Hexaware — a cash machine the PE owner is selling
Hexaware Technologies Ltd
Snapshot
Hexaware writes software and runs the back-office technology for big Western companies — mostly American banks, insurers, and a couple of giant US mortgage agencies. You hand it a chunk of your IT work, it does it cheaper from India, and bills you for it. It is asset-light (almost no factories or machines), throws off a lot of cash, and earns a high return on the money tied up in it. Market cap ₹30,526 cr, share price ₹500 (52-week range ₹900 / ₹400 — so it sits near the bottom), P/E 20.8, price-to-book ~4.9×, return on equity 24.9%, return on capital 30.1%, dividend yield 2.3%, and effectively no debt. As of 2026-06-20, from the screener snapshot.
What kind of animal it is: a Good business with Great-business plumbing — fat returns on tiny capital, gushing cash — but no real pricing power, and right now a bruised, freshly-relisted private-equity asset whose owner (Carlyle) still holds three-quarters of it and will keep selling.
The verdict in one box
| Lens | Result |
|---|---|
| QGLP score (Motilal Oswal) | 15 / 25 — Quality 8.5/12 · Growth 3/6 · Longevity 2.5/5 · Price 1/2 |
| Buffett rubric | 5.5 / 10 PASS |
| Business bucket | Good (Great financial DNA — asset-light, 30% RoCE — but no franchise/pricing power, real disruption risk) |
| Wealth-creator type | Enduring track record · Consistent earner — with a live question over whether AI shortens the runway |
| Economic Profit | +₹815 cr/yr (RoE 24.9% − CoE 12% on ₹6,316 cr net worth) — genuinely creating value |
| Margin-of-safety price band | ₹380–₹460. CMP ₹500 is fair on the quality, slightly demanding on near-term numbers, but cheap vs its own peer group |
A Good business with excellent cash economics, that IS a wealth creator — currently priced fairly-to-slightly-rich on its own arithmetic, yet cheap next to its sector. The two questions that cap it are not financial: who ends up owning it, and what AI does to the work it sells.
In plain English
Imagine you own a very good plumbing firm. It needs almost no equipment, it collects its bills faster than it pays its own, and for every ₹100 of owners’ money left inside it, it earns about ₹25 a year. That’s Hexaware. Over the last decade its sales grew from ₹2,582 cr to ₹13,430 cr, profits more than quadrupled, and it converted essentially all of that profit into real cash. On the numbers alone, this is a lovely little compounding machine.
But three things keep it from being a slam-dunk, and you should look at each square in the eye.
First, who owns it. Hexaware has had no founder at the wheel for thirteen years. It was taken private by one private-equity firm in 2020, sold to another (Carlyle) in 2021, and floated back onto the market in February 2025 at ₹708 a share. Carlyle sold its own shares in that IPO — not a single rupee went into the company — and it still owns about 74%. Worse, in November 2025 Carlyle pledged that entire stake as collateral for a $1.25bn loan. A private-equity fund holding three-quarters of a company, borrowed against, will sell down. That’s a permanent weight on the share price until they’re gone — and nobody knows the timetable.
Second, the price fell for a reason. The stock is down ~44% from its high because in early February 2026 the company reported a soft quarter and — the real shock — guided next year’s profit margin down, to 13–14%. Growth, which had been running at 16% a year, slowed to a crawl (~1% in the last quarter). Management says this is a specific problem: one of its two biggest clients, a US mortgage agency, slashed its budget, and there were seasonal furloughs. That’s plausible — and fixable. But it left a stock priced for steady growth suddenly looking expensive on the spot numbers.
Third — and this is the one that really matters — what AI does to this kind of work. Hexaware’s pitch is “AI for IT operations: leave it to us.” The trouble is that running IT operations and processing back-office tasks is exactly what AI automates best. The CEO himself admits AI is a “20–40% dampener” — meaning if a client renews the same work, it now needs 20–40% fewer people, so Hexaware bills less. He argues AI also opens big new work (rebuilding old systems, replacing expensive software). Both are true. Which one wins is genuinely unknown today, and it is the whole ballgame for the next decade.
