where it sits on the strategic map): sized, force-scored, and framed.

Reading
Pages 07: 09
Key Figure
$2.1tn TAM · 2030
What You’ll Take Away
01
The market, sized
TAM/SAM/SOM funnel by segment, with the analyst variance reconciled.
02
The structure, scored
Porter’s Five Forces, and why the toll-booths win, not the platforms.
03
The strategic frame
A formal SWOT and a PESTLE for the macro, political and legal lens.
01: The MarketMarket Sizing: TAM / SAM / SOM
Market Sizing

01TAM, SAM & SOM

India fintech is a $2.1tn opportunity but a ~$47bn revenue business today: a 45x gap that is the whole bull case, and the whole trap. Confuse the two and you over-size every segment ~10x.

The Sizing Funnel: Opportunity vs Revenue vs ObtainableFIG. 1.1
TAM = total digital-financial-services opportunity; SAM = fintech revenue pool; SOM = today’s captured revenue. Note the unit shift from market-opportunity to revenue.
TAM · $2.1tn opportunity by 2030 total digital-financial-services market (Inc42) SAM · $190–250bn revenue 2030E fintech revenue pool (BCG/EY/Inc42) SOM · ~$47bn today 2025 revenue · ~35% CAGR
SOURCE: Inc42 State of Indian Fintech 2025 (TAM); BCG/EY (revenue); Inc42 (~$47bn 2025). TAM is market-opportunity, not revenue: the two are ~45x apart.13,41
TABLE 1.1 · Segment sizing: current market & 2030 trajectory
SegmentMarket 2025ForecastCAGRGrade
Digital payments~$6.8B~$7.9B (2026)16–22%B
Digital lending (platform)~$0.5B (’24)$2.45B (2030)31.5%B
Lending (TAM)—$1.3T (2030)>60% of TAMC
Wealthtech~$0.15B (FY24)21% CAGR+258% fund.B
Insurtech$0.9–10B*$11.9B (2033)29–55%C
Neobanking—$253B TAM (’30)~20%C
Fintech revenue~$47B$190–250B (’30)~35%
*Insurtech 2024 baseline ranges $0.9–10B across analysts (~5x) by scope (solutions-only vs end-to-end). Treat all segment forecasts as directional; analyst TAM/CAGR figures vary widely and are graded C unless a named filing supports them.41,43
The Revenue Split That Matters (vs the TAM Trap)

Of the ~$190bn 2030 revenue pool, only lending’s ~50% share is analyst-disclosed (§12); the rest is modelled. Never apply the $2.1tn TAM segment shares to the $190bn revenue base: a ~10x error that recurs across India fintech decks. The investable number is the revenue pool and who captures its margin (§06), not the headline TAM.

Gravitywell Research · Confidential07 / 33
01: The MarketIndustry Attractiveness
Industry Structure

02Porter’s Five Forces

The same five forces score very differently across the three tiers of the stack: which is precisely why the toll-booths are attractive and the platforms are not. Higher force = lower attractiveness.

TABLE 2.1 · Five Forces by tier (■ High force / ▪ Medium / ◦ Low: High force = worse for incumbents)
ForceRails (UPI/AA)Toll-booths (CDSL/CAMS)Platforms (apps)
Competitive rivalry◦ Low (NPCI mono)◦ Low (duopoly)■ High (price war)
Threat of new entry◦ Low (gatekept)◦ Low (licence moat)▪ Med (Navi, super.money)
Supplier power▪ Med (banks)◦ Low■ High (NPCI + banks)
Buyer power■ High (zero-MDR)◦ Low (captive)■ High (multi-homing)
Substitutes▪ Med (cards)◦ Low (compulsory)▪ Med (cash 60% rural)
AttractivenessUtility: not investable directly★★★★★ Most attractive★☆☆☆☆ Least attractive
Toll-Booths: Win

Low rivalry (regulated duopolies), high entry barriers (RBI/SEBI licences), captive buyers on compulsory activity, no substitute. The structurally attractive tier, at a price (§08).

Rails: Utility

NPCI’s monopoly kills rivalry, but buyer power (zero-MDR merchants) extracts all the surplus to the citizen. Not directly investable; the value leaks downstream.

Platforms: Lose

Brutal rivalry, powerful suppliers (NPCI + banks), multi-homing buyers with zero switching cost, cash still a substitute. Margins compressed to ~0 (§07). Avoid as a toll-booth proxy.

The Five Forces verdict is the report’s thesis in one frame: the only tier where all five forces favour the incumbent is the regulated infrastructure. Everything bolted to the rails fights a structurally hostile market.
Gravitywell Research · Confidential08 / 33
01: The MarketSWOT & PESTLE
Strategic Frame

03SWOT & PESTLE

Strengths
  • World-leading DPI; 1.4B identity base; 757m UPI txns/day.
  • Regulated toll-booths at 41–84% EBITDA.
  • ₹81.6L cr MF AUM & SIP ₹30,954cr/mo tailwind.
Weaknesses
  • Zero-MDR: no payments-layer margin.
  • ~79% concentration; >70% unsecured fintech books.
  • Fraud ₹48k cr; ~350m excluded; thin data monetisation.
Opportunities
  • ₹30L cr MSME credit unlock (AA/ULI).
  • Tiered MDR ~₹6–16k cr/yr; credit-on-UPI.
  • DPI export; wealthtech (+258% funding).
Threats
  • DPDP ring-fence; IEX market coupling.
  • Ant-style forced intervention on concentration.
  • NPCI single point of failure; US Section 301.
TABLE 3.1 · PESTLE: the macro, political & legal lens
FactorSignalNet
PoliticalZero-MDR populism into 2029; DPI as soft power; tiered-MDR recommended but not enactedMixed
EconomicGDP $3.9tn, real growth 7.6%, repo 5.25%, bank credit +15.9%, household savings recoveringTailwind
Social89% account ownership but 16% inactive; ~350m excluded; fraud & literacy gapsMixed
TechnologicalUPI/AA/ULI mature; AI 21% adoption / 11% production; agentic & CBDC nascentTailwind
LegalDPDP Rules, Digital Lending & Co-Lending Directions, 125% risk-weight: compliance opex risingHeadwind
EnvironmentalGreen-fintech / climate-tech $5.1bn (2nd globally); DPI rails for green subsidies & carbon credit; SEBI BRSREmerging
Net PESTLE read: a strong economic & technological tailwind, a hardening legal headwind, and a politically-gated payments-pricing question, the configuration that keeps the toll-booths compounding while the rails stay margin-free.26,27,42
Gravitywell Research · Confidential09 / 33
Part II · The Stack
Sections 04: 05

The Architecture
of the OS

Seven layers, modelled on a protocol stack: publicly developed, privately provisioned, regulatorily governed. And the rail that conquered the world by volume while earning nothing: UPI.

Reading
Pages 10: 11
Key Figure
757m UPI txns/day
What You’ll Take Away
01
A layered map of the stack
Identity bedrock to markets & wealth, with each layer’s competitive structure.
02
UPI in global context
Benchmarked vs Visa, Mastercard, PIX, FPS, and the 2026 share reversal.
03
The zero-MDR trap
Why the most-used rail is the least-investable: valued in §08.
02: The StackThe 7-Layer Financial Stack
Architecture

04The 7-Layer Financial Stack

Modelled on the OSI protocol stack: each layer publicly developed, privately provisioned and regulatorily governed. Maturity rises from a universal identity bedrock to a near-monopoly markets layer.

TABLE 4.1 · The India Stack, layer by layer (top = market-facing)
#LayerKey infrastructureStatus
7Markets & WealthCDSL · NSDL · CAMS · KFin · BSE · NSE · MCX · IEX · GIFT/IFSCAMonopoly
6Programmable MoneyDigital Rupee e₹ (circulation fell 24% in FY26) · e-RUPI · UMIEarly
5Commerce ProtocolONDC (Beckn) · 450M+ txns · mobility ~56% · ~85% retail sellers inactiveStalling
4Credit InfrastructureOCEN / ULI (₹1.75L cr) · CIBIL · Experian · Equifax · CRIF · GST SahayProven
3Consent Data LayerAccount Aggregator (DEPA) · 474M consents · 820+ FIs · NBFC-AAsScaling
2Payments RailUPI (757m/day) · IMPS · AePS · FASTag · Bharat BillPay · NPCI · 703 banksDominant
1Identity BedrockAadhaar 1.4B · eKYC ($0.15/ID) · DigiLocker 676M · eSign · UIDAIUniversal
Cross-cutting: credit bureaus · AI underwriting (§19) · embedded-finance APIs · agentic payments (§21). e₹ retail circulation fell to ₹772cr (Mar 2026, −24% YoY); ONDC growth has migrated from retail to transit ticketing.4,5
Built Public

Layers 1–3

Identity, payments and consent: universal reach, near-zero unit cost, no native margin. The data they generate is the contested prize (§09, §14).

Contested Middle

Layers 4–6

Credit, commerce and programmable money: public protocol meets private execution. Mixed traction: ULI scaling, ONDC stalling, e₹ shrinking.

The Rent Layer

Layer 7

Markets & wealth: regulated near-monopolies on compulsory activity. The structural home of excess returns, and mostly listed (§06–08).

