Coforge — a relay team that runs without an owner
Coforge Limited
Snapshot
Coforge is a mid-sized Indian IT services firm — the old NIIT Technologies, renamed in 2020 — that writes software and runs technology for big banks, airlines, insurers and travel companies (British Airways, ING, Sabre, SITA). Market cap ₹62,922 cr, share price ₹1,463, 52-week range ₹1,008–₹1,994, P/E 38.3, price-to-book 5.2×, return on equity 20.6%, return on capital 23.5%. In one phrase: a Great, asset-light franchise run by a hired relay team that has been sprinting for a decade — high returns, almost no debt, a fountain of cash — now priced like the market expects the sprint to continue.
As of 2026-06-21, from screener snapshot.
The verdict in two boxes — the business first, the price second
Box 1 — The business (durable):
| Lens | Result |
|---|---|
| Business-quality score | 18.5 / 23 (Quality 8.5/12 · Growth 6/6 · Longevity 4/5) |
| Buffett rubric | 6 / 10 PASS |
| Business bucket | Great (asset-light, high return on capital, free-cash machine) |
| Wealth-creator type | Enduring · Consistent (with one real asterisk: AI disruption) |
| Economic Profit | ≈ ₹700 cr (RoE 20.6% − CoE 12% on ~₹8,000 cr net worth) — clearly creating value |
A Great business and a genuine wealth creator — independent of what it costs today. The only thing that could un-make that verdict is the one thing nobody can yet measure: whether AI eats the work it sells.
Box 2 — The price today (a current phenomenon):
| Reading | Result |
|---|---|
| CMP | ₹1,463 (as of 2026-06-21) |
| Price pillar | 0.5 / 2 (PEG ≈ 1.7× · 5-yr payback ≈ 4.5×) |
| Margin-of-safety band | ₹765–₹955 (where PEG ≈ 1× / P/E ≈ 20–25×) |
| Mr. Market’s mood now | Fair-to-demanding — cooled from euphoria (down ~27% off the ₹1,994 high) but still paying a full price for quality + the fastest growth in its peer group |
| CMP vs the band | Demanding on the strict framework; fair-for-the-growth versus peers |
Today the market is charging a premium price for a premium grower. That’s a mood — born of Coforge being the fastest-growing large IT name in India and the AI-winner narrative — and it can de-rate next week without the business changing one bit.
In plain English
Imagine a company that has no owner. Most Indian businesses have a “promoter” — a founding family with their name on the door and their wealth in the stock. Coforge has none. A private-equity firm (Baring) bought it in 2019, cleaned it up, and sold every share by 2023. What’s left is a company owned entirely by mutual funds and foreign investors and run by a hired management team. So when you buy Coforge, you’re not betting on a founder. You’re betting on a relay team — and on the board that hires and watches them.
That team, led by chief executive Sudhir Singh since 2017, has done something rare. They took a sleepy ₹400-million-dollar firm and grew it almost seven times in ten years — revenue compounding ~22% a year, profit ~24% a year, for nine straight years. In the year just ended (FY26), revenue grew 29% in dollar terms — faster than almost any large Indian IT company. They did it by getting very good at a few industries (banking, airlines, insurance, healthcare) rather than trying to do everything, and by keeping their best people: staff turnover is the lowest in the industry at under 11%. That’s the boat — and it’s a good one. IT services need almost no factories or machines, so the business throws off cash: nearly ₹1,200 cr of free cash last year, and management now promises to turn 100% of profit into cash.
The moat, though, is real but narrow. Coforge keeps customers because switching software vendors is painful and because it knows their business deeply — not because it owns a patent or a brand nobody can copy. That has been enough to earn 20–30% on capital for a decade. The open question — the only one that matters — is whether artificial intelligence breaks the model. Indian IT has always sold people’s time: hundreds of engineers billed by the hour. If AI lets one engineer do the work of five, the whole industry’s revenue could shrink even as the work grows. Management is blunt about this — they say “labour as a default has been disrupted” — and they’re racing to sell outcomes and AI platforms instead of hours. So far they’re winning the narrative (25 analyst awards, a record order book). Whether they win the economics is a 3–5 year story nobody can score yet.