So: a genuinely good, cash-rich business, run by a competent long-tenured CEO, sitting at a not-crazy price — wrapped inside a private-equity exit and pointed straight at the most uncertain technology shift its industry has ever faced. The quality is real. The questions are real too.
Sitting down with the management
If Buffett and Raamdeo Agrawal sat across from this team for an afternoon, they’d come away with respect for the operators and a furrowed brow about the owners — and those are two different things here.
The operator. R Srikrishna — “Keech” — has run Hexaware since 2014. IIT Madras, IIM Calcutta, ex-HCL. That’s eleven years of continuity through two private-equity owners, which is rare and valuable: the captain didn’t change when the ship was sold twice. On his watch revenue roughly doubled, headcount climbed to ~34,000, and — a detail Buffett would notice — voluntary attrition is ~11%, below an industry that often runs 13–20%. People don’t leave well-run firms in droves. On the February 2026 call he did the thing honest managers do and bad ones avoid: he guided margins down in plain sight rather than dressing it up. When a J.P. Morgan analyst bluntly asked “why shouldn’t we conclude AI is hurting you more than your peers?”, he answered the question instead of dodging it. That’s the candor test, and he mostly passes it.
The owner is the problem, and it’s structural. There is no founder-promoter. The “promoter” today is CA Magnum Holdings — a Carlyle vehicle. Three facts tell the whole capital-allocation story of the private years. One: dividend payout spiked to 75% in 2022 and 53% in 2023 — precisely the years Carlyle’s holding company was carrying ~$1bn of buyout debt that those dividends helped service. Two: the February 2025 IPO was 100% an offer-for-sale — Carlyle cashing out, nothing for the company. Three: in November 2025 Carlyle pledged its 74.55% stake against a $1.255bn refinancing. None of this is illegal or even unusual for private equity. But it tells you what this asset has been for: returning cash upstream to its owner. You are a passenger in someone else’s exit.
The governance plumbing — mixed. On the good side: the auditor is B S R & Co. (KPMG network) with a clean, unmodified opinion; the chairman is independent; there’s no SEBI run-in, restatement, or pending mess. On the cautious side: roughly half the board are Carlyle nominees, one of whom sits on the audit committee; the company leans on “adjusted” and “normalized” margins more than its larger peers do (always read those twice); and this past quarter carried an earnout reversal, an impairment, and a bumped-up bad-debt provision — all disclosed, none alarming on its own, but a reminder to watch the cash, not the adjusted slide.
The forensic red-flags — mostly quiet. Reported profit turns into real cash year after year (operating cash flow has run ~100–110% of operating profit for a decade) — the single most reassuring number in the whole file, and the opposite of the “credit the P&L, debit the balance sheet” trick. Receivables are under control (debtor days 56). The one flag that genuinely fires is promoter pledging — the entire stake is encumbered — plus the steady, certain future selling that comes with any PE owner.
Would Buffett and Agrawal shake hands on this management? On the operators, probably yes — competent, candid, low-drama, good with cash. On the ownership structure, they’d want to wait: you can’t fully judge a company’s stewardship while a leveraged financial owner is still heading for the exit. The one thing that would change their mind: Carlyle finishing its sell-down to a normal level, the pledge released, and a stable institutional shareholder base taking its place. Until then, the asset is excellent and the cap table is a question.
What’s on the horizon (live-issues tracker)
Three live threads will decide the next one-to-three years.
1. AI — the crux (the full interrogation is below). 🟡
2. The Carlyle overhang. 🟡 Carlyle owns ~74% and has pledged it against debt. The plain mechanics: a private-equity fund’s job is to sell, not to hold forever, so a large block of stock is certain to come to market — the only unknowns are timing, price, and whether it’s done in tidy institutional blocks or messy dumps. Every block sale caps the price near-term. The bull reframe: each clean placement to long-only funds actually improves the shareholder base and removes the very overhang that’s depressing the multiple — so the overhang is self-curing, just slowly. Watch for: block-deal filings, any change in the pledge, and the promoter stake ticking down from 74%.