Gravitywell Research · Confidential10 / 33
02: The StackUPI at Global Scale
Payments Rail · Global Benchmarks

05UPI Is the World’s Payment OS

By daily volume, UPI has surpassed Visa. By real-time share, it is India’s economy: ~85% of digital payments and ~60% of global real-time transactions, across nine live international corridors.

22.72bn
Txns · Jun 2026 · ₹28.9L cr
757m/d
Daily record · Jun 2026
~29%
P2M share of UPI value
9
Live international corridors
Daily Transaction Volume: Global NetworksFIG. 5.1
Approximate transactions per day, millions. UPI has passed Visa on volume while remaining a domestic, zero-MDR rail.
UPI · India Visa · global Mastercard PIX · Brazil FPS · UK 757m ~639m ~480m ~243m ~90m ~60% of global real-time ~7.3bn/mo · fastest-scaling rail
SOURCE: NPCI (Jun 2026); IMF; RBI Annual Report; BCB (PIX). Visa/Mastercard are global, all-rail; UPI & PIX are domestic real-time.2,3
UPI Share · Volume · the 2026 reversal
AppDec 2025May 2026
PhonePe (Walmart)45.35%46.26%
Google Pay (Alphabet)34.64%32.75%
Paytm7.65%7.91%
Navi / super.money / others12.36%13.08%
~79% in two foreign-controlled apps (May 2026: below 80% for the first time): the concentration the 30% cap (deadline 31 Dec 2026, deferred 3×) is meant to break. The Ant parallel: §18.2,32
Investment Implication

Zero-MDR caps app profit at break-even. P2M is ~63% of UPI volume but only ~29% of value (tiny-ticket QR dominates): the number that governs any MDR calculation (§08). Credit-on-UPI (~16% of card spends now on RuPay) is the live monetisation wedge.31

228bn
UPI txns · CY2025
~85%
Of India digital payments
9
Live corridors (UAE, SG, Nepal…)
703
Banks live on NPCI rails
Gravitywell Research · Confidential11 / 33
Part III · Where the Rent Accrues
Sections 06: 10

The Toll-Booths
& the Data Market

Profits concentrate in a thin layer of regulated near-monopolies: mostly listed, richly priced. The data rent forms on top, ring-fenced by DPDP. And the value chain shows exactly where the margin pools sit.

Reading
Pages 12: 17
Key Figure
41–84% EBITDA · 42–64x
What You’ll Take Away
01
Where margin concentrates: & its ₹ size
Nine pools, the tier line, the absolute revenue, and the value chain.
02
The comps and the valuation
FY26 multiples, multiple history, cyclicality and the MDR option.
03
The AA / ULI unlock
India’s open-finance ecosystem vs Plaid: net of the DPDP wall.
03: Where the Rent AccruesThe Profit-Pool Map & Its ₹ Size
Value Capture · Profit-Pool Analysis

06Where Economic Rent Accrues

The rails are zero-margin utilities. Profit concentrates in regulated near-monopolies, but the “toll-booth” label holds only for the licensed infrastructure, not the platforms. And the pools, in ₹, are smaller than the throughput implies.

High rent · 40–84% EBITDA Medium · 15–40% Thin / competitive
TABLE 6.1 · Profit pools: margin, ₹ revenue pool, and whether the thesis applies
PoolEBITDAIndustry rev poolStructureToll?
Power/Comm. Exch.73–84%≈₹2,900cr AIEX · MCXYes
Depositories~51%≈₹2,900cr ACDSL · NSDL · 22.5Cr dematYes
Stock Exchange48–64%BSE ₹4,834cr ABSE listed · NSE unlistedYes
Fund RTAs41–45%≈₹2,800cr ACAMS (~68%) · KFin (~32%)Yes
Credit Bureaus~50–60%≈₹2,000cr BCIBIL · Experian · Equifax · CRIFYes
PA / PG gateways~25%₹3.5–4.5k cr net CRazorpay · Pine Labs · CashfreePartial
Broking platforms~35% opcompetitiveAngel One · Zerodha · GrowwNo
Payment/insur. plat.~6–10%thinPaytm · PB FintechNo
UPI / ONDC apps~0 / neg.~0 MDRPhonePe · GPay · ONDC appsNo
The whole digital-payments gross revenue pool is ≈₹2.5–4.2L cr C, but India earns ~0 MDR on ~$3.4tn UPI throughput, so monetisation has migrated downstream to lending/insurance. Each toll-booth pool is only ₹2,000–4,800cr absolute: high-margin, slow-compounding annuities, not vast markets.
Keystone Thesis: Refined

The “Visa, Moody’s and Bloomberg” of India are CDSL, CIBIL and CAMS: toll-booths at 41–84% EBITDA. The thesis does not extend to the platforms (Paytm, Angel One, PB Fintech). Buy the licensed infrastructure; the platforms are a thinner, more competitive trade (§02 Five Forces confirms it).

Gravitywell Research · Confidential12 / 33
03: Where the Rent AccruesThe Listed Toll-Booths
Public Comparables · FY26

07The Listed Toll-Booths

The thesis is testable: most names are listed with audited FY26 filings. The infra layer confirms 41–84% EBITDA and regulatory moats. The platforms do not qualify.

TABLE 7.1 · India financial-infrastructure & fintech comps: FY26 (USD at ₹94.7)
CompanyMkt capRev (ops)EBITDAPAT (YoY)P/E
BSE$16.8B₹4,834cr48–64%+88%~64x
MCX$7.7B₹2,302cr73%+138%~54x
CDSL$3.0B₹1,145cr~51%−14%~61x
CAMS$2.1B₹1,516cr45%+2%~43x
NSDL IPO Aug’25$1.9B₹1,530crcustody+11%~47x
KFin$1.6B₹1,301cr41%+3%*~44x
IEX moat at risk$1.2B₹616cr~84%+15%~23x
Angel One$3.3B₹5,140cr~35% op−22%~34x
Paytm$7.3B₹8,437cr~6%turned +~95x
PB Fintech$7.9B₹6,794cr~7–10%+90%~112x
*KFin FY26 growth is largely inorganic (Ascent acquisition); organic Q4 ~4.6%. Consolidated; total income > ops-revenue for cash-rich exchanges. CAMS did a 1:5 split, Angel One 1:10 (FY26): split-adjust per-share data. P/E noisy for newly-profitable Paytm / PB Fintech.9
The One Eroding Moat

IEX posts 84% EBITDA but faces market coupling: CERC’s order (Jul 2025) routes price discovery through Grid-India. APTEL dismissed IEX’s appeal (Feb 2026); CERC’s draft notification (20 Apr 2026) codifies Grid-India as the coupling operator, removing the main legal off-ramp. Coupling first hits the Day-Ahead Market: ~81% of IEX volume, so FY26’s +17% volume does not yet reflect it; analysts model share falling toward ~50% by FY28. Down ~34% YoY (₹212→₹125); P/E ~23x. The cheapest infra multiple is a regulatory-threat discount.

The Tailwind Under the RTAs

MF AUM ₹81.58 lakh cr (May 2026); SIP ₹30,953cr/month (+16% YoY); 27.65cr folios. Demat ~22.9cr (May 2026; CDSL 18.4cr + NSDL 4.5cr) feeds CDSL/NSDL, though FY26 new-account additions slowed ~22%. And SEBI’s F&O clampdown cut NSE F&O turnover −18% in FY26: these revenues are cyclical (§08).29

All eleven are investable on public markets. The infra cluster (BSE, MCX, CDSL, CAMS, KFin) is a richly-priced compounder basket; the platforms (Angel One, Paytm, PB Fintech) are competitive and should not be underwritten as toll-booths.
Gravitywell Research · Confidential13 / 33
03: Where the Rent AccruesValuation & What’s Priced In
Valuation Framework

08Valuation & What’s Priced In

High margins are not the trade: the entry multiple is. These names are growth-cyclical, beta to market volumes, and serially de-rated by regulators. The market already prices regulatory threat over growth.

TABLE 8.1 · Multiple history, returns & cyclicality (split-adjusted)
Co.P/E now5yr P/E range5yr rev CAGRROEDiv yldBeta driver
CDSL~61x17–90x~27%24.5%~1.0%Demat + txn/IPO volumes
CAMS~43x30–50x~17%~37%~1.7%MF AUM (telescopic)
KFin~44x42–64x~22%24.5%~1.3%MF AUM + fund-admin
BSE~64x23–80x~50%~40%~0.25%Index-options premium ▲▲
MCX~54x30–55x~43%~56%~0.2%Commodity ADT ▲▲
IEX~23x23–71x~14%~39%~2.7%Power volume; reg. de-rate
Angel One~34x10–35x~32%~16%~1.5%Retail F&O orders ▲▲
The paradox: IEX (slowest grower, ~14%) is cheapest (~23x); BSE/MCX (fastest, ~43–50%) the dearest (~54–64x), the market is pricing regulatory threat, not growth. De-raters: SEBI true-to-label cut CDSL ~−33% (2025); TER consultations cut CAMS/KFin −24–28%; IEX market coupling −45–50%; Angel One’s F&O framework cut Q1 FY26 PAT −61%.30
MDR Reform: Sensitivity Model C

UPI value ~₹360L cr/yr · P2M ~29% of value · large-merchant ~50% (est, range 30–60%) · MDR 0.2–0.3%:

MDR@30%@50%@60%
0.20%~₹6,200cr~₹10,400cr~₹12,500cr
0.25%~₹7,800cr~₹13,000cr~₹15,700cr
0.30%~₹9,400cr~₹15,600cr~₹18,800cr

Columns = large-merchant share of P2M value (weakest input; §Methodology). Range ~₹6,200–18,800cr/yr.