The big move right now is Encora — a ~$2.35 billion acquisition (closed April 2026) of an AI-led engineering and data-services firm, that bulks Coforge up for exactly this AI future. Here’s the catch that matters: Coforge paid for it almost entirely in its own shares — issuing 9.4 crore new shares (₹17,000 cr) to Encora’s sellers, who now own roughly a fifth of the whole company. That’s a big slice handed over, and it will dent return on equity in the near term (a much larger equity base for not-yet-much-larger profit). Management promises it’s profit-accretive within ~18 months. It’s the second big bet in two years; the first (Cigniti, a testing firm, ~$220m) worked beautifully — they nearly doubled its margins. The bull case is Encora repeats that. The bear case — and several brokerages agree — is that paying ~21× cash profit for a business growing slower than Coforge itself, with one-fifth dilution, is a rich price whose payoff is years away.
The tension in one line: this is a genuinely excellent business at a price that already assumes it stays excellent. At 38× earnings and ~5× book, you’re paying up for quality and growth that are real — but with no margin of safety if either the growth or the AI story stumbles.
Sitting down with the management
Here is the thing you must hold in your head the whole time you read this: there is no promoter. When Buffett sizes up a company he wants to know whether the people running it think like owners and eat their own cooking. At Coforge, the “owners” are index funds. The people running it own shares only through stock options. So the question isn’t “do I trust the founder?” — it’s “do I trust this hired team, and the board that holds the leash?”
On the team, the record speaks loudly and well. Sudhir Singh has run this since 2017 and, with a small core group (John Speight, Saurabh Goel, and a few others), delivered a nine-year revenue compounding of 21.7%, profit 24.1%, free cash flow 19% — and, more tellingly, kept the promises he made on these very calls. Five years ago, with COVID hammering the travel business, he told analysts Coforge would grow anyway — and it did. Two years ago, when they bought a 12%-margin testing company (Cigniti) and the market worried it would drag everything down, he said they’d fix it — and they did, lifting that unit to a 19% margin and tripling its two biggest clients. A management that says what it will do and then does it, again and again, is the single most valuable thing in this section. It is the closest a hired team gets to “skin in the game.”
The board is a genuine strength, not a rubber stamp. The chairman is O.P. Bhatt, the former chairman of State Bank of India — about as heavyweight and independent as Indian boards get. Anil Chanana (ex-HCL CFO) and other credible names sit alongside. Sudhir Singh was re-appointed for a clean five-year term to January 2030, so there’s continuity. Accounts are clean: reported profit turns into cash year after year (operating cash flow runs 90–110% of operating profit), the balance sheet isn’t quietly bloating apart from the goodwill that acquisitions naturally create, and the debt is small and conservatively managed (repaid bonds, funded from internal cash). The one screener “con” — a low tax rate — is explained honestly: software-export tax breaks plus, this year, a one-time ₹181 cr deferred-tax write-back from the Cigniti merger that flattered the headline profit. They told you about it on the call. That’s candour.
Now the two honest concerns. First, pay and dilution. A promoter-less company rewards its managers with stock options — and the more options issued, the more the pie is sliced away from outside shareholders. This isn’t abstract at Coforge: the CEO’s pay hit ₹105 cr in FY24 (the second-highest of any Indian IT chief that year), the bulk of it the value of stock options exercised — and the FY24 AGM asked shareholders to let his remuneration exceed the usual 5%-of-profit ceiling and to grant options worth over 1% of the company. (It fell back to ₹36 cr in FY25.) Most of this is shares, not cash — alignment, in theory — but it is alignment bought by diluting the people who already own the company. Stack on top the Encora deal, which handed ~20% of the company to outside sellers in one stroke, and you have a business where the share count is the number to watch like a hawk. This is the structural cost of the ownerless model. Second, key-man risk dressed as a strength. The “relay team” is really a star runner — Sudhir Singh is the Coforge story. The bench is deep and professional, but a promoter-less company leaning this heavily on one CEO is one resignation away from a re-rating. The board’s job is to make the company bigger than the man; it isn’t there yet.
Would Buffett and Agrawal shake hands on this management? Probably yes, with a raised eyebrow on pay. They’d admire the execution and the candour and the SBI-grade chairman; they’d dislike paying a hired team like owners while diluting the real owners. The one thing that would change their mind: a pattern of guidance misses, or ESOP grants that outrun performance. Neither has happened yet.
What’s on the horizon (live-issues tracker)
1. The crux — does AI eat the labour-arbitrage model? 🟡
This is the whole ballgame, so it gets the full interrogation below.
