Priced for AI disintermediation its own segment data refutes · +165.1% upside
Market far below the MC band — P(below) = 0%Intrinsic value $86.48/share (after a 5% governance discount; pre-haircut $91.03) vs market $32.62. The Monte Carlo p5 tail still sits at $62 — every one of 1,000 correlated stress draws, β swept 0.70–1.30 and the AI-melt left tail fattened, lands roughly double today's price.
What it sells, where it sells
Operating segments
The whole thesis is a race inside one consolidated segment: can the ~24% ATS leg (+17%, >2× revenue/headcount) scale and lift the blend faster than agentic AI compresses the ~76% Core base? Margins and ROIC expanded 2023–25 as the AI mix rose — the empirical counter to the disintermediation short.
Country mix (revenue-weighted CRP input)
Weights are client/billing geography (71% North America, 22% Europe, 75% billed in USD per the 10-K) — not the India-58.7% service-delivery footprint, which is the cost base. So the weighted CRP is a near-zero +42bp add-on over the US mature ERP; India enters only as a 5% revenue tail, not the delivery 58.7%.
Background — five things to know before the case
Context the share price doesn't carry on its face. Skim this once and the bull / bear bullets below stop reading as inside baseball.
- Record 2025, yet de-rated ~30% YTD on a sector AI fear. Revenue +6.6% to $5.08B, GAAP operating margin at a decade-high 14.8%, adjusted EPS +11.3% (5th straight year EPS outgrew revenue). The stock trades at ~9× GAAP / ~8.6× adjusted EPS because the market prices BPO/IT-services as labor-arbitrage work AI will commoditize.
- The AI leg is winning, not losing. Advanced Technology Solutions (Data & AI, Digital, Advisory, Agentic) grew +17% to $1.204B (24% of revenue, >1/3 of new bookings), guided “at least high-teens” for 2026, at >2× revenue/headcount. HFS ranks Genpact a top-tier “GCC Orchestrator”; expanded Google Cloud agentic-finance alliance.
- Clean, dispersed ownership — Bermuda is a GE-spinoff tax legacy, not a controlled structure. Bain Capital fully exited; 96% institutional (BlackRock, Nalanda); insiders hold only ~1.58%. No controlling shareholder, no related-party drag.
- Share count down ~21.6% over a decade via buyback-and-cancel. 225.2M → 176.6M diluted; ~$2.46B repurchased since 2015, and the company kept buying into the 2025 AI-fear sell-off at ~$46 avg. A dividend initiated 2017 has grown ~10%/yr (4th+ straight double-digit hike); ~50% of cash flow returned.
- Disciplined, never-impaired M&A. ~25 tuck-ins over 20 years (Rightpoint, Enquero, Hoodoo), no mega-deal blow-ups and no material goodwill impairment despite $1.78B of goodwill. ROIC expanded to ~14–15% (from ~11% three years ago), comfortably above WACC.
Where we diverge from the sector
Each row asks: if we'd used the sector median for this one assumption instead of our override, how much would the share price change? Negative = our override is more conservative (worth less); positive = more aggressive (worth more). Skim down the rightmost column to see which bets are doing the work.
The story & the five claims
The 10-year story this DCF is built on, plus the five anchor claims that translate the story into engine inputs.
The 10-year story
Genpact is a disciplined, asset-light digital-operations and IT-services compounder that the market de-rated ~30% this year on a sector-wide AI-disintermediation fear — the worry that generative and agentic AI will commoditize the labor-arbitrage work that built the firm. Its own segment data says the opposite happened: as the high-value AI mix rose to 24% of revenue, operating margin and ROIC both expanded into 2025. The base case isn't a landslide — it's that the mix shift wins by a nose. Blended revenue compounds ~7% a year through year 5 (high-teens ATS lifting a low-single-digit legacy base), then fades to the 4.2% USD risk-free by year 10, taking revenue from $5.08B today to roughly $9.4B. GAAP operating margin grinds modestly from its 14.8% decade high toward ~16% as the richer mix earns its keep, reinvestment stays asset-light at 3.0× sales-to-capital, and the central risk — does agentic AI compress the 76% Core base faster than the 24% AI leg scales? — is priced in the Monte Carlo left tail, not waved away.
Two debates worth pressure-testing
- We discount at β 1.0, not the 0.62 regression. The multi-source 5Y regression β (StockAnalysis/Yahoo/SWS ~0.62) is a stale, pre-AI-fear calm-regime read. Using it to discount post-disruption flows is the core error. Council rebased to a market-consistent 1.0 — a higher discount rate that lowers our value vs the regression.
- Terminal ROC disciplined to 15%, not 36%. The margin triple (0.16 × 3.0 × 0.75) auto-resolves to a 36% perpetual return on capital — a fantasy franchise for a competed, AI-disrupted services firm. Pinned to ~current book ROIC: still value-creating vs ~8% terminal WACC.
- AI risk priced in the flows, not bolted on the end. The MC revenue (low 0.03) and margin (low 0.13) draws carry deliberately fattened left tails. The melt is the central fundamental call — modeled honestly, not waved away with a flat discount.
- Margin target capped at 16% GAAP, not the 17.5% adjusted. A GAAP target should not assume away the amort + SBC that the adjusted figure adds back. Sell-side anchors on the adjusted framing.
- S2C at 3.0×, below the 5.19× sector median. The Damodaran median ignores the goodwill that funds a capability-acquirer's growth; a fading, honest incremental efficiency is the conservative read.