Flows at ~85–95% to pre-tax profit (UPI MDR ≈0 today). Leverage tracks large-merchant P2M share: PhonePe (~46% of UPI volume) ≫ Paytm (~8%) > Razorpay > Pine Labs (acquirers capture merchant-side economics). Common error: using 63% volume share doubles it; use 29% by value.

Swing Variables: What Moves the Answer
1
Market-volume cycle
BSE/MCX/Angel revenue is high operating-leverage beta to F&O/ADT.
2
MDR reinstatement
Binary; ±₹6–16k cr to the payments industry.
3
IEX market coupling
Binary; ~−45% already in the price.
4
SEBI fee caps (TER / true-to-label)
Recurring yield drag on CDSL/CAMS/KFin.
5
AA/ULI credit scale-up
Slow-burn upside to bureaus & infra-SaaS.
Verdict: the infra basket is a quality-at-a-price trade, not a value one. Entry on market-cycle drawdowns (when F&O/ADT volumes compress the cyclicals) beats chasing the compounder narrative at 60x. IEX is cheap for a reason.
Gravitywell Research · Confidential14 / 33
03: Where the Rent AccruesLiquidity, Float & Fair Value
Tradability & Entry · For the Allocator Who Must Size It

08 · iiLiquidity, Float & Fair Value

The thesis names “the right multiple”, but a desk also needs float, liquidity and derivatives to size the position. All seven trade single-stock F&O; FII holdings sit well below caps; the fair-value bands are GW judgements, grade C, not price targets.

TABLE 7.2 · Liquidity, ownership & fair value: listed toll-booths (as of 1 Jul 2026, USD at ₹94.7)
CompanyMkt capFree floatFII% (cap)F&OP/E nowGW fair P/E C
BSE$16.9B~100%19.4% (49%)Yes~64x45–55x
MCX$7.6B~100%26.1% (49%)Yes~54x40–50x
Angel One$3.2B~71%12.8% (—)Yes~33x22–30x
CDSL$2.9B85%11.4% (—)Yes~61x40–50x
CAMS$2.1B~100%44.4% (—)Yes~41x38–45x
KFin$1.6B~77%26.3% (—)Yes~44x38–46x
IEX moat at risk$1.2B~100%14.2% (—)Yes~23x18–25x
No-promoter names (BSE, MCX, CAMS, IEX) run ~100% free float, institutionally held; CDSL float 85% (BSE holds 15% strategic); KFin float ~77% (General Atlantic 22.9%); Angel One float ~71% (promoter 28.8%). FII sits well below the 49% exchange cap: no cap binds RTAs or brokers, so headroom is ample everywhere. Basket ≈ $35bn, led by BSE. ADV (traded value) not shown: no source publishes clean multi-session figures; all seven are liquid, F&O-eligible large-caps. Shareholding as at Mar-2026 (Q4FY26) filings.9,30
How to Read the Fair-Value Band

Bands anchor to 5-yr P/E ranges de-rated for regulatory state (grade C). BSE (64x) and MCX (54x) sit above band: a cyclical premium; enter on volume drawdowns, not at the peak. Angel One (33x) screens cheap but F&O is ~47% of income and structurally capped: the discount is deserved. IEX (23x) is a coupling-overhang trap, not value. The compounders (CDSL/CAMS/KFin) are quality at a full price.

2026 Regulatory Scorecard

Structural hits already in price: IEX (CERC coupling; APTEL appeal dismissed Feb 2026) and Angel One (SEBI F&O framework; FY26 the earnings trough). Feared-but-defused: CAMS/KFin TER cuts (~10bps net, GST excluded: RTAs rallied) and BSE (F&O framework absorbed, re-rated on Sensex-options share). CDSL mid (true-to-label margin drag, PAT −14%). MCX the upside outlier (options boom, PAT +138%, coal-exchange optionality, Apr 2026).

Gravitywell Research · Confidential15 / 33
03: Where the Rent AccruesAccount Aggregator & the Data Market
Data Capital Markets · Open-Finance Benchmark

09The World’s Largest Open-Finance Ecosystem

India’s Account Aggregator leads the world in users; the weakness is thin per-consent monetisation versus Plaid, and DPDP now walls the consent-manager off from monetising the data itself.

TABLE 9.1 · Open finance benchmarked: India (AA) vs USA (Plaid)
MetricIndia · Account AggregatorUSA · Plaid
Scale of adoption474M cumulative consents500M+ accounts connected
Accounts enabled2.88B+ for data sharing*~50% of US bank accts (est.)
Institutions on-network820+ FIs12,000+ FIs
Monetisation₹1.47L cr disbursed · H1 FY26$546M est. ARR · 2025
Market value~10% of personal loans (vol.)$8B valuation · Feb 2026
*“Accounts enabled” (2.88B) ≠ “accounts linked by users” (~305M); FIP/FIU counts vary 135–999 by definition B. AA leads on mandate-backed reach; Plaid on revenue-per-connection: the question is whether AA monetisation converges as ULI scales.10
Credit Gap · The Runway
₹30L Cr

MSME addressable gap: ~24% of debt demand (SIDBI×Crisil, May 2025). Not the ₹69L cr gross-demand figure.7

ULI Pipeline
₹1.75L cr

64 lenders (41 banks + 23 NBFCs); 3.2M loans to Oct 2025; still in pilot, no public launch.8

The DPDP Ring-Fence: read this before underwriting AA monetisation

DPDP Rules (notified 13 Nov 2025) require a Consent Manager to be India-incorporated, ₹2cr net worth, operate data-blind, and: critically: a consent manager cannot also be the data fiduciary/processor monetising the same data. Value migrates to whoever owns the analytics/underwriting between consent and disbursement (the §06 infra-SaaS pool). CM obligations bite ~Nov 2026; full compliance May 2027; Board not yet staffed.17 Contrast §18: this mandate-led model out-scaled the EU’s PSD2.

Gravitywell Research · Confidential16 / 33
03: Where the Rent AccruesValue-Chain Economics & GTM
Value Chain · Business Models

10Value-Chain Economics & GTM

Trace a rupee of financial activity up the stack and the margin pools light up unevenly: the rails earn nothing, the licensed infrastructure earns most, and the capital intensity inverts the returns.

The Value Chain: Margin Pool by NodeFIG. 10.1
EBITDA margin (bar height) × indicative ₹ pool, traced along the stack. Public rails → private toll-booths.
Rails (UPI)~0% UPI apps~10% Acquirers~25% RTAs~45% Bureaus~55% Exch/Dep~70%+ ₹0 pool ₹2.8kcr ₹2.9kcr
SOURCE: §06 profit pools; FY26 filings. Margin rises monotonically up the chain; the ₹ pools stay modest: annuity, not scale.9
TABLE 10.1 · Business model · GTM · cost structure · capital intensity
ModelGTMCost driverCapital intensity
LendingDistribution + co-lending (bank balance sheet)Credit losses · cost of capitalVery high
PaymentsScale land-grab on a free gov’t railIncentives · cashbacks · cloudLow
Infra SaaS / APILand-and-expand into FIs (B2B)R&D · enterprise salesMedium
Toll-boothsRegulatory licence; captive flowTech + compliance (fixed)Low (asset-light)
The Distribution / Balance-Sheet Split

The dominant operating model is bank/NBFC-as-balance-sheet, fintech-as-distribution (detailed §19). The fintech sources and services; the regulated lender holds the loan and the NPA, posting ≤5% first-loss (DLG). Co-lending AUM est. ₹1.8–2.0L cr (FY26); PhonePe alone distributes for 56 lending partners (₹14,270cr cumulative). For an allocator: the credit risk sits on bank books: own the distribution and the infrastructure, not the balance sheet.38

Gravitywell Research · Confidential17 / 33
Part IV · Capital, Funding & Exits
Sections 11: 13

What You Can
Actually Buy

The private-market map with real unit economics, the funding rotation, the M&A wave and the down-round reset, with returns math, and three probability-weighted trajectories to 2030–2040.

Reading
Pages 18: 20
Key Figure
$51bn cumulative VC
What You’ll Take Away
01
The investable universe, with P&L
Who is actually profitable, who is burning, on what scale metric.
02
Funding, M&A & returns
The rotation, the consolidation wave, and IRR by scenario.
03
Probability-weighted scenarios
Bull/base/bear with weights and explicit triggers.
04: Capital, Funding & ExitsPrivate-Market Map & Unit Economics
Investment Landscape · Private Markets

11Private-Market Map & Unit Economics

Valuations alone mislead. Below each mark is the FY25 P&L: who is profitable, who is burning, and on what scale metric. Profitable infra (Perfios) and lending (KreditBee) sit beside cash-hungry rails (M2P, Cred).