The crux in one sentence: Coforge is a wealth creator if and only if AI turns out to be a bigger tailwind (more transformation work to win) than headwind (fewer billable hours per rupee of client value).
The mechanism, in plain terms. Indian IT has always run on one engine: take a task a Western company did with expensive local staff, and do it with many more, cheaper, Indian engineers billed by the hour. The “moat” was a wall of trained people. Generative AI threatens the engine directly — if a coding agent makes one engineer 30–50% more productive (Coforge itself claims 25–35% productivity uplift in development, 40–60% in code generation), then the same client outcome needs fewer billed hours. Revenue per project can fall even as the number of projects rises. That’s the deflation fear, and it’s not theoretical — it’s arithmetic.
Test the analogy. Is this like the shift from on-premise software to cloud, which everyone feared would shrink IT services but actually grew it (someone had to migrate everything)? Partly. New technology waves have historically created more integration work than they destroyed. But there’s a difference: cloud changed where the work ran; AI changes how much human labour the work needs. Cloud didn’t reduce headcount-per-rupee; AI is explicitly designed to. So the comforting analogy only half-holds — which is exactly why this is “mixed,” not “fine.”
The named threat — and the proof it’s already moving.
| Who can take share | Armed with | The proof point (2025–26) |
|---|---|---|
| Accenture | Consulting front-end, scale | GenAI bookings ~$2.2bn in a single quarter (Q1 FY26), up from ~$100m two years earlier — openly “taking significant share” at the CxO layer Coforge wants |
| TCS / big-3 | Balance sheet, captive AI platforms | TCS cut >30,000 jobs in six months (an AI-led restructuring) while revenue grew — the clearest evidence the headcount-to-revenue link is breaking, and that scale players will pass the savings into pricing |
| AI-native / hyperscaler delivery | OpenAI/Anthropic/Microsoft/Google tools sold direct | Could disintermediate the simplest “build” work — no clean proof of Coforge-specific loss yet |
But the link isn’t uniformly broken: Infosys added ~13,000 staff over two quarters, branding itself the client’s “AI partner.” Even the giants disagree on whether AI shrinks or grows the labour base — which is exactly why this is unsettled.
The honest read: the near-term threat to a specialised mid-cap like Coforge is less “a startup eats me” and more “the big-3 use AI to compress prices at renewal across the whole industry.” Brokers are already pricing this — Kotak models ~3.5% pricing erosion showing up first at large-contract renewals, and sees sector growth stuck near 3–5%. Coforge’s defence is to climb into domain-heavy, outcome-priced work where the value sits in knowing banking/insurance/travel, not in writing code (its six-moats pitch: deep domain, client intimacy, outcome-priced “ModSquad” pods, the OneAI platform with 100+ domain agents, 11,000 AI-trained staff). Whether it holds the price line is unproven.
The precedent. The closest real-world precedent is every prior tech wave in IT services (client-server → web → cloud → mobile): each time, deflation fears ran ahead of reality and the incumbents that re-skilled fastest came out larger. The pool, too, is still growing — Gartner sees Indian IT services spend up ~11% in 2026. The counter-precedent is industries where automation genuinely shrank the labour pool (call-centre voice, basic testing). Coforge sits between: routine “build and maintain” work will deflate; “reinvent the enterprise with AI” work is a new, larger pool. Which dominates decides the next five years. Management’s own arithmetic, stated openly: a 25–30% productivity headwind over 3–5 years, which they expect to more than offset with AI tailwinds in existing (+20–30%) and new (+30–40%) services — and they’re targeting $5bn revenue by FY30 (~19% CAGR) on that bet.
Follow-on questions, answered. Is the damage to share or to the fee? Mostly to the fee (price per outcome at renewal), not yet to share — Coforge is still gaining share (top accounts +40% YoY). Which segment is hit first? The most rules-based, headcount-heavy work — insurance claims/underwriting and the BPO layer of banking — deflates first; Coforge’s crown-jewel Travel/Transportation (personalisation, dynamic pricing, agentic booking = new spend) and public-sector modernisation lean the other way. Is the incumbent’s response credible? Coforge’s is more credible than most — record order book ($1.75bn, +16% YoY), margins rising not falling (EBIT 10.7% → 14.4% → 16.6% exit), 25 analyst leader-recognitions, and Jefferies naming it a mid-cap top pick even while flagging AI deflation. A company being deflated does not usually post its best-ever margins. Has anyone actually switched away? No visible Coforge share loss yet. The honest verdict: not “too hard” — but genuinely mixed. The evidence so far (rising margins, rising order book, share gains) says Coforge is, for now, on the right side of the AI line. But “for now” is doing real work in that sentence, and a key vulnerability — rising client concentration (top-10 ~31% of revenue) — means one big renewal repriced down would sting. This is a 🟡 to re-check every quarter, not a settled 🟢.