- Agentic AI melt accelerates. The 76% Core base rolls over faster than the 24% ATS leg can scale; the blend falls below 4% and margins compress. This is the short thesis the market is pricing — modeled as the fattened MC left tail. Even so, the AI-melt bear floors near $39/share, still above today's $32.62.
- β is genuinely 1.3, not 1.0. If the Damodaran-global Computer Services bottom-up (β_lev 1.30) is the right read, WACC rises and value falls to ~$79/share — the bottom of the β-sweep, still >2× the price.
- Pricing power erodes as AI deflates services. If AI turns domain workflows into a race to the bottom, the ~16% margin grind reverses toward the low-double-digits and ROIC drifts toward WACC. The terminal-ROC discipline (15%) already leans against this.
- Terminal value is ~71% of EV. A large share of the value sits beyond year 10 — sensitive to the terminal growth (4.2% risk-free) and ROC (15%) assumptions. Both are deliberately conservative, but a long-duration DCF is always exposed here.
- SBC + routine insider selling leak cash to minorities. ~$90M/yr stock comp (~1.8% of revenue) and option-exercise insider selling are the leakage the 5% governance haircut captures. If they run hotter, the haircut is light.
Risks to thesis (tail, not bear case)
The thesis-breaker: AI commoditizes the 76% Core base faster than the 24% ATS leg scales. Priced as the fattened MC left tail; even then the bear floors near $39 — above the $32.62 price.
If the true β is the 1.30 bottom-up, not the 1.0 council rebase, WACC rises and value falls to ~$79. The full 0.62–1.30 range is swept in the MC β-axis — all of it clears the price.
~71% of EV sits past year 10, exposed to the 4.2% terminal growth and 15% terminal ROC. Both are deliberately conservative, but a long-duration DCF is always sensitive here.
Short interest ~10.9% of float (4.8 days to cover). A sentiment reversal could re-rate fast in either direction — a fundamentals call wrapped in a crowded short.
58.7% of delivery sits in India — a cost-base (not revenue) risk. Wage inflation or fiscal-incentive changes could pressure margin, but the decade record is of margin expansion.
~$90M/yr stock comp + routine option-exercise insider selling. Net share count is still falling via buyback-and-cancel; the 5% governance haircut covers the residual.
10-year forecast
Revenue $5.44B → $9.99B over 10y (~7% Y1-5 blend — high-teens ATS lifting a low-single-digit Core base — fading to the 4.2% USD risk-free terminal). GAAP operating margin grinds from 14.9% Y1 to 16.0% by Y10 — modestly above the 2025 decade high, short of the 17.5% adjusted.
Monte Carlo distribution
Even at the 5th-percentile outcome ($62) — with β swept to 1.30 and the AI-melt revenue/margin tails fattened — intrinsic value still exceeds today's $32.62 by ~90%. The disagreement isn't whether Genpact is cheap, but by how much.
Mean $80.74 ± $11.44/sh, 1000 iterations (0 failed). P(intrinsic < market $32.62) = 0.0%.
Cost of capital build
| Risk-free rate | 4.20% |
| Mature-market ERP | 4.23% |
| Levered β | 1.00 |
| Weighted CRP | 0.42% |
| Cost of equity | 8.85% |
| Pre-tax cost of debt (synth Aaa/AAA) | 4.60% |
| D / V | ~22% |
| WACC | 7.67% |
Full year-by-year DCF
| Year | Revenue | Op mgn | EBIT | EBIT(1−t) | Reinvest | FCFF | PV |
|---|---|---|---|---|---|---|---|
| 1 | $5.44B | 14.88% | $809M | $612M | $119M | $494M | $459M |
| 2 | $5.82B | 15.01% | $873M | $661M | $127M | $534M | $461M |
| 3 | $6.22B | 15.13% | $942M | $713M | $136M | $577M | $462M |
| 4 | $6.66B | 15.26% | $1.02B | $769M | $145M | $624M | $464M |
| 5 | $7.12B | 15.38% | $1.10B | $830M | $155M | $674M | $466M |
| 6 | $7.62B | 15.50% | $1.18B | $893M | $166M | $727M | $466M |
| 7 | $8.16B | 15.63% | $1.27B | $961M | $178M | $784M | $467M |
| 8 | $8.73B | 15.75% | $1.37B | $1.04B | $190M | $845M | $468M |
| 9 | $9.34B | 15.88% | $1.48B | $1.11B | $204M | $910M | $468M |
| 10 | $9.99B | 16.00% | $1.60B | $1.20B | $218M | $981M | $468M |
Methodology & flags
Damodaran FCFF DCF, 10y explicit + perpetuity, in USD. Risk-free
4.20% (local-currency government bond). Synthetic credit Aaa/AAA.
CRP 0.42% from revenue-weighted country mix × Damodaran 2026 CRPs.
Terminal ROIC faded to 15.00%; 5% governance haircut applied
post-DCF. Monte Carlo: 1000 iterations. Engine v1.0.0 · result:
/Users/valentin/Documents/notes_jiliac_labs/Finance/Damodaran/valuations/g/output/2026-06-07-result.json
- R&D cap: OFF · Lease cap: OFF · Failure: OFF · ESO: OFF
- Governance haircut: 5% applied post-DCF ($91.03 > $86.48)
- Sensitivity tornado: not run