TABLE 11.1 · Selected India fintech: valuation & FY25 unit economics
CompanyValuationFY25 revPATScale metric
PhonePe~$15B₹7,115cr+₹630cr adj.†UDRHP filed
Razorpay~$5–6B₹3,783cr−₹1,209cr†TPV ~$180B
Cred~$4.5B₹2,735cr−₹1,457crTPV ₹8.5L cr
KreditBee$1.5B₹2,712cr+₹473crAUM ₹10,102cr
Pine Labs$3.3B₹2,274cr−₹145cr‡GTV ₹11.4L cr
Perfios~$1B₹670cr+₹104crprofitable SaaS
Juspay$1.27B₹514cr+₹62cr‡Series D
M2P Fintechundisc.₹506cr−₹256crEBITDA −₹223cr
†PhonePe ₹2,358cr & Razorpay losses are one-time ESOP/redomicile charges: both operationally profitable (PhonePe adj. EBIT +₹117cr; Razorpay payments EBITDA+). ‡Pine Labs Q1 FY26 turned net-positive; Juspay PAT flattered by deferred-tax. Entity traps: Razorpay India-entity (PAT+ since FY23) vs consolidated; KreditBee group vs standalone.14,40
The Allocation Read

Separate the operationally-profitable (Perfios, KreditBee, PhonePe adj., Razorpay payments) from the genuinely loss-making (M2P, Cred). The clean infra-SaaS compounder is Perfios; the scaled lending book is KreditBee (ROMA 4.8%). Razorpay’s listing (end-2026) is the cohort’s price-discovery event. Underwrite to adjusted, operating economics: not headline reported losses or 2021 marks.

Gravitywell Research · Confidential18 / 33
04: Capital, Funding & ExitsFunding, M&A, Exits & Returns
Capital Flows · Consolidation · Returns

12Funding, M&A, Exits & Returns

$51bn has flowed into India fintech; the boom reversed in 2026, capital is rotating to wealthtech, consolidation is rising, and the 2022 marks are resetting 30–45% at exit.

Annual Fintech FundingFIG. 12.1
US$bn raised. Peak 2021; reset since.
2021$8.4 2022$4.1 2023$2.7 2024$2.3 2025$2.4 26E$1.5
Tracxn / Venture Intelligence (annual); Inc42 H1-2026 ≈ $1.3bn, −19% YoY. Provider totals differ by taxonomy.14,41
M&A & Consolidation · 2025–26DEALS
Meta → CRED$900M @ $4.5bn
Slice → NESFBbank licence
InCred Money → Stockobroking
Partners Group → Infinity Fincorp$230M (75%)
Cumulative (Tracxn, May’26)306 acq · 76 IPOs
Theme: VC slowdown is pushing full-stack consolidation (payments+lending+bank licence); Meta’s $900M into CRED at a $4.5bn mark (Jun 2026) reprices the reset cohort. EV/revenue multiples are not publicly disclosed on Indian fintech deals: size on operating economics, not rumoured marks.45
IRR by Scenario C: illustrative 5yr
ExposureBullBaseBear
Listed infra (public eq.)~22%~12%~2%
Private infra-SaaS~30%~18%~5%
MDR option (apps)~40%+~8%neg.
Unsecured lenders~20%~10%neg.
GW model. Assumptions: entry at current multiples (infra 40–64x; IEX ~23x), 5-yr hold, exit multiple per §13 scenario (base de-rates to 35–45x), revenue CAGR ~35% base / ~45% bull / ~20% bear. India VC pooled returns are dispersion-heavy: top-quartile drives the index (Bain India VC 2026).
The Down-Round Reset

2022 marks clear 30–45% lower at liquidity: Pine Labs ($5–6B→$3.3B), Razorpay ($7.5B→$5–6B), Cred ($6.4B→$3.5B→$4.5B). Sector revenue ~$47bn (2025) → $190bn (2030) ≈ 35% CAGR. Underwrite to reset marks, not 2021 vintages. Funding rotated to wealthtech (+258%) as lending contracted.15

GIFT City: IFSC banking assets $106.7bn (Feb 2026, 7x in five years; rising toward ~$111bn by Q4-FY26); ~1,150 entities, 37 banks: the offshore structuring venue for the exit cohort.12
Gravitywell Research · Confidential19 / 33
04: Capital, Funding & ExitsScenarios to 2030–2040
Scenario Analysis · 2030: 2040

13Three Trajectories: Probability-Weighted

Weights are Gravitywell judgements, not market-implied. The base case dominates; the bull case needs a tiered MDR and a real export pipeline; the bear case is the unsecured tail plus regulatory overhang.

TABLE 13.1 · Scenario matrix: India’s financial OS to 2030–2040
IndicatorBull · 25%Base · 50%Bear · 25%
Fintech revenue · 2030E$250B+$190B (BCG)†$120B
Credit-to-GDP · 2035E60%50%42%
India Stack export nations50+20+ (MOSIP/UPI)Sub-10
Annual fintech VC$5–10B$2–4B$1–2B
Infra comps re-rateHold 50–70xDe-rate to 35–45x25–30x
†$190B is BCG’s Aug-2024 estimate (~35% CAGR), predating the 2025–26 slowdown. Probability-weighted 2030E revenue ≈ $188B: base-case-anchored.13
Bull · 25% · DPI Exporter

ULI closes most of the MSME gap; tiered MDR reinstated; a real export pipeline converts (not just MOUs); e₹ finds product-market fit. Public market-cap creation across the infra basket.

Base · 50% · Domestic Champion

Strong domestic success; ONDC a transit niche; AA to ~25% penetration; selective export. Razorpay lists, PhonePe re-files; ULI scales but the MSME gap persists; comps de-rate to 35–45x.

Bear · 25% · Fragmented Rails

ONDC stalls; DPDP/probe overhang; zero-MDR maintained; unsecured NPA spike; concentration entrenches (cap deferred again); e₹ stays sub-0.1% of money supply.

Buy Signal (flips toward bull)

ULI past ₹2.5L cr; AA monthly disbursements > ₹25,000cr; a tiered-MDR notification; a disclosed-value export deal; DPDP Board operational.

Risk Watch (flips toward bear)

Unsecured slippage acceleration; household debt past ~48% of GDP (already 45.5%, Sep 2025); IEX coupling go-live (APTEL appeal dismissed Feb 2026); a 4th deferral of the 30% cap; a systemic UPI outage.16

Gravitywell Research · Confidential20 / 33
Part V · Risk, Regulation & the Citizen
Sections 14: 17

The Rules,
the Tail & the Customer

The regulatory regime hardened in 2025. The credit risk lives in the unsecured tail. The citizen carries a fraud-and-exclusion burden, and the customer, finally, gets a voice.

Reading
Pages 21: 24
Key Figure
1-in-5 defrauded
What You’ll Take Away
01
The regulatory calendar
DPDP, Digital Lending, MDR and the 30% cap: dated.
02
The unsecured tail & the citizen
Where the NPAs form; fraud, inclusion and the exclusion tail.
03
The customer voice
Segmentation, NPS, switching barriers and what users actually pay.
05: Risk, Regulation & the CitizenThe Regulatory Calendar
Policy & Compliance

14The Regulatory Calendar

The biggest change since the OS thesis was written is regulatory, not technological. Three new regimes (data, digital lending and payment pricing) reshape every monetisation path in this report.

TABLE 14.1 · The live regulatory calendar (as of 1 Jul 2026)
RegimeStatus / key datesBite
DPDP Rules 2025Notified 13 Nov 2025; CM obligations ~Nov 2026; full compliance May 2027; Board not yet staffed₹250cr cap
Digital Lending Directions 2025Issued 8 May 2025; multi-lender rules 1 Nov 2025; public DLA directory live Jul 2025DLG 5%
Co-Lending Directions 2025Effective 1 Jan 2026; all lending (not just PSL); ≥10% risk retentionNew
Unsecured risk-weights125% (Nov 2023) stands; only MFI (→100%) & bank→NBFC eased Feb 2025125%
UPI MDRZero-MDR intact; tiered-MDR proposed by Parl. panel (12 Mar 2026); RuPay-CC-on-UPI >₹2k from 1 Jun 2026Proposal
NPCI 30% market capDeadline 31 Dec 2026; PhonePe ~46%; deferred 3×; 4th plausiblePending
SEBI F&O / TERF&O framework phased to Dec 2025 (NSE F&O −18%); base-expense framework eff 1 Apr 2026Enforced
The Fiscal Trade-Off

Zero-MDR is funded by the UPI incentive: ₹2,000cr allocated FY27 vs the industry’s ~₹4,500cr ask (incentives cover only ~11% of cost). DFS told Parliament zero-MDR is “financially unsustainable”; RBI Governor: “UPI is not truly free.” A tiered MDR is the structural fix: politically gated (§19).31

For the Investor

Every digital-lending and data thesis now carries new opex: DLG capped at 5%, ≥10% co-lending risk retention, data localization, neutral-comparison mandates, and DPDP consent/breach machinery (₹250cr ceiling). Price compliance as a structural cost, not a one-off.17,18

Gravitywell Research · Confidential21 / 33
05: Risk, Regulation & the CitizenCredit Risk & the Unsecured Tail
Credit Cycle · Asset Quality

15Credit Risk & the Unsecured Tail

Aggregate leverage is benign and system GNPA is at a multi-decade low. The danger is concentrated: the unsecured consumption tail, small-ticket digital loans, and the multi-lender borrower.