3. The two-GSE client concentration. 🟡 Hexaware’s top-ten clients are ~36% of revenue, and its two biggest are large US housing-finance agencies (GSEs). One of them cut its budget hard in 2025 — that single client is ~70 basis points of full-year growth, and it’s the main reason growth stalled while peers ran ahead. This is the textbook risk that looks like an AI story but is really a concentration story. It cuts both ways: if that GSE’s budget recovers and a long-pending consolidation deal lands, growth snaps back fast; if Hexaware loses the still-undecided consolidation deal (it’s the incumbent, but management has prudently assumed a downside), it’s another leg down. Watch for: the GSE consolidation decision, and whether Financial Services keeps leading growth.
The crux, interrogated: does AI deflate or expand this business?
In one sentence: Hexaware works as a long-term investment if and only if AI ends up growing the pool of IT-services work faster than it shrinks the billable hours inside Hexaware’s existing contracts.
The mechanism, in plain terms. Indian IT has always billed roughly in proportion to headcount: more work = more people = more revenue. Generative AI breaks that link. If a tool lets four engineers do what six used to, then a client renewing the same scope of work needs ~33% fewer billed hours — and Hexaware’s revenue on that contract falls by that much. The CEO put the number himself: a 20–40% dampener on same-scope renewals. That is not a rumour; that is the incumbent telling you his own renewals shrink.
Test the analogy. Is this like the cloud shift of the 2010s — scary, then expansionary? Partly. When companies moved from their own servers to the cloud, the old work (running data centres) shrank, but a bigger new line (cloud migration and management) appeared, and Indian IT grew straight through it — Infosys’s “digital” revenue went from nothing to 38% of the company in a few years. Automation (RPA) in 2017–19 was the same: it killed narrow rule-based tasks and the industry kept compounding to ~$181bn of exports. So the precedent says: each tech shift compresses the old line and grows a new, larger one — net expansionary, painful for laggards. But — and this is where the analogy bends — RPA and cloud automated infrastructure and rules. GenAI attacks knowledge work itself, the actual coding and processing that is the product. It may be faster and deeper than anything before. The honest read: the precedent is reassuring but not conclusive, because this disruption reaches one layer closer to the core.
Map the threat by name. It isn’t a rival IT firm — it’s the AI labs and the productivity itself.
| Threat | What it does | Proof point |
|---|---|---|
| Anthropic (Claude “Cowork” agents) | Feb 2026 launch of agents that run whole professional workflows (legal, data, ops) — attacks the service layer, not just coding | Triggered a ~6% one-day, ~8% one-week crash in Nifty IT; ~$50bn erased across Feb 2026 [HARD on moves] |
| In-house AI tooling at clients | Clients use AI to do work they used to outsource | TCS headcount fell ~23,460 in FY26; HCLTech CEO models 2–3%/yr price deflation [HARD/SOFT] |
| AI-native / modernization-led peers (Coforge, Persistent) | Same end-market, growing 20–35% while Hexaware stalled | Coforge FY26 ~+29%, Persistent ~+17% vs Hexaware +7.6% [HARD] |
The real-world precedent, scored. European day-ahead power coupling, the dot-com build-out, RPA, cloud — incumbents who owned the new tool captured the surplus; those who clung to the old billing model got commoditised. Applied here: the firms positioning as “we’ll build and run your AI” (modernization, agent platforms) are growing; the ones most exposed to plain maintenance and back-office processing are most at risk.
The follow-on questions, answered.
- Is the damage to volume or to price? Both — fewer hours (volume) and, per HCLTech, gentle price deflation. Volume is the bigger near-term hit.
- Which part of Hexaware is hit first? Its IT-operations / business-process-services book and routine maintenance — the very “AI for IT, leave it to us” pitch. That’s the irony: its headline positioning sits on the most-automatable work.
- Which part is protected? Net-new builds, legacy-system rewrites (“you need someone who understands the old code first — that’s our secret sauce,” says the CEO), and running the AI agents themselves. These can grow.
