Last Price
₹1,135
Jeff the Financier · Institutional Equity Research · Fallen-Quality Series
A 34%-ROE, net-cash, 100%+-FCF-conversion franchise at ~15.5x trailing earnings — a ~40% discount to its own 10-year median — with management buying back ₹18,000 Cr of stock at ₹1,800, 59% above today's price. The market is pricing structural impairment; the order book ($14.9B TCV) is pricing a recovery. We side with the cash signal.
Section 01
Infosys is a 34% ROE (≈28–30% ex-buyback), 40% ROCE, net-cash, 126%-FCF-conversion franchise trading at 15.5x trailing earnings — roughly the bottom decile of its own 10-year range (median ~23x) — with a 7.5% FCF yield, a 4.1% dividend yield, and a board that authorised an ₹18,000 Cr buyback at ₹1,800 in Nov-2025, a 59% premium to today's ₹1,135. The stock has round-tripped from a 35–40x "digital supercycle" darling to an 18–20x "AI eats services" pariah. The entire question is whether the AI/GCC threat is structural (then the cheap multiple is deserved) or cyclical (then this is a generational entry). Our read: the macro slowdown is cyclical, the AI-deflation is partly structural — but the market is pricing INFY as if the franchise is impaired, which the record $14.9B FY26 deal TCV (+28% YoY, 55% net-new) and the buyback signal do not support. Probability-weighted fair value ₹1,458 (+29%); we initiate BUY / ACCUMULATE, conviction 3/5 — high-conviction on value, deliberately tempered by an unresolved structural debate whose diagnostic is the Q1 FY27 print on 23 July.
This is the name in the basket most directly in the crosshairs of the broader AI-capability-buildout debate — capability compounding, "unhobbling" from chatbot to autonomous agent — precisely the mechanism behind the bear case: if AI coding agents collapse the marginal cost of software toward "the cost of electricity," the labour-arbitrage pyramid is the first business model to erode. But the same frame cuts the other way. That world is one of trillions in enterprise AI deployment, and someone has to integrate, govern and run agentic AI inside regulated BFSI/healthcare/government estates. Infosys's Topaz/Cobalt orchestration layer is a bet on owning that deployment toll-booth rather than being disintermediated by it. The investment question reduces to: does Infosys end up as the disrupted (the typist) or the deployer (the plumber of the AI build-out)? At 15.5x, the market has priced the first; we think the truth is in between, with a free option on the second.
Feature · Peak → Trough → Now
Infosys's arc is the cleanest "euphoria → capitulation" round-trip in the Indian large-cap universe: a 35–40x digital-supercycle darling that became an 18–20x "model-destruction" pariah, even as the order book hit a record. The gap between the multiple and the bookings is the opportunity — and the risk.
At the peak, sentiment over-extrapolated a pandemic demand pull-forward into a permanent 15% growth rate. At the trough, sentiment is over-extrapolating one OpenAI press release and a real-but-gradual AI-deflation into "model destruction." Both are emotional over-corrections; the truth is a low-to-mid-single-digit grower with a 7.9% shareholder yield. The mean sell-side target of ₹1,722 (+52%) is one of the widest price-to-consensus gaps in the stock's history — a measure of how far fear has run ahead of estimates. The reconciliation event is Q1 FY27 (23 July): hold/raise guidance and the structural thesis takes its first real hit; cut below 1.5% CC and even 15x is too expensive.
Section 02
Infosys is the #2 Indian IT-services franchise by revenue ($20.16B FY26), differentiated by best-in-class capital efficiency and the sector's highest revenue-per-employee (~$61K vs TCS ~$50K, HCL ~$48K) — the clearest quantitative fingerprint of AI-assisted productivity offsetting volume headwinds. The pandemic supercycle (FY22 +19.7% CC) gave way to an inventory-correction trough (FY24 +1.4%) and a stabilisation (FY25/26 +3–4% CC).
| Metric | FY21 | FY22 | FY23 | FY24 | FY25 | FY26 |
|---|---|---|---|---|---|---|
| Revenue (USD $B) | 13.56 | 16.31 | 18.21 | 18.56 | 19.28 | 20.16 |
| CC revenue growth | +6.1% | +19.7% | +15.4% | +1.4% | +4.2% | +3.1% |
| EBIT margin | ~24.5% | ~23.0% | 21.0% | 20.7% | 21.1% | 20.3% |
| PAT (₹ Cr) | ~19,423 | ~22,146 | ~24,108 | ~26,248 | ~26,750 | 29,474 |
| FCF / PAT conversion | ~95% | ~95% | ~95% | 102% | 145% | 126% |
| Revenue / employee ($K) | ~52 | ~57 | ~59 | ~59 | ~59 | ~61 |
| Metric | FY22 | FY23 | FY24 | FY25 | FY26 |
|---|---|---|---|---|---|
| ROE | ~28% | ~31% | ~29% | ~28% | ~34%* |
| ROCE | ~38% | ~40% | ~38% | ~38% | ~40% |
| ROIC | ~32% | ~33% | ~30% | ~32% | ~34% |
| ROIC − WACC spread | +19.5% | +20.5% | +17.5% | +19.5% | +21.5% |
Section 03
No formal SOTP applies to a single-segment IT franchise; the relevant decomposition is by vertical and geography, plus the TCS relative-value frame. BFSI (28%) is the swing vertical — a US-banking discretionary unlock is the single largest cyclical lever. Concentration risk is real: 88% of revenue is North America + Europe.