45.5%
Household debt / GDP (Sep 2025)
1.8%
System GNPA (Mar 2026) · multi-decade low
53%
Of retail slippages are unsecured
36%
Fintech-lending YoY growth
Where the NPAs Form
  • Unsecured loans drive 53% of retail slippages; ~76% of slippages at private banks. The first default is the unsecured loan; it cascades to secured.
  • Fintech books are >70% unsecured; impairment elevated among borrowers with 5+ lenders.
  • Small-ticket PLs (<₹10K): +44% delinquency jump in the late-2023/24 vintage: NBFC/fintech subprime, outside the top-100 cities.
  • Credit-card NPAs +28% YoY; the one retail segment still not eased from the 150% risk-weight.
The Microfinance Cycle

Peak stress FY25: GLP −13.9%, PAR 31–180d peaked ~6.3% (Mar 2025), ~4M borrowers exited. MFIN guardrails 2.0 (3-lender cap, ₹2L household limit, Apr 2025) plus the Feb-2025 risk-weight cut reopened funding. Q4 FY26 inflection: GLP ₹3.25L cr, +3% QoQ (first growth in 7 quarters); PAR 31–180d back to ~2.0%. A real cycle, not a blip.20

RBI’s Verdict, and the Caveat

RBI’s FSR (June 2026) puts household debt at 45.5% of GDP (Sep 2025, up from 41.9%): rising but still moderate by EM standards, ~69% prime-concentrated and asset-backed; non-housing retail is 58.4% of household borrowings (Mar 2026) and retail unsecured GNPA sits at 1.7%. The flagged risk is the unsecured consumption tail, not the aggregate. For a fintech-lending thesis, that tail is the business, and the co-lending model (§19) pushes the NPA onto the bank’s book, not the fintech’s.16

Gravitywell Research · Confidential22 / 33
05: Risk, Regulation & the CitizenThe Citizen: Inclusion, Fraud & Exclusion
The End-User Lens · Public Interest

16The Citizen: Inclusion, Fraud & Exclusion

Access is won; usage and trust are not. The headline 89% account ownership hides an inactivity, fraud and exclusion tail that is the sector’s real reputational and policy risk.

89%
Adult account ownership (Findex 2025)
16%
Accounts inactive · ~26% PMJDY inoperative
₹48kcr
Bank fraud · FY26 (+46% YoY)
~350m
Feature-phone users · app-excluded
Fraud & Predation
  • Bank fraud ₹48,021cr FY26 (+46% YoY, but value inflated by 314 legacy cases reclassified post-SC; case count fell to 10,114); cyber fraud ₹22,495cr / 28.15 lakh cases (2025), 76% investment scams.
  • Digital-arrest losses fell −66% (2025) after awareness campaigns: a rare win.
  • Loan-app harassment/suicide cluster (Kerala, H1 2026); MeitY blocked 87 apps, RBI→Google 47.
  • Mitigation: mandatory 2FA on all digital payments (1 Apr 2026); verified-payee-name; NPCI MuleHunter.AI (23 banks).
Grievance & Redress

RBI Ombudsman complaints 13.34 lakh in FY25 (+13.6%): loans & advances (29%) and credit cards (+20%) the top categories; banks 81.5%, NBFCs 14.8%. DPDP adds 90-day grievance resolution and 72-hour breach notification, but the Data Protection Board is not yet staffed, so redress efficacy is unproven.25,17

The Exclusion Tail: who the OS leaves out
~350m

Feature-phone users locked out of app-based UPI/AA (Counterpoint 2025); UPI123Pay exists but publishes no adoption data.

27%

Financial literacy: last measured 2019 (NCFE FLIS); no national survey since, NFLIS 2025 still in field. Digital literacy ~37%; ~90% not English-proficient.

56 / 84

Women vs men mobile ownership (%). Ownership gap ~nil; the gap is now usage, elderly access and vernacular UI.

RBI Financial Inclusion Index reached 67.0 (Mar 2025). The frontier is no longer the bank account: it is whether the elderly, the feature-phone user and the non-English speaker can transact safely without being defrauded or excluded.21,33

Gravitywell Research · Confidential23 / 33
05: Risk, Regulation & the CitizenThe Customer: Demand-Side, NPS & Pricing
Demand-Side · Customer Intelligence

17The Customer

Who uses the OS, how satisfied they are, what makes them switch, and what they actually pay. The demand side the investor lens usually skips, and where the next 200 million users will be won or lost.

Segmentation & Behaviour
  • Urban 62% / rural-semiurban 38% prefer UPI; small-town adoption 42%→50% (FY24→25).
  • Gen Z (18–35) the highest-adoption cohort; UPI Lite / tap-and-go skew.
  • Women <30% of UPI users: a 200m+ addressable gap.
  • Income tiers: affluent→credit/CRED; middle→UPI+wallet; low→PMJDY+agent-assisted.
  • Multi-homing: users hold 2–3 apps; switching is low (habit + interoperability).
Satisfaction & NPS B
Paytm P2M NPS (RedSeer)42
PhonePe · Trustpilot1.6 / 5
Google Pay · Trustpilot1.4 / 5
Rural prefer UPI (EY-CII)38%
UPI tech-decline rate~0.8%
Merchant NPS strong (Paytm 42); consumer Trustpilot low but selection-biased toward complainants. Gateway success 92–96%.44,46
Pain Points & Switching Barriers

1-in-5 users defrauded in 3 years; 51% don’t report; UPI fraud 10.64 lakh cases / ₹805cr+ (Apr–Nov 2025). Top decline cause: bank-server timeout (35–45%). Switching is blocked by habit, rewards (cited by 30% of young users) and lack of cross-app interoperability: a moat for the incumbents, a barrier for entrants. The blockers on the next 200m: literacy (<50% rural understand KYC/PIN), language (~90% not English-proficient), feature phones, and trust after a failed transaction.44,33

What Users Actually Pay B
P2P / P2M UPI₹0 (zero-MDR)
App-based personal loan APR12–28%
Margin-trading (MTF) interest~15% p.a.
Equity-delivery brokerage₹0 (Zerodha)*
Credit-on-UPI / BNPLopaque
*“Zero brokerage” ≠ zero cost: statutory STT/stamp/GST/DP charges remain. Credit-on-UPI & BNPL APRs are not transparently disclosed to consumers: a redress & conduct risk.29,46
The Competitive Read by Segment

Payments: Paytm leads merchant NPS in “Bharat”; PhonePe/GPay lead volume. Credit: CRED owns the premium (25m members, 750+ score gate). Wealth: Groww wins Gen-Z mindshare; Zerodha stays broker-first. The strategic battle is the super-app: Paytm (payments+lending+insurance) vs PhonePe (payments+credit) vs CRED (credit-first + UPI), fought on trust and rewards, not price (which is zero).44,46

Gravitywell Research · Confidential24 / 33
Part VI · Forces, Impact & the Trade
Sections 18: 22

The Mirror,
the Impact & the Trade

What the world’s other state-built rails teach. The structural forces beneath the surface. The economic and ESG impact. The technology horizon. And, finally, the trade.

Reading
Pages 25: 29
Key Figure
~2% of GDP
What You’ll Take Away
01
International comparables & forces
PIX, the Ant cautionary tale; banks, AI & the single point of failure.
02
Impact & the horizon
Jobs, GDP, ESG, and the 5–10-year technology roadmap.
03
The trade
Model positioning, a dated catalyst calendar, a stakeholder playbook.
06: Forces, Impact & the TradeInternational Comparables
Comparative · State-Built Payment Rails

18International Comparables: PIX · Ant · PSD2

India is not the first to build a state rail. Brazil shows how to monetise one; China shows what happens when private apps grow too powerful on it; the EU shows how mandate-without-monetisation fails. And the export bull case shrinks under scrutiny.