- Has anyone actually moved yet? The deflation is real and showing up (peer headcount cuts, Hexaware’s own dampener). The expansion is mostly pipeline and platform launches — promised, not yet in the revenue line. That asymmetry — pain booked, gains pending — is exactly why the sector de-rated.
- Is Hexaware’s stall an AI problem or a client problem? On the evidence, mostly a client problem (the GSE cut, furloughs, lower licence revenue) that coincided with the AI scare — which is why the market punished it doubly. Notably, even fast-growing Persistent is decelerating beneath strong headlines, which slightly exonerates Hexaware.
Honest verdict on the crux: not “too hard,” but genuinely two-sided. The disruption is real and arrives one layer closer to the core than past shifts — yet the historical pattern (compress-old, grow-new) has never once failed the industry, and Hexaware’s specific stall is better explained by one client than by AI. The bet is on whether this team can pivot from selling hours to selling outcomes fast enough. That’s a judgement about execution, not arithmetic.
The watch-list
- Carlyle stake drops meaningfully below 74% via clean block deals — and the pledge is released.
- CY26 revenue growth prints clearly above 7.6% (management’s own bar), with acceleration each quarter after Q1.
- EBIT margin troughs in H1 CY26 and exits the year above CY25 — proof the rebadging/ramp drag was temporary, not structural.
- The GSE consolidation deal is decided — and which way.
- AI “new revenue” (legacy modernization, “Zero License”) shows up as named, sized deal wins converting to revenue, not just pipeline.
- Operating cash flow stays ~100% of operating profit — the integrity tell. If profit stops becoming cash, walk.
QGLP scorecard (the Motilal Oswal lens) — the receipts
RoE = profit earned per ₹100 of owners’ money. RoCE = profit per ₹100 of all capital used. OPM = operating profit margin. PEG = price-to-earnings divided by growth rate; below 1 is cheap-for-the-growth.
| # | Question | Score | Evidence |
|---|---|---|---|
| Quality of Business | 5/6 | ||
| 1 | Large opportunity? | 1 | Global IT-services TAM >$2tn, AI/digital growing; huge runway |
| 2 | Favourable industry structure? | 0.5 | Oligopoly at the top, but mid-tier competitive; OPM slipped 17%→14% |
| 3 | Defensible moat? | 0.5 | Switching costs real; RoCE 26–37% for 12 yrs (returns moat real) — but no pricing power, narrowing under AI |
| 4 | High return ratios (>15%)? | 1 | RoE 24.9%, RoCE 30.1%; both >15% for a decade |
| 5 | Asset-light? | 1 | Capex ~₹163 cr on ₹13,430 cr sales; FCF ₹1,576 cr > PAT ₹1,368 cr |
| 6 | Favourable terms of trade? | 1 | Working-capital days now −16 (was +75 in 2022) — collects before it pays |
| Quality of Management | 3.5/6 | ||
| 7 | Unquestionable integrity? | 0.5 | Clean audit, cash-backed profit — but entire promoter stake pledged; heavy “adjusted” margins |
| 8 | Proven execution? | 0.5 | Revenue doubled 2020–25, CEO 11 yrs — but missed CY25, guided CY26 margins down |
| 9 | Growth mindset / vision? | 1 | Investing in AI platforms, bolt-on M&A, new verticals, private-markets push |
| 10 | Superior capital allocation? | 0.5 | Net cash, no dilution, sensible cash-funded M&A — but IPO 100% OFS; one acquisition (Softcrylic) missed earnout |
| 11 | Clear succession? | 0.5 | Deep professional bench — but PE-controlled board, ownership in flux |
| 12 | Minority interests protected? | 0.5 | High dividend, independent chair — but ~75% PE owner, half the board, sell-down overhang |
| Growth | 3/6 | ||
| 13 | Structural tailwind? | 0.5 | IT grows ~1–1.5× GDP; currently soft demand + AI uncertainty |