2. The Encora integration 🟡 (early)
This is the second-biggest issue and a near-twin of the crux. Coforge announced (26 Dec 2025) and closed (23 Apr 2026) its acquisition of Encora — a US-based, AI-led engineering, data and cloud firm bought from private-equity owners Advent and Warburg Pincus for an enterprise value of ~$2.35bn (~$1.9bn equity). Encora brings ~$600m of revenue (FY26E), a ~19% margin, ~9,500 people including a 3,100-strong nearshore Latin-America base, and the new CTO on the calls. It’s Coforge’s biggest bet by far — and the way it was paid for is the whole story: an all-stock share-swap, 9.4 cr new shares at ₹1,816 (₹17,000 cr) handed to the sellers, who now hold ~20–21% of the enlarged company, plus a $550m three-year loan to clear Encora’s debt. So this single deal both dilutes existing owners by a fifth and adds interest cost.
How’s it going? Too early — it consolidates only from FY27 (~11 months). Management guides FY27 group EBIT to 15.5% with Encora vs 16.5–17% standalone (Encora drags margin down before synergies), with G&A synergies of 20–25% to close the gap, and calls the deal EPS-accretive within ~18 months. The bull case is earned: they did exactly this with Cigniti (~11% → ~19% margin, top clients tripled), so the playbook is proven. The bear case has named backers: Morgan Stanley calls it “bold” but near-term EPS-dilutive (and warns of a share overhang when the sellers’ lock-in ends); DAM Capital and ICICI Direct flag the dilution plus the $550m interest bill; the core worry is paying ~21× EBITDA / 3.9× sales for a business growing only ~7–10% organically — slower than Coforge itself. Watch the FY27 margin bridge and whether reported return on equity holds up against the much bigger equity base.
3. The margin step-up 🟢
Management has guided — and the Q4 exit rate already shows — a structural margin reset: FY27 EBITDA of 20.5–21%, driven by AI-automating their own back-office, SG&A leverage, and Encora synergies. This is the most concrete, near-term, checkable promise, and the Q4 FY26 print (EBIT 16.6%, a record) is delivering against it. 🟢 on track.
The watch-list:
- Revenue per employee / headcount-delinking: does revenue keep growing while headcount stays flat? (FY26: 35,777 staff, +436 in Q4 — still adding.) Falling rev/head would be the first hard sign of AI deflation.
- FY27 group EBIT lands ≥15.5% with Encora in — confirms the integration math.
- Order book stays > $1.7bn and large-deal count holds (21 in FY26).
- Share count: ESOP dilution per year (is it < 1.5%?) on top of the one-off ~20% Encora issuance.
- Top-account growth holds above ~25% — but watch client concentration (top-10 ~31%): a single large renewal repriced down would now move the needle.
- Any visible client loss to an AI-native competitor — none yet; watch for the first.