TABLE 18.1 · State-built / open-finance rails compared
SystemScaleModelLesson for India
Brazil · PIX~7.3bn txns/mo; ~83–93% of adultsFree P2P; ~0.22–0.33% MDR borne by banks; credit layerMonetise the credit layer, not the free rail
China · Alipay+WeChat>90% of mobile pay; >$80tn (2024)Private duopoly on a state rail; Ant IPO killedThe state will intervene once concentration turns systemic
EU · PSD2559 TPPs; ~13% retail; ~19% monetiseMandated free API access; no charging mechanismMandate without monetisation stalls; AA’s central mandate beat it
UK · Open Banking16.5M users; 351M payments (2025)VRP & first paid open-banking API model (2025)Paid API rails are the route to monetisation
The Ant Cautionary Tale

Ant’s $34.5–37bn IPO was halted 3 Nov 2020, days before listing; forced restructuring, a ~$985M fine (2023), Jack Ma ceding control, and a ~70% valuation collapse (~$280–315bn → ~$78.5bn). India’s mirror: ~79% of UPI in two foreign-controlled apps, with the 30% cap deferred three times. China shows the intervention is forced and brutal once concentration is systemic.34,35

The Export Reality-Check

Cross-border UPI ran ~755,000 transactions / ~$29.5m in FY25, nearly doubling to ~1.49m / ~₹330cr in FY26-to-date (through Dec 2025): still against ~757m domestic transactions per day. India–Nepal launched the first two-way P2P corridor (Jun 2026); the rest are tourist/diaspora acceptance, not selling UPI abroad. MOSIP (the more credible export, ~29 countries, 185m IDs) is donor-funded and given away. The bull case is soft power and pipeline optionality: not demonstrated revenue.28,36

Brazil also shows the geopolitical tail: a US Section 301 investigation (determination 1 Jun 2026) proposed 25% tariffs citing payment policies that displaced US card networks, a warning as UPI internationalises.37
Gravitywell Research · Confidential25 / 33
06: Forces, Impact & the TradeStructural Forces
Beneath the Surface

19Structural Forces: Banks, AI & Systemic Risk

Three forces the headline numbers miss: who actually holds the credit risk, how far AI really is, and the fragility of a one-operator rail.

Banks vs Fintech: Distribution, Not Disintermediation

The dominant model is bank/NBFC-as-balance-sheet, fintech-as-distribution. The regulated entity is lender of record and holds the NPA; the fintech sources and posts ≤5% first-loss (DLG). PhonePe (56 lending partners, ₹14,270cr cumulative disbursals) and Paytm (lending rev ₹611cr Q2 FY26, ~80% under DLG) are distributors, not lenders. Co-lending AUM ~₹1.1L cr (CRISIL, Mar 2025) → GW est. ₹1.8–2.0L cr FY26 C (no published FY26 print: treat as modelled). RBI’s Co-Lending Directions 2025 (eff 1 Jan 2026) now mandate ≥10% risk retention.38

AI: Mostly Pilot Behind the Hype

RBI’s FREE-AI survey (Aug 2025): ~21% of REs use AI, 13.7% for underwriting, but only ~15% use interpretability tools. EY: 74% started PoCs, only 11% in production. Fraud-AI (NPCI MuleHunter.AI) live at 23 banks; agentic-UPI pilots (Razorpay+NPCI+OpenAI/Anthropic) run in closed user groups with no agent-specific regulation; for now NPCI’s platform rules set the limits, not a regulator. FREE-AI’s line: “accountability must remain human.”39

Systemic & Governance: The Single Point of Failure

One operator (NPCI) runs UPI, IMPS, RuPay, NACH, AePS, FASTag. Four major UPI outages in 14 months (incl. a 5-hour event, 12 Apr 2025; nationwide, 10 Feb 2026). RBI’s New Umbrella Entity scheme, meant to create a competitor, was shelved (2023); no live alternative exists. NPCI itself competes on its own rail via BHIM.

Governance cautionary tales: Paytm Payments Bank licence cancelled 24 Apr 2026; BharatPe/Grover settled (2024); Fino Payments Bank MD arrested (Feb 2026). Political economy: zero-MDR is statutory, populist and pro-trader; with 2029 approaching, “the government is taxing UPI” optics outweigh the shrinking ₹2,000cr subsidy: which is why a tiered MDR keeps being recommended but not enacted.

Net: the credit risk sits on bank balance sheets (good for fintech equity, watch the banks); AI is a multi-year capex story, not a here-and-now moat; and the rail’s single-operator design is an under-priced systemic and governance risk.
Gravitywell Research · Confidential26 / 33
06: Forces, Impact & the TradeESG, Economic Impact & Employment
Public Value · Sustainability

20ESG, Economic Impact & Employment

The policy case for the OS is not the equity story: it is jobs, GDP, leakage savings and inclusion. And ESG is moving from absent to mandated.

~250k→400k
Fintech jobs · 2025→2026E
~2%
DPI ecosystem · share of GDP
$39bn
DBT leakage savings (Aadhaar)
$2.6bn
Climate-tech funding · 2025 (Tracxn)
Economic Impact & Jobs B
  • 250k+ jobs (2025) → ~400k (2026E), +7.5% hiring; 7,500+ firms, 30 unicorns; Tier-2/3 cities ~35% of new postings.
  • DPI contribution 0.9% of GDP (2022) → 2.9–4.2% (2030E) (NASSCOM-ADL); ecosystem ~$100bn (~2% GDP, Redseer).
  • Aadhaar-DBT leakage savings ~$39bn cumulative; 500m+ brought into banking (majority women).
  • Digital economy crossed $200bn, trajectory to ~$1tn.
ESG & Sustainability B

SEBI BRSR / ESG-disclosure mandates now bind listed FIs. India climate-tech funding $2.6bn (2025, Tracxn; $12.8bn cumulative across 1,583 firms): note broader “cleantech” trackers scope it higher. DPI rails (UPI+Aadhaar+AA) increasingly carry green subsidies, ESG tracking and carbon-credit settlement. Financial inclusion is the headline impact lever: account ownership 50%→89% (2011→25), gender parity, 500m+ banked. Gaps: no standardised ESG metrics, thin MSME capacity, few fintechs publishing KPIs.47

For the Policymaker

The OS is a fiscal asset (jobs, GDP, leakage savings, inclusion) funded by a fragile mechanism (the ₹2,000cr zero-MDR subsidy covers ~11% of industry cost). The policy frontier: price the rail without taxing the citizen, mandate ESG/inclusion outcome reporting (not just access), and fund the infrastructure gaps (vernacular, feature-phone, elderly) that the market will not.31,33

Gravitywell Research · Confidential27 / 33
06: Forces, Impact & the TradeTechnology Roadmap & Disruptive Horizon
5–10 Year Outlook

21Technology Roadmap & Disruptive Horizon

Where the OS goes next, and what could disrupt it. The near term is monetisation of existing rails; the long term is AI-native, agentic and tokenised finance.

TABLE 21.1 · The technology horizon
HorizonWhat shipsInvestable signal
Near · 1–3yrCredit-on-UPI scale; ULI public launch; UPI international (Nexus, 20+ countries by FY29); AA monetisationLive
Mid · 3–5yrAgentic payments (Razorpay+NPCI pilots); AI underwriting mainstream (FREE-AI); CBDC programmability (purpose-bound); ONDC financial servicesPilot
Long · 5–10yrTokenised deposits / UMI; cross-border DPI rails; AI-native finance; quantum-safe securityEmerging
Disruptive Innovations to Watch
  • Agentic commerce: AI agents transacting on UPI Circle; NPCI’s FiMI LLM; no agent-specific regulation yet.
  • AI underwriting: alt-data + AA; the fraud-AI arms race (MuleHunter vs synthetic identity).
  • CBDC programmability: purpose-bound money (PDS, agri DBT) despite shrinking circulation.
  • Tokenisation: UMI for tokenised deposits/CDs in wholesale CBDC.
Disruptive Risks

A super-app consolidation that entrenches the foreign duopoly; a Big-Tech / AI-platform shift that moves the value layer above the rails; an AI-driven fraud escalation that outruns detection; and a data-localisation / geopolitical fracture (US Section 301) that fragments cross-border ambition. The OS’s openness is its strength and its disruption surface.

The 5–10-Year Arc

By 2030, expect: a monetised payments layer (tiered MDR), a scaled data-credit market (AA/ULI past the MSME gap), agentic and AI-native rails in production, and a still-listed toll-booth complex compounding: re-rated by the cycle, not the story. The disruption that matters is not a new app but control of the intelligence layer sitting on top of public rails the state will keep free.

Gravitywell Research · Confidential28 / 33
06: Forces, Impact & the TradeThe Trade: Positioning & Catalysts
Positioning · Catalysts · Playbook

22The Trade

The decision-ready synthesis: how to position, what dates to trade around, and what each stakeholder should do. Illustrative framing for professional investors: not investment advice.

Model PositioningFRAMING
▲
Core long · listed infra
CDSL, CAMS, BSE, MCX: own the moat; enter on market-cycle drawdowns, not at 60x.
OWN
▲
Private · profitable infra
Perfios, KreditBee; wealthtech. Operating profit, not 2021 marks.
OWN
◆
Option · MDR beneficiaries
PhonePe, Razorpay, Paytm: small, asymmetric, binary.
OPTION
▼
Avoid/underweight
IEX until coupling resolves; pure unsecured lenders; zero-MDR consumer apps.
AVOID
Catalyst CalendarDATED
SEBI base-expense framework1 Apr 2026 ✓
Co-Lending Directions live1 Jan 2026 ✓
Govt response to tiered-MDR2026 (watch)
Razorpay listing~end-2026
NPCI 30% cap deadline31 Dec 2026
DPDP CM obligations~Nov 2026
DPDP full complianceMay 2027
IEX DAM coupling go-liveappeal dismissed
PhonePe IPO re-filepaused (watch)
Tripwire KPIs · monitor
Unsecured retail GNPA>1.7% → derisk
Household debt / GDP>48% → derisk
Next RBI FSRDec 2026
Stakeholder PlaybookACTION
PE / VC

Underwrite to adjusted operating economics & reset marks. Demand unit economics & IRR-by-scenario, not valuations. Use GIFT City for structuring; model DPDP/DLG compliance as opex. Exit via the ’26–27 IPO window & the M&A consolidation wave.