| 14 | Volume-led growth? | 0.5 | Historically volume-led; AI now a 20–40% price/productivity dampener |
| 15 | Operating leverage? | 0 | OPM compressed 17%→14%; CY26 EBIT guided 13–14%, lower |
| 16 | Manageable leverage? | 1 | Net cash; borrowings ₹681 cr (mostly leases) vs ₹6,316 cr net worth |
| 17 | Market-share gains? | 0.5 | Pipeline >$4bn, consolidation wins — but growth lagged peers |
| 18 | Earnings growth >15%? | 0.5 | 5-yr PAT CAGR ~17% — but forward guided to low-teens |
| Longevity | 2.5/5 | ||
| 19 | Relevant in 10–15 yrs (low disruption)? | 0.5 | Survives, but genuine GenAI disruption risk — “lots of technology” |
| 20 | Extend competitive-advantage period? | 0.5 | Returns moat real but narrowing under AI |
| 21 | Sustain growth-advantage period? | 0.5 | Long runway IF AI expansionary; short if deflationary |
| 22 | Diversification headroom? | 0.5 | Heavy Americas (74%)/BFSI — room to widen geo + new TPP vertical |
| 23 | Adaptive culture? | 0.5 | Low attrition, AI first-mover, recovered from 2016/COVID dips |
| Price | 1/2 | ||
| 24 | Valuation reasonable? | 0.5 | P/E 20.8; PEG ~1.2 on trailing growth; well below peers (Coforge 38×, Persistent 39×) |
| 25 | Margin of safety (PEG<1 or payback<1)? | 0.5 | PEG ~1.2, payback ~3.1× — neither clears the strict bar, but down 44% from high |
| Total | 15 / 25 |
The pattern is clear: Quality-of-business is the strength (5/6 — this is a beautiful set of cash economics). Management quality is dragged down not by the operators but by the private-equity ownership. Growth and Longevity are where the real doubt sits — a decelerating moment plus a genuine AI question. Price is the only thing that’s actually attractive relative to peers, even if not cheap on its own arithmetic.
Buffett lens (the Berkshire-letters read)
| # | Test | Verdict | Evidence |
|---|---|---|---|
| 1 | Good boat? (business > management) | PARTIAL | A Good business with Great economics — RoCE 30%, tiny reinvestment — but no franchise |
| 2 | Moat + pricing power | PARTIAL | Switching costs + RoE>CoE for a decade; but OPM falling, AI dampener = no pricing power |
| 3 | See’s test (high returns, little capital) | PASS | Capex ~1% of sales; FCF/PAT >100%; RoCE 30% — a fountain of cash |
| 4 | One-dollar test | PASS | Reserves ₹1,230→₹6,255 cr over a decade while paying ~50% out; no dilution |
| 5 | Owner-oriented, candid management | PARTIAL | CEO candid (guided down honestly) — but PE owner ran it for cash upstream, IPO 100% OFS |
| 6 | Integrity / no “credit P&L, debit B/S” | PASS | OCF ~100–110% of OP for a decade; debtor days 56; clean audit |
| 7 | Circle of competence / predictability | FAIL | IT services mid-AI-transition — cannot confidently say what this looks like in 10 yrs |
| 8 | Mr. Market — gift or trap? | PARTIAL | Down 44% from high, P/E below peers, fearful — but PEG ~1.2, not dirt cheap |
| 9 | Patience / compounding runway | PARTIAL | High RoCE + reinvestment runway — but growth decelerating, runway clouded by AI |
| 10 | The honest red flag | (see below) | — |
Count: 3 PASS + 5 PARTIAL (×½) + 1 FAIL = 5.5 / 10 — “a real business with real gaps.”
The See’s test, spelled out. This is where Hexaware genuinely shines. See’s Candy needed almost no extra capital to grow and threw off cash for decades. Hexaware is cut from similar cloth: in CY25 it spent only ~₹163 cr of capex to run ₹13,430 cr of sales, and converted ₹1,368 cr of profit into ₹1,576 cr of free cash. It earns ~30% on capital employed without swallowing capital to do it. On this test, it’s close to a textbook Buffett business.