QGLP scorecard (the Motilal Oswal lens) — the receipts
| # | Question | Score | Evidence |
|---|---|---|---|
| Quality of Business | 4.0/6 | ||
| 1 | Large opportunity? | 1 | Global digital-transformation + AI-reinvention TAM is enormous; Coforge is only ~$2.8bn revenue inside it |
| 2 | Favourable industry structure? | 0.5 | IT services is competitive, but Coforge holds pricing discipline — OPM stable 14–18% for a decade |
| 3 | Defensible moat? | 0.5 | RoCE 19–32% for 12 of 12 years (numeric moat ✓), but the moat is narrow (switching cost + domain), and AI contests it |
| 4 | Return ratios > 15%? | 1 | RoE 20.6%, RoCE 23.5% — both well above 15% for a decade |
| 5 | Asset-light? | 1 | Capex light; FCF ₹1,197 cr FY26; the operating business swallows little capital |
| 6 | Negative working capital (Terms of Trade)? | 0 | IT services banks its customers — debtor days 88, positive working capital. Structural, not a flaw |
| Quality of Management | 4.5/6 | ||
| 7 | Unquestionable integrity? | 1 | Profit cash-backed (CFO/OP ~90–110%); clean audit trail; one-time tax gain disclosed openly |
| 8 | Proven execution? | 1 | 9-yr revenue CAGR 21.7%, hit guidance repeatedly — exceptional |
| 9 | Growth mindset & vision? | 1 | Fastest-growing large Indian IT; contrarian bets (COVID travel, Cigniti) paid off |
| 10 | Superior capital allocation? | 0.5 | Cigniti worked (12→19% margin); but equity-funded M&A + ESOP dilution + Encora unproven |
| 11 | Clear succession? | 0.5 | Deep professional bench, but CEO is the story — key-man risk in an ownerless firm |
| 12 | Minority interests protected? | 0.5 | Healthy historical payout (~43% median), but FY26 cut to ~9% to fund Encora; CEO pay ₹105 cr in FY24 + ESOP/Encora dilution (watch) |
| Growth | 6.0/6 | ||
| 13 | Structural tailwind? | 1 | Digital + AI transformation growing well above nominal GDP |
| 14 | Volume-led growth? | 1 | Wallet-share / volume led (29% USD growth), not price |
| 15 | Operating leverage? | 1 | EBIT 10.7% → 14.4% → 16.6% exit; margins expanding with revenue |
| 16 | Manageable leverage? | 1 | Near debt-free; borrowings ₹728 cr vs reserves ₹9,470 cr (D/E ~0.08) |
| 17 | Market-share gains? | 1 | Top-10 accounts +40% YoY; gaining share in a growing category |
| 18 | Earnings growth > 15%? | 1 | PAT CAGR 24% (9-yr), 28% (5-yr); forward ~18–22% |
| Longevity | 4.0/5 | ||
| 19 | Relevant for 10–15 yrs (low disruption)? | 0.5 | IT services endures, but the labour-arbitrage model faces real AI disruption |
| 20 | Can extend its moat (CAP)? | 0.5 | Domain specialism + low attrition help; AI pressures the edges |
| 21 | Can sustain growth runway (GAP)? | 1 | Only ~$2.8bn of a vast TAM; low wallet penetration |
| 22 | Headroom to diversify? | 1 | US/UK/EU/APAC; multiple verticals; Encora adds engineering + nearshore |
| 23 | Adaptive, resilient culture? | 1 | Proven through COVID, the turnaround, the AI pivot; lowest attrition in industry |
| Business-quality total | 18.5 / 23 | Quality 8.5 · Growth 6 · Longevity 4 | |
| Price pillar (separate) | 0.5 / 2 | PEG ≈ 1.7×, payback ≈ 4.5× — no margin of safety | |
| Canonical QGLP /25 | 19 / 25 | (noted for fidelity; the headline is the 18.5/23) |
The pillar pattern: Growth is flawless and Quality is strong; the two dents in the business score are both the same thing seen twice — the narrow, AI-contested moat (Q3, Q19, Q20) — and the ownerless-model governance cost (Q10–Q12). Price is the only pillar that fails outright. This is a textbook “wonderful business, unwonderful price” profile.
Buffett lens (the Berkshire-letters read)
| # | Test | Verdict | Evidence |
|---|---|---|---|
| 1 | Good boat (business > management)? | PASS | Great — asset-light, RoCE 23.5%, throws off free cash |
| 2 | Moat + franchise + pricing power? | PARTIAL | RoE > cost of capital 10/10 yrs; stable OPM — but narrow moat, AI-contested |
| 3 | See’s test (high returns, little capital)? | PARTIAL | Organic business is asset-light & FCF-rich, but growth has leaned on equity-funded M&A |
| 4 | Capital allocation (one-dollar test)? | PARTIAL | Cigniti created value; market value up 7× in 10 yrs — but recent dilution + Encora unproven |
| 5 | Owner-oriented, candid management? | PARTIAL | Remarkably candid, delivers guidance — but ownerless, with pay/ESOP scrutiny |
| 6 | Integrity / forensic (no “credit P&L, debit B/S”)? | PASS | Profit converts to cash; balance sheet clean ex-goodwill; one-time gain disclosed |
| 7 | Circle of competence / predictability? | PARTIAL | ”If there’s lots of technology, we won’t understand it” — AI makes the 10-yr picture genuinely hard |
| 8 | Mr. Market — gift or trap now? | PARTIAL | Off 27% from its high, but still 38× — not a fearful price |
| 9 | Patience / compounding runway? | PASS | Long runway, small share of a huge TAM, high RoE to reinvest into |
| 10 | The honest red flag | (see below) | AI deflation + a full price — the two things that could break the thesis |
Score: 6 / 10 PASS (3 full + 6 partials). A real business with two real gaps: predictability (AI) and price.