Policymakers

Quantify the zero-MDR welfare-vs-fiscal trade-off; staff the DPDP Board; enforce or retire the 30% cap; harden NPCI’s single point of failure; measure inclusion & ESG outcomes per rupee; fund the exclusion-tail gaps.

Industry & End Users

Operators: own a layer (distribution or infra), not the balance sheet; partner on co-lending. Users: know the true cost behind “free,” verify lenders on the RBI DLA directory, use 2FA, assert DPDP rights, demand vernacular & feature-phone access.

Gravitywell Research · Confidential29 / 33
AppendixMethodology & Data-Quality
For The Technical Reader

APPENDIX · iMethodology & Data-Quality

Data-Quality Grades: used throughout
A Audited / primary

Listed filings, RBI/NPCI/SEBI releases. Lock decimals to source.

B Primary, estimated split

Real source, but a sub-component is apportioned or survey-based (incl. NPS).

C Modelled / illustrative

GW model (MDR, segment split, TAM, IRR, scenarios). Directional.

Point-in-time & FX

Data as of 1 July 2026. FY26 = year ended Mar 2026. USD at ₹94.7/$ (RBI ref ~₹94.4, 29 Jun 2026). Figures are point-in-time and may restate; prior values are preserved in the §Change Log, not silently overwritten.

Market sizing (§01)

TAM/SAM/SOM figures are mostly grade C: analyst variance is wide (fintech market $51bn vs TAM $2.1tn vs revenue $190–250bn; insurtech baseline ranges 5x). We show ranges, never a false-precise single number, and flag the ~10x TAM-vs-revenue trap.

The MDR model (§08)

₹360L cr UPI value × ~29% P2M-by-value (not 63% by volume) × large-merchant share 30–60% × MDR 0.2–0.3%, flowing ~85–95% to pre-tax profit. Output ₹6–16k cr/yr is grade C; the large-merchant split is the weakest input.

Valuation & returns (§08, §12)

Multiples are split-adjusted (CAMS 1:5, Angel One 1:10). P/E ranges are 3–5yr from secondary aggregators: trust the band, not the decimal. IRR-by-scenario is a GW model C, entry-multiple- and exit-window-dependent.

Customer data (§17)

NPS/Trustpilot are grade B: Trustpilot skews to complainants (selection bias); Paytm NPS 42 is RedSeer merchant survey. Consumer pricing (APRs) is indicative; credit-on-UPI / BNPL rates are not transparently disclosed.

Impact & ESG (§20)

Employment (250k→400k) and DPI-GDP share (0.9%→2.9–4.2%) are grade B (NASSCOM-ADL, ET, Redseer). Few fintechs publish ESG KPIs; figures are sector-level.

Unit economics (§11)

Reported vs adjusted matters: PhonePe and Razorpay losses are one-time ESOP/redomicile charges. Entity traps (Razorpay India vs consolidated; KreditBee group vs standalone) are footnoted in-table.

Frameworks (§02–03)

Five Forces, SWOT and PESTLE are GW analytical judgements synthesised from the sourced data in §06–20, not third-party scores.

Primary-source verification

Several RBI/PIB/MoSPI/NPCI primaries were access-limited at writing; load-bearing figures are corroborated across multiple secondary outlets quoting the same release. Verify decimals against primaries before onward use.

AI-assisted analysis

Drafting, synthesis and cross-source reconciliation were AI-assisted (LLM). Every load-bearing figure was human-verified against the cited primary release; where the latest release could not be confirmed, the last-known value is retained and dated rather than estimated. No figure is model-generated without a source.

Report confidence: Medium-High

Payments (NPCI), macro (RBI/NSO), listed comps (FY26 filings) and the June-2026 RBI FSR are grade A and current: high confidence. TAM / 2030 revenue, the MDR opportunity, co-lending FY26 AUM and IRR-by-scenario are grade C models: directional only. Net: Medium-High.
Gravitywell Research · Confidential30 / 33
AppendixChange Log & Data Vintage
Point-in-Time Discipline

APPENDIX · i-bChange Log & Data Vintage

Change LogPOINT-IN-TIME
Ver.DateWhat changed
v1.0–1.2Jun 2026Initial build: frameworks, 7-layer stack, comps, scenarios.
v1.3Jun 2026Added §01–03 sizing/frameworks, §10 value chain, §12 M&A, §17 customer, §20 ESG, §21 roadmap.
v1.41 Jul 2026Refreshed to 1 Jul cutoff: Jun-2026 UPI (22.72bn / ₹28.9L cr / 757m-per-day); May-2026 app share (PhonePe 46.26 / GPay 32.75, combined below 80% first time); RBI FSR June 2026 (household debt 41.3→45.5%, GNPA 2.1→1.8%); RBI Annual Report FY26 bank fraud (₹36,014→48,021cr); co-lending ₹1L cr (Apr-24)→₹1.1L cr (Mar-25); climate-tech re-scoped to $2.6bn (2025). Corrections: microfinance stress restated to MFIN PAR (was a 16% GNPA figure not in the source); demat growth reframed to the FY26 slowdown; IEX coupling timeline (APTEL dismissal, Feb 2026). Added: §08·ii liquidity/fair-value exhibit; expanded MDR sensitivity grid; §22 tripwire KPIs; confidence rating; AI-assisted disclosure; this change log & vintage table.
Published values are point-in-time. Where a source revised (e.g. RBI restated FY25 bank fraud to ₹32,803cr), the current print carries the new figure and the prior value is preserved here: history is never silently overwritten.
Data Vintage & FreshnessAS OF 1 JUL 2026
DomainKey metricAs ofGradeFreshness
Payments (§05)UPI volume/value/dayJun 2026ACurrent
App share (§05)PhonePe/GPay/PaytmMay 2026ACurrent
Macro (§03)GDP, repo, CPI, FXJun 2026ACurrent
Credit risk (§15)Household debt, GNPAJun-2026 FSRACurrent
Fraud (§16)Bank fraud FY26Mar 2026ACurrent
Listed comps (§07–08)FY26 filings, pricesJun/Jul 2026ACurrent
Funds (§07)MF AUM / SIPMay 2026A1 mo
Funding (§12)H1-2026 fintech VCJun 2026BCurrent
Microfinance (§15)MFIN GLP / PARQ4 FY26AQuarterly
Co-lending (§19)AUM (CRISIL)Mar 2025BStale ~15mo
Inclusion (§16)RBI FI-IndexMar 2025AAwaiting
Literacy (§16)Financial literacy2019 (NCFE)CNo newer survey
AI adoption (§19)RBI FREE-AIAug 2025A10 mo
TAM / 2030 (§01)Revenue forecastAug 2024CModel
Freshness is disclosed, not masked. Three items are knowingly stale: co-lending AUM (no FY26 print exists), financial literacy (no national survey since 2019; NFLIS 2025 in field), and the Aug-2024 TAM (predates the 2025–26 slowdown; §13 probability-weights it). RBI Ombudsman FY26 and the Mar-2026 FI-Index publish ~Aug 2026, after this cutoff.
Gravitywell Research · Confidential31 / 33
AppendixSource Register
Citations

APPENDIX · iiSource Register

Superscript numerals refer to the primary sources below, verified to the most recent publication available at writing (1 July 2026). Where a fresher release did not yet exist, the last-known value is retained and dated (see §Data Vintage).

1 GW India Stack framework; UIDAI; DigiLocker; MeitY DPI disclosures.

2 NPCI: UPI: 22.72bn/₹28.92L cr/~757m-per-day (Jun 2026, released 1 Jul via IANS/ANI wire); 228bn CY2025; app share Dec 2025 & May 2026 (PhonePe 46.26% / GPay 32.75%, combined below 80% first time). Cross-border FY26 ~1.49m txns/₹330cr (to Dec 2025); India–Nepal 2-way corridor Jun 2026.

3 IMF; RBI Annual Report; Worldline India Digital Payments (Jan 2026): P2M 63% vol / 29% value.

4 RBI Annual Report FY26: e₹ circulation ₹771.7cr (−24% YoY).

5 ONDC / Beckn: 450m+ cumulative; mobility ~56%; ~1.16 lakh active retail sellers.

6 iSPIRT: India Stack / OCEN architecture.

7 SIDBI × Crisil (13 May 2025): MSME addressable gap ₹30L cr.

8 RBI / RBIH: ULI ₹1.75L cr / 3.2m loans (GFF 3 Oct 2025); 64 lenders.

9 FY26 listed filings: CDSL, NSDL, CAMS, KFin, BSE, MCX, IEX, Angel One, Paytm, PB Fintech; demat ~22.5cr (SEBI).

10 Sahamati: AA: 474m consents (May 2026); 2.88bn enabled; 820+ FIs; ₹1.47L cr H1 FY26.

11 Sacra; Business Standard: Plaid: $8B valuation, $546m ARR, 12,000+ FIs.

12 IFSCA; Tribune: GIFT City IFSC banking assets $106.7bn (Feb 2026).

13 BCG × GFF, Building Bridges (Aug 2024): ~$190bn revenue by 2030; EY corroborates; ~$47bn (2025, Inc42).

14 Tracxn (Jan/May 2026): $2.4bn (3rd globally); valuations; cumulative 306 acq / 76 IPOs / 30 unicorns / $51.1bn VC.

15 Venture Intelligence (Jan 2026): $2.82bn 2025; payments $1.06bn / lending $752m / wealthtech $547m (+258%).

16 RBI FSR (June 2026, released 30 Jun): household debt 45.5% of GDP (Sep 2025); system GNPA 1.8% (Mar 2026); retail unsecured GNPA 1.7%; non-housing retail 58.4%. Unsecured 53.1% of slippages & fintech +36% (>70% unsecured) carried from Dec-2025 FSR (not re-stated in June).

