Genpact Limited G

Genpact Limited (NYSE: G) — deep-value analysis

Analysis date: September 8, 2026 · Reference price: … (09/07/2026) · Market cap: … USD · Shares outstanding: 168,028,805 (06/30/2026) · Average diluted shares Q2 2026: 170.4 mil · Single currency: USD reporting / USD trading, no FX exposure in the valuation

Primary sources: 10-K FY2025 (filed 26.02.2026, CIK 0001398659), 10-Q Q1 2026 (08.05.2026), 10-Q Q2 2026 (07.08.2026), DEF 14A (11.03.2026), SEC EDGAR companyfacts (XBRL), the Q2 2026 earnings release (06.08.2026), data pack data-pack-G-20260908.md. Market figures from yfinance and stockanalysis.com at 07.09.2026.

Why the analysis is being redone. The mechanical triage deep_delta.py decided FULL because a claim underpinning the valuation broke: operating_margin. This report checks the claim from scratch and answers in two layers — one technical (the falsifier was poorly constructed) and one economic (the margin really is compressing, but where no one was looking). Meanwhile the price rose from … (14.08.2026) to … (09/07/2026), i.e. …, which alone consumes two-thirds of the safety margin estimated in August.

---

Executive summary

The thesis in brief. Genpact is a former GE captive that became the third-or-fourth-largest global BPM player, with … TTM revenue, 14.83% GAAP operating margin, 18.2% ROIC (39.6% ex-goodwill), and a normalized free-cash-flow yield of 7.9% at today's price. The market pays …× TTM GAAP profit and 7.95× EV/EBITDA (excluding operating leases) for a business growing 6.9% a year and returning 6.9% of market cap annually gross (4.94% buybacks … dividend), with net debt of 0.76× EBITDA. At … the stock is 23.3% below the 52-week high of … and 39% above the low of …

What's newly known versus the August analysis, and why it matters. Three things, all verified in the filings:

1. The margin is compressing at the segment level, not the group level. In Q2 2026, the combined adjusted operating income (AOI) of the three reportable segments grew 0.2% (… vs …) on revenue growing 7.1% — i.e. segment AOI margin fell from 18.30% to 17.12%, minus 118 basis points. The group, however, reports AOI … and a margin of 17.4% (vs 17.3%), because the "unallocated corporate expenses" line swung from −… to +…, a … swing. The company explicitly states what's in that line: "an adjustment to allowances for credit losses." The credit provision was … in H1 2026 versus … in H1 2025. All of the group's reported margin increase comes from a smaller receivables expense, not operations. Two of three segments have declining margins: Consumer and Healthcare −228bp, High Tech and Manufacturing −137bp.

2. Unbilled work is exploding, and no one is watching it. Classic DSO analysis looks at receivables. At Genpact, the item that's growing isn't the receivable, it's "deferred billings" — work performed but not yet billed, parked in "Other assets." Gross balance: … (Dec. 2024) → … (June 2025) → … (Dec. 2025) → … (June 2026). … per year, on revenue of …. Combined with net receivables, total collection exposure rises from … (June 2025) to … (June 2026), …, and extended DSO moves from 104.9 to 114.2 days — and to 119.5 days if you add back the … of receivables sold through factoring and derecognized from the balance sheet.

3. The company is voluntarily cutting its own Core Business Services revenue. At the August 6 earnings call, management guided CBS "flat to slightly down" in Q3 2026, explaining it's voluntarily exiting contracts "not aligned with the Agentic Operations strategy," with the impact concentrated in H2. This changes the interpretation: CBS's deceleration from … (2025) to … (Q2) and then to zero is no longer just suffered cannibalization, it's also chosen pruning. It's better news for margin and worse news for 2027's top line than the previous analysis assumed.

The FCF bridge and the valuation base. Reported TTM free cash flow (simple FCF …, FCFE …) contains … USD of net growth in contract liabilities — customers' money, not Genpact's profit, which management itself qualifies as unguaranteed. The owner-earnings base I built the valuation on is …, versus 510 in August. The resulting yield, 7.89%, compares to 9.12% published by stockanalysis.com — my normalization is 13.5% more conservative than consensus, not off by an order of magnitude. The figure passes the external check.

Estimated value. Five independent models give a range from … (DCF bear) to … (DCF bull), with a median of … and a central intrinsic value of …/share. The Monte Carlo simulation over 20,000 scenarios gives a median of … MOS, a … probability of undervaluation, and a P10-P90 range from … to …. The 11-analyst average consensus target is … (range 31-58), so sell-side sees 13% more value than I do.

Verdict: HOLD IF YOU ALREADY OWN, DON'T BUY AT …. In August, at … the thesis supported a first tranche. At … it no longer does: the median fell from … to …, the probability of undervaluation from …% to …%, and the model that assumes no growth at all (EPV) gives …, i.e. today's price is exactly the current earnings power capitalized. Nothing broke in the business — the safety margin got consumed, and in parallel two weaker quality signals were added (segment margin, unbilled work). Re-entry threshold: below …, where the median returns above 25% and EPV leaves 16%. Threshold to add to an existing position: the Q3 reporting (November 5, 2026) with 9M cumulative CFO above … and net receivables + unbilled work under …. Liquidation threshold: 9M CFO below …, or segment AOI margin below 17.0% for two more quarters.

---