The See’s test, in numbers. See’s Candies took $25m to buy and only $32m of extra capital over 35 years, yet threw off $1.35bn. Coforge’s organic business has the same shape — it earns 23% on capital and barely consumes any, generating ₹1,197 cr of free cash on ₹16,403 cr of sales. The asterisk: unlike See’s, Coforge has chosen to also spend heavily on acquisitions (Cigniti, Encora) to grow faster, funded with equity and debt. So it’s a See’s-quality core wrapped in a more capital-hungry growth strategy. Pure See’s would be a PASS; this hybrid is a strong PARTIAL.
The one-dollar test. Has each rupee retained created at least a rupee of market value? Emphatically yes over a decade — book value compounded ~20% a year and the market value rose roughly seven-fold, far more than the earnings retained. The recent wrinkle is dilution: issuing new shares (for the QIP and to fund deals) means the per-share compounding is what matters now, and ESOP grants chip at it. Over ten years: a clear pass. Over the last two: keep watching the share count.
The framework metrics
- Economic Profit = Net Worth (~₹8,000 cr avg) × (RoE 20.6% − CoE 12%) = ≈ ₹690 cr of true profit above the cost of owners’ money. Solidly value-creating on trailing numbers — but a warning on FY27: the all-stock Encora deal adds ~₹17,000 cr of equity for not-yet-much-more profit, so reported RoE will compress (likely toward the low-to-mid teens) until Encora’s earnings ramp. The spread stays positive, but thinner; the EP engine pauses while the deal is digested.
- Terms of Trade = Debtors ÷ Creditors > 100% (debtor days 88, positive working capital). Not favourable in the FMCG sense — IT services fund their customers. Structural for the sector.
- 5-yr Payback = ₹62,922 cr market cap ÷ projected 5-yr cumulative profit (~₹13,800 cr at 18% PAT CAGR off a ~₹1,640 cr base) = ≈ 4.5×. Far above the < 1× multi-bagger signal — you are paying for the growth, not getting it free.
- PEG = P/E 38.3 ÷ ~22% forward growth = ≈ 1.7×. Above the < 1× discipline line.
- RoE − CoE spread = 20.6% − 12% = +8.6%; RoCE > 15% in 12 of 12 years shown — a durable, genuine moat by the numbers.
- Consistent vs Volatile = Net profit rose almost every year FY15→FY26 (₹122 cr → ₹1,745 cr), no fall > 10%, terminal ≫ initial. CONSISTENT — value it on P/E (which the market does).
CoE assumed 12% (Indian benchmark; studies use 10–15%). Forward growth 18–22% assumed, below management’s own trajectory, to stay conservative.
Peer comparison
| Company | Mkt cap (₹cr) | CMP (₹) | P/E | P/B | RoE | RoCE | OPM | Sales (₹cr) |
|---|---|---|---|---|---|---|---|---|
| Coforge | 62,922 | 1,463 | 38.3 | 5.2 | 20.6% | 23.5% | 18% | 16,403 |
| Persistent Systems | 76,177 | 4,829 | 39.4 | 9.7 | 27.3% | 34.4% | 19% | 14,748 |
| Mphasis | 43,280 | 2,268 | 22.9 | 4.0 | 18.5% | 22.8% | 19% | 15,880 |
| Hexaware | 30,526 | 500 | 20.8 | 4.9 | 24.9% | 30.1% | — | 13,836 |
What it says: Coforge sits at the top of the cohort on P/E (38×), level with Persistent and well above Mphasis/Hexaware (~21–23×). But the absolute P/E read flips when you look closer: Persistent is dearer on price-to-book (9.7× vs Coforge’s 5.2×) and earns a higher RoCE, while Mphasis and Hexaware are cheaper but growing far slower. Coforge’s distinctive edge is the fastest revenue growth in the group (29% USD) plus a sharply rising margin — it’s the growth, not the returns, you’re paying the premium for. So: expensive in absolute terms, but reasonable for a sector-allocator choosing the fastest grower in premium mid-cap IT. The patient value-investor and the sector-rotator will honestly disagree here, and both are right about different questions.