17 MeitY: DPDP Rules 2025 (13 Nov 2025); Consent Manager wall; ₹250cr penalty.

18 RBI (Digital Lending) Directions 2025 (8 May 2025): DLG cap 5%; DLA directory.

19 RBI: risk-weight 125% (Nov 2023); Feb 2025 partial rollback.

20 MFIN Micrometer Q4 FY26 (57th ed., as of 31 Mar 2026): GLP ₹3.25L cr (+3% QoQ, first growth in 7 qtrs); PAR 31–180d ~2.0% (peaked ~6.3%, Mar 2025).

21 World Bank Global Findex 2025: ownership 89%; 16% inactive; FI-Index 67.0 (RBI).

22 PIB / DFS: PMJDY ~56.98cr accounts; ₹2.68L cr deposits; ~26% inoperative.

23 RBI Annual Report FY26: bank fraud ₹48,021cr (+46% YoY on restated FY25 base ₹32,803cr; value inflated by 314 legacy cases; 10,114 cases). RBI digital-payment fraud 293 cases/₹29cr. UPI-specific ₹805cr/10.64 lakh (Apr–Nov 2025, NPCI/Parliament: different dataset).

24 I4C / MHA: cyber fraud ₹22,495cr / 28.15 lakh cases (2025); digital-arrest −66%.

25 RBI Ombudsman Annual Report 2024-25: 13.34 lakh complaints (+13.6%).

26 NSO / MoSPI (5 Jun 2026): nominal GDP ~₹347L cr (~$3.9tn); real growth ~7.6% FY26.

27 RBI MPC (5 Jun 2026): repo 5.25%; CPI 3.93%; INR/USD ~94.7.

28 NIPL / IBEF: 9 live corridors; cross-border UPI ~755k txns / ~$29.5m (FY25).

29 SEBI / AMFI: F&O framework; NSE F&O −18% FY26; MF AUM ₹81.58L cr, SIP ₹30,954cr/mo; brokerage schedules.

30 Company filings & aggregators: P/E ranges, ROE, div yields; SEBI true-to-label (Oct 2024); CERC market coupling (Jul 2025).

31 Parliamentary Standing Committee on Finance (12 Mar 2026); Budget 2026 ₹2,000cr UPI incentive; PCI 0.3% proposal.

32 NPCI: UPI 30% cap, deadline 31 Dec 2026 (3rd deferral).

33 NCFE FLIS 2019 (financial literacy ~27%: no post-2019 national survey; NFLIS 2025 in field); Counterpoint 2025 (feature-phone ~350m); TRAI (Dec 2025, 1.258bn wireless); digital literacy ~37%.

34 BCB (PIX); ProMarket; EBANX (Nov 2025): ~7.3bn txns/mo; ~0.22–0.33% MDR; PIX Automático.

35 Caixin; CNBC; SCMP: Ant IPO halt (Nov 2020); $985M fine (2023); ~70% collapse.

36 European Commission COM(2023) 365; McKinsey (2025); Open Banking Ltd (Feb 2026): PSD2; UK VRP.

37 USTR: Section 301 Brazil determination (1 Jun 2026).

38 CRISIL (13 Aug 2025): co-lending AUM ~₹1.1L cr (Mar 2025); RBI (Co-Lending Arrangements) Directions 2025 (issued 6 Aug 2025, eff 1 Jan 2026, ≥10% retention); PhonePe/Paytm lending disclosures.

39 RBI FREE-AI report (13 Aug 2025); EY AIdea of India (Mar 2025); RBIH MuleHunter.AI; NPCI/Razorpay agentic-UPI pilots.

40 MCA / Tofler / Entrackr / DRHPs: FY25 unit economics: Razorpay, Perfios, M2P, Pine Labs, Juspay, KreditBee, Cred, PhonePe.

41 Inc42 State of Indian Fintech 2025; IMARC; Straits Research; Grand View Research: TAM/segment sizing & CAGRs.

42 NASSCOM-ADL (2024); Redseer; ORF: DPI GDP contribution; DBT savings; digital economy.

43 Nexdigm; Market Research Future: insurtech & wealthtech sizing (wide variance).

44 EY-CII Financial Inclusion (Dec 2024); RedSeer (NPS); Trustpilot; National Herald / RBI: UPI fraud, decline rates.

45 Fintechfutures; Inc42 (H1-2026 fintech funding ~$1.3bn, −19% YoY); Bloomberg/Reuters (Meta–CRED $900M @ $4.5bn, Jun 2026); M&A (Slice-NESFB, InCred-Stocko, Partners Group-Infinity: values largely undisclosed); Bain India VC 2026.

46 Zerodha; Groww; ProductGrowth: brokerage & MTF pricing; gateway success rates.

47 SEBI BRSR/ESG; Economic Times HR survey (jobs); Tracxn India Climate-Tech (Jun 2026): $2.6bn (2025), $12.8bn cumulative; DBT Mission / PIB: ₹3.48L cr cumulative savings (~$39bn).

Additional: Watal Committee (2017, NPCI governance); MediaNama (outage RCAs); KPMG India Fintech 2025; Deloitte (rural fintech).

Report GWR-2026-IN-001 · v1.4 · India Fintech Series · July 2026 · Institutional / For Professional Investors · Not investment advice
Gravitywell Research · Confidential32 / 33
AppendixGlossary & Acronyms
Decoding The Jargon

APPENDIX · iiiGlossary & Acronyms

Account Aggregator (AA / DEPA)
Consent-based open-finance framework letting individuals share financial data via licensed NBFC-AAs: Layer 3.
BRSR
Business Responsibility & Sustainability Report: SEBI’s mandatory ESG-disclosure framework for listed companies.
Consent Manager (DPDP)
A registered, data-blind intermediary managing user consent; under DPDP Rules it cannot also monetise the data: the ring-fence around AA economics.
DLG / FLDG
Default/First-Loss Default Guarantee: fintech cover to a lender, capped by RBI at 5% of the disbursed portfolio.
DPDP
Digital Personal Data Protection Act 2023; Rules notified 13 Nov 2025, phasing to May 2027; ₹250cr penalty ceiling.
Five Forces
Porter’s framework (rivalry, entry, supplier & buyer power, substitutes) used in §02 to score industry attractiveness by tier.
Market coupling
Routing all power-exchange bids through a single operator for price discovery: the reform threatening IEX’s moat.
MDR
Merchant discount rate: fee to accept a digital payment. UPI’s zero-MDR caps payment-app profitability.
MTF
Margin Trading Facility: broker-funded leverage for equity buying; ~15% p.a. interest.
NPS
Net Promoter Score: customer-loyalty metric (−100 to +100); used in §17.
P2M vs P2P
Person-to-Merchant vs Person-to-Person UPI. P2M is ~63% of volume but only ~29% of value: governs MDR maths.
TAM / SAM / SOM
Total Addressable / Serviceable Available / Serviceable Obtainable Market: the sizing funnel in §01.
ULI
Unified Lending Interface: RBI/RBIH credit rail pulling consented data for rapid underwriting.
UPI
Unified Payments Interface (NPCI): India’s real-time retail payments rail (Layer 2); the world’s largest by daily volume.

Institutional Abbreviations

ADTAverage daily turnover BNPLBuy-now-pay-later CAGRCompound annual growth rate CERCCentral Electricity Regulatory Commission DBTDirect Benefit Transfer DPDPDigital Personal Data Protection Act 2023 DPIDigital public infrastructure FSRRBI Financial Stability Report GTV/TPVGross / Total transaction value IRRInternal rate of return NPCINational Payments Corporation of India PESTLEPolitical-Economic-Social-Tech-Legal-Environmental RTARegistrar & Transfer Agent SAM/SOMServiceable available / obtainable market SWOTStrengths-Weaknesses-Opportunities-Threats TAMTotal addressable market TERTotal Expense Ratio UIDAIUnique Identification Authority of India ULIUnified Lending Interface
Gravitywell Research · Confidential33 / 33
Gravitywell ResearchInstitutional Investment Intelligence
End of Report

Own the moat.
Pay the right multiple.

India built the operating system and gave the rails away. Capital compounds at the regulated choke-points it kept, but they are listed, cyclical and richly priced. The new data rent is walled off as it forms; the citizen carries a risk tail; and the export story is, so far, diplomacy. Own the moat the state built; pay the right multiple; underwrite the rail, and the risk: it has not yet priced.

Gravitywell Research · GWR-2026-IN-001 · v1.4 · July 2026 Confidential · For Professional Investors