The one-dollar test, spelled out. Buffett asks: for every rupee a company keeps instead of paying out, did it create at least a rupee of value? Over the decade, Hexaware retained roughly half its profits (paying the other half as dividends) and grew its reserves five-fold while holding a 25–30% return on equity — without issuing new shares. Each retained rupee kept earning ~25%. That passes cleanly. The asterisk is the owner’s use of cash, not the company’s: the IPO returned money to Carlyle, not to the business, and the high-payout years helped service the owner’s buyout debt. The company allocated well; the owner allocated for itself.
The honest red flag (test 10). The single strongest reason this might not be a wealth creator from here: its core “AI for IT operations” business sells exactly the work that AI is best at automating, and it’s owned by a financial seller, not a builder. If the 20–40% renewal dampener bites faster than new AI-led work ramps, revenue and margins compress together — and a private-equity owner heading for the exit has every incentive to manage the story and the cash rather than make the patient, painful investments a decade-long platform pivot demands. The numbers don’t refute this yet; they show a stall that’s currently client-specific, but a margin guide that’s already pointing down. This is the bear case’s strongest form, and it’s not silly.
The framework metrics
- Economic Profit = Net Worth × (RoE − 12%) = ₹6,316 cr × (24.9% − 12%) = +₹815 cr/yr → comfortably creating value above the cost of owners’ money. Top-tier on this measure.
- Terms of Trade = working-capital days −16 (improved from +75 in 2022) → favourable; debtor days 56. It now runs on slightly negative working capital — a small cash engine.
- 5-yr Payback = Mcap ₹30,526 cr ÷ projected cumulative 5-yr PAT (~₹9,735 cr at 12% PAT CAGR) = ~3.1× → above the <1× multibagger signal. (A <1× payback is rare and reserved for deep-value; this isn’t that.)
- PEG = P/E 20.8 ÷ ~17% trailing PAT growth = ~1.2× (≈1.7× on forward low-teens growth). A mild premium, not a bargain.
- RoE − CoE spread = +12.9%; RoE has almost certainly been >15% in all of the last 10 years (RoCE was 26–37% throughout).
- Consistent / Volatile test = CONSISTENT (pass). Over 11 reported years, net profit fell only once (a 3% dip in COVID-2020), never >50%, and ended ~4× higher. Value this on earnings (P/E), not book.
Peer comparison
| Company | Mkt cap | CMP | P/E | P/B | RoE | RoCE | OPM | Latest sales |
|---|---|---|---|---|---|---|---|---|
| Hexaware | ₹30,526 cr | ₹500 | 20.8 | 4.9× | 24.9% | 30.1% | 14% | ₹13,430 cr (CY25) |
| Coforge | ₹62,922 cr | ₹1,463 | 38.3 | 5.2× | 20.6% | 23.5% | 18% | ₹16,403 cr |
| Mphasis | ₹43,280 cr | ₹2,268 | 22.9 | 4.0× | 18.5% | 22.8% | 19% | ₹15,880 cr |
| Persistent | ₹76,177 cr | ₹4,829 | 39.4 | 9.7× | 27.3% | 34.4% | 19% | ₹14,748 cr |
The relative read flips the absolute one, and that’s the most useful thing in this table. On its own arithmetic Hexaware is only fairly priced (PEG ~1.2). But against its asset class it’s the cheapest name by a mile — 20.8× versus Coforge and Persistent near 38–39× — while earning higher returns on capital than Coforge or Mphasis. The catch is honest: it earns the discount, because it’s the slowest-growing of the four right now (CY25 +7.6% vs Coforge ~+29%, Persistent ~+17%) and it’s a couple of percentage points light on operating margin. So the question a sector-allocator must answer is the mirror of the value-investor’s: is the growth gap temporary (a client blip) — in which case the discount is a gift — or structural (it’s the AI loser of the group) — in which case the discount is deserved? The evidence leans toward “mostly temporary,” but it’s not settled.