Latest quarter & what’s happening now
Q4 FY26 (reported 5 May 2026). Revenue grew 21.2% YoY in USD; the full year grew 29.2% — exceptional for the sector. EBIT margin hit a record 16.6% (vs 12.3% a year ago). Reported PAT of ₹666 cr for the quarter was flattered by a one-time ₹181 cr deferred-tax reversal from the Cigniti merger (reported tax rate −7%; normalized ~22%) — so read the underlying profit, not the headline. Order book at a record $1.75bn (+16% YoY); 21 large deals signed in FY26; attrition 10.8% (industry-low); free cash flow 110% of PAT. Management guided FY27 to a structural margin step-up (EBITDA 20.5–21%) and raised its through-cycle FCF/PAT target to 100%. (HARD: results, order book, guidance per the Q4 FY26 call.)
Live catalysts: Encora consolidation from FY27 (MEDIUM — guidance); the margin reset playing out quarter by quarter (MEDIUM); continued large-deal momentum (HARD — $648m Q4 intake).
Where the two lenses agree — and disagree
They agree on the spine: Great boat, exceptional execution, value-creating economics, and a price that offers no margin of safety. QGLP scores the business 18.5/23 and flunks only Price; Buffett buckets it Great and also fails it on Mr. Market.
They diverge in tone on one test, and it’s the important one. The QGLP checklist, being a scorecard, rewards Coforge’s flawless growth and rising margins and lands at a near-textbook 18.5. Buffett’s lens — which explicitly distrusts technology bets (“if there’s lots of technology, we won’t understand it”) and weights predictability heavily — drags the same company down to 6/10 by docking it on circle-of-competence. That gap is the signal: the numbers say “wonderful compounder,” the temperament says “wonderful, but I genuinely cannot tell you what this looks like in ten years because of AI.” Trust the Buffett flag as the risk warning and the QGLP score as the quality reading — they’re not in conflict, they’re answering “how good is it?” and “how sure can I be?” respectively.
The price as a current phenomenon
This judges the price, not the business — the business verdict above is already settled.
The margin-of-safety band. The framework’s arithmetic is unsentimental: to satisfy QGLP’s Price pillar (PEG ≈ 1× at ~20% growth, i.e. P/E 20–25×), the stock would need to trade around ₹765–₹955 on today’s earnings power (₹1,640 cr). That’s 35–48% below the current ₹1,463. A 5-year payback under 1× is not remotely available here. So on the strict value lens, this is demanding — you’d want a sector-wide recession or an AI scare to hand you that band.
One fairness caveat on the headline 38×: that multiple divides the post-Encora market cap (the ~20% extra shares are already in it) by pre-Encora earnings (FY26 closed before Encora consolidated). Once Encora’s profit lands in FY27, the denominator grows and the forward P/E compresses meaningfully — likely into the high-20s/low-30s. So the stock is genuinely less expensive on a forward basis than the trailing 38× suggests; the margin-of-safety band above is the patient-buyer’s floor, not a forecast.
Mr. Market’s mood. The crowd was greedy at the ₹1,994 high (the AI-winner, fastest-grower narrative in full bloom). It has since cooled ~27%, so the mood is now fair-to-demanding rather than euphoric — but it is not fearful. At 38× earnings you are being asked to pay for continued excellence, and the market is currently confident it will get it. That confidence is a mood: a broad IT de-rating, one soft quarter, or a louder AI-deflation scare could re-price it toward the band above without changing a single thing in the franchise.
The tension, plainly. Coforge is a wonderful business sitting at an unwonderful-to-fair price. It is not the bruised-blue-chip bargain setup; it’s the opposite — a high-quality grower priced like one. The reading can flip next week. The business won’t.