| Vertical (FY26) | % mix | Read |
|---|---|---|
| Financial Services | 28.0% | Above-avg; US bank IT recovering |
| Manufacturing | 15.9% | >2x co. avg; deal ramps |
| Energy/Utilities/Res. | 13.2% | Above avg |
| Retail | 12.8% | In-line/soft |
| Communications | 12.4% | >2x co. avg |
| Hi-Tech | 7.7% | Soft; highest GCC risk |
| Life Sciences | 7.3% | Stable/regulatory |
| Geography (FY26) | % mix | Read |
|---|---|---|
| North America | 55.7% | Core; discretionary cautious |
| Europe | 32.6% | Resilient; cost-takeout deals |
| Rest of World | 9.1% | Smaller, faster-growing |
| India | 2.6% | Govt/enterprise, recovering |
TCS trades at ~16.2x P/E and ~10.6x EV/EBITDA on a 25% EBIT margin and 48% ROE; Infosys at 15.5x P/E on a 20.3% margin. Historically Infosys traded at a 10–15% P/E discount to TCS to reflect execution-risk; today the gap is ~4%, while TCS's margin premium is widening. The relative-value case (re-rate Infosys to 18x on FY27E EPS ~₹80 → ₹1,440) is arguably stronger than the absolute case, because TCS — the higher-quality franchise — is barely cheaper.
| Company | Trailing P/E | Fwd P/E | EV/EBITDA | EBIT mgn | CC growth FY26 | Div yld | FCF yld |
|---|---|---|---|---|---|---|---|
| Infosys | 15.5x | ~14.5x | 10.6x | 20.3% | +3.1% | 4.1% | 7.5% |
| TCS | 16.2x | 14.1x | 10.6x | 25.0% | −0.5% | 4.9% | 6.0% |
| HCL Technologies | 18.0x | 16.1x | 10.2x | 17.2% | +3.9% | 4.6% | 6.2% |
| Wipro | ~16.4x | ~15.5x | ~10.1x | 17.3% | +4.0% | ~2.5% | ~5.5% |
| Accenture (ACN) | 12.8x | 11.0x | 7.4x | 15.7% | ~+5% | 5.0% | 13.1% |
| Cognizant (CTSH) | ~10.4x | ~9.5x | 6.2x | 16.0% | +4–6% | ~3.5% | ~8.0% |
Section 04
This section replaces the standard reverse-DCF narrative because, for Infosys, the entire valuation hinges on one variable: is AI a 3–5% annual revenue headwind forever, or a transitional disruption that ultimately expands the TAM? We frame both, then back out what the price implies.
Kotak: enterprise AI adoption cuts IT-services revenue ~3–3.5% p.a. for three years. HCLTech's CEO: contracts repricing 2–5% lower at renewal for equal scope. GCCs now rival Indian IT exports and insource the high-value work. If $10B of a ~$200B TAM is structurally at risk and Infosys holds ~10% share, ~$1B (~5% of revenue) is exposed. At 1–3% structural CC growth, fair multiple is 15–18x — i.e., not cheap; fair value ~₹820–900.
Emkay (Overweight, ₹1,650–1,750), Nomura ("major tech shifts expand the TAM"). $14.9B TCV at 55% net-new = clients signing new engagements, not just renewing. Three FY26 acquisitions add ~200–225bps to FY27 reported CC, so organic guidance of 1.5–3.5% understates momentum. Every enterprise adopting agentic AI needs integration, governance and legacy-modernisation — Topaz's addressable demand. Re-rate to 24x FY27E EPS → ₹1,920.
At CMP ₹1,135, WACC 12.5%, the reverse-DCF backs out essentially ~3% CC growth with margins holding ~20% and an ~11x exit EV/EBITDA — i.e., the market is pricing flat-to-modest real growth, not a collapse, but also not the structural-deflation doomsday (which would imply ~₹820). The consensus base (5.5% CC, 21% margin, 14x exit) implies ~₹1,520 (+35%). The asymmetry is moderately favourable: you are paid to wait via a 7.9% shareholder yield while the structural-vs-cyclical question resolves.
| Scenario | 5-yr CC CAGR | Terminal EBIT mgn | Exit EV/EBITDA | Implied value | vs CMP |
|---|---|---|---|---|---|
| Implied at CMP | ~3.0% | ~20% | ~11x | ₹1,135 | = CMP |
| Consensus base | ~5.5% | ~21% | ~14x | ₹1,520 | +35% |
| Bull | ~7.5% | ~22% | ~17x | ₹1,900 | +67% |
| Bear (structural AI) | ~1.5% | ~19% | ~9x | ₹820 | −28% |
Section 05
Three 12-month scenarios. The bear seriously models AI structural compression; the base assumes cyclical stabilisation; the bull assumes a BFSI unlock + AI-deal margin accretion. Probability-weighted fair value = ₹1,458 (+29%).
(0.20 × ₹820) + (0.55 × ₹1,480) + (0.25 × ₹1,920) = ₹1,458 (+29%). Note the buyback at ₹1,800 sits between base and bull — management has effectively planted a flag at a 24x-ish multiple. The distribution is right-skewed; investors would be underwriting a 20% chance of a −28% drawdown against a 80% chance of flat-to-+69%.
Section 06
Section 07
Rating: BUY / ACCUMULATE, conviction 3/5. Base target ₹1,480 (+30%); prob-weighted FV ₹1,458 (+29%); street mean ₹1,722 (+52%). Quant conviction score 61/100 — high on valuation, capital return and balance-sheet quality; docked for the genuinely unresolved AI-structural risk and soft near-term growth. This is a value-with-income play, not a multibagger call.
Section 08
Section 09