Conviction texture
The bull case, at its strongest. This is the best-executing mid-cap IT franchise in India, run by a team that has hit its numbers for nine straight years, gaining share, with a record order book, margins structurally stepping up, almost no debt, and a credible — analyst-decorated — AI strategy that is winning deals rather than losing them. If AI proves a tailwind (more transformation spend) for the domain-led work Coforge specialises in, this compounds at 20%+ for years and today’s 38× looks cheap in hindsight. The Cigniti playbook proves they can buy, fix, and cross-sell; Encora is the next leg.
The bear case, at its strongest (Buffett’s mandatory red flag). You are paying 38× earnings and 5× book for a business whose century-defining risk is unmeasurable. The entire Indian IT model rests on billing human hours, and the company’s own slides admit AI can do the work with 25–30% fewer of them — while a peer (TCS) has already cut 30,000 jobs and brokers model price erosion landing first at renewals. The defence — climb to outcome-based, domain-heavy work — is exactly what every competitor including the deep-pocketed big-3 and Accenture is also attempting, and the routine work will deflate. Layer on: an ownerless governance model that paid its CEO ₹105 cr in one year while diluting the real owners; a $2.35bn, all-stock Encora deal that handed ~20% of the company to outside sellers and will depress return on equity until it ramps; rising client concentration (top-10 ~31%); and a one-time ₹181 cr tax gain dressing up this year’s headline profit. That’s a full price with no margin of safety against a real — if unhurried — structural threat, on a balance sheet mid-digestion.
What the numbers actually support. The quality is not in doubt — 12 years of 20–30% returns on capital, cash-backed profit, and the sector’s best growth are facts, not hopes. The AI threat is, so far, not showing up in the numbers (margins and order book are rising, not falling), which tilts the evidence toward “Coforge is on the right side of the line — for now.” The price is the clear weak point: excellent business, full valuation, zero margin of safety.
Three things to watch that would tip it: (1) revenue de-linking from headcount — the first hard sign AI is deflating the model (bullish if rev/head rises, bearish if revenue stalls while headcount holds); (2) the FY27 margin bridge landing at guidance with Encora in; (3) any de-rating toward the ₹765–₹955 band, which would convert a wonderful business at a full price into a wonderful business at a fair one.
No buy/sell/hold — the reader decides.
Sources
- Screener.in — Coforge Ltd (consolidated): https://www.screener.in/company/COFORGE/consolidated/ — ratios, P&L, balance sheet, cash flow, shareholding (fetched 2026-06-21).
- Coforge Q4 FY26 earnings call transcript, 5 May 2026 — margins, order book, Encora, AI strategy, deferred-tax reversal.
- Coforge Q3 FY26 earnings call transcript, 23 Jan 2026 — Encora “defining moment”, order book +30%, AI market-context commentary.
- Coforge Annual Report FY25 — board composition (Chairman O.P. Bhatt, took over 29 Jun 2024), CEO re-appointment to Jan 2030, capital management, audit trail.
- Peer snapshots (screener.in, 2026-06-21): Persistent Systems, Mphasis, Hexaware Technologies.
- Encora deal: Coforge press releases (announce 26 Dec 2025, close 23 Apr 2026); Bloomberg ($2.35bn EV, share-swap); deal financials ~$600m rev / ~19% EBITDA / ~9,500 staff (BusinessToday); analyst caution (Morgan Stanley / DAM / ICICI Direct via Moneycontrol). Tagged HARD on terms, SOFT/MEDIUM on multiples & synergy.
- CEO compensation: FY24 ₹105.1 cr / FY25 ₹36.1 cr (Trendlyne; Business Standard “highest-paid IT CEOs FY24”); FY24 AGM remuneration/ESOP resolution (Coforge AGM Notice 2024). Exact proxy-advisor (IiAS/SES) recommendations and AGM dissent % not independently confirmed.
- AI-disruption crux: Kotak (~3.5% renewal pricing erosion, Apr 2026); TCS −30,000 jobs (PeopleMatters, Jan 2026); Infosys +13,000 staff (SEC 6-K, Q3 FY26); Accenture GenAI bookings ~$2.2bn/qtr (ConstellationR); Gartner India IT services +11.1% 2026; Jefferies mid-cap pick. Coforge $5bn-by-FY30 target & AI productivity math (Q4 FY26 transcript / BusinessToday interview).
- Assumptions: Cost of Equity 12%; forward PAT growth 18–22% (below management’s own ~19% FY30 trajectory, deliberately conservative). One-off ₹181 cr deferred-tax gain normalised out of FY26 earnings power.