Latest quarter & what’s happening now
Q4 CY25, reported 5 Feb 2026. Revenue $389m, down 1.5% sequentially — blamed on seasonal furloughs, a $7m drop in licence revenue, and FX. Full-year CY25 revenue grew 7.6% (7.1% constant-currency). Reported EBITDA margin 17.1% (up 120bps for the year), but cluttered with one-timers (a Softcrylic earnout reversal, an offsetting impairment, a bad-debt provision). The market-moving item was guidance: management switched to reporting EBIT and guided CY26 EBIT margin to just 13–14% — below CY25 — citing first-half deal ramp-ups with heavy “rebadging” (absorbing clients’ staff) that depress margins before “right-shoring.” They promised growth above CY25’s 7.6% and a stronger H2 margin exit. Two concall takeaways: (1) the CEO’s central message was AI — he called it both a near-term “dampener” and the biggest new opportunity, and pointedly said Hexaware has been “first off the block” on AI offerings (legacy modernization, vibe-coding, “Zero License” to replace SaaS); (2) the order pipeline crossed $4bn for the first time, with a large big-tech consolidation deal won (though without a guaranteed book of work) and a still-pending GSE consolidation decision. The stock fell ~17–20% over the two days after.
Where the two lenses agree — and disagree
They agree on the shape: a financially excellent, cash-generative business (QGLP’s Quality-of-Business 5/6; Buffett’s See’s and one-dollar tests both PASS) that is not a clean compounder you can close your eyes and own (QGLP 15/25; Buffett 5.5/10).
The most important divergence is internal to the Buffett lens, and it’s the signal: the business economics pass emphatically (tests 3, 4, 6), but predictability fails outright (test 7). QGLP’s checklist, by scoring growth and quality somewhat mechanically, is gentler on this — it can’t fully “see” that the entire ten-year thesis rests on an unresolved technology question. Trust the Buffett flag here. When a checklist says “good business, decent growth” but the letters say “I can’t tell you what this looks like in ten years,” the letters are catching the thing that matters. The second, quieter divergence: QGLP rewards the high dividend and clean balance sheet (capital allocation) more than the Buffett read does — because Buffett looks past the company’s tidy books to the owner’s behaviour, and sees a financial seller, not a partner.
Margin-of-safety price band
Not a recommendation — just the arithmetic. To satisfy QGLP’s Price pillar strictly (PEG ≤ 1× on a realistic low-teens forward growth), you’d want a P/E around 13–15× — roughly ₹300–₹390 on current earnings. The Buffett Mr.-Market read on CMP ₹500 is “fearful, but not yet a fat pitch”: the stock is down 44%, priced below every mid-cap peer, and the bad news (one client, near-term margins) looks more cyclical than terminal — but PEG ~1.2 means you’re not being handed a bargain, you’re being handed fair value on a quality business at a fearful moment.
Blending the two: the quality becomes comfortably owned around ₹380–₹460 — where you’d be paying ~16–19× for a 25%-RoE, net-cash, cash-gushing business that’s cheap to its peers. Below ~₹400 (it touched ₹400 this year) it starts to look genuinely cheap if growth re-accelerates. At ₹500 it’s fair on the franchise and cheap on the comp table, but slightly demanding on the spot numbers — which is exactly what you’d expect of a good business caught between a client air-pocket and an AI question mark. This is not a wonderful business at an unwonderful price; it’s a good business at a fair price, where the upside is a re-rating toward peers if the stall proves temporary, and the downside is the AI dampener and the Carlyle selling proving more than temporary.
Conviction texture
The bull case, at its strongest: You’re buying a 25%-return-on-equity, net-cash, free-cash-gushing business at 20× — half the multiple of peers earning lower returns — because of a single client’s budget cut that the company says is fixable and a sector-wide AI scare that has hit the price faster than it’s hit the fundamentals. The CEO is a steady eleven-year operator, attrition is industry-low, the pipeline just crossed $4bn, and AI’s “new work” (rebuilding legacy systems someone has to understand first) plays to a genuine strength. If CY26 growth beats 7.6%, margins trough and recover, and Carlyle sells down cleanly into eager institutions, this re-rates toward 28–35× and the overhang lifts at once. That’s a lot of upside from a fearful price.
The bear case, at its strongest: This is a twice-flipped private-equity asset whose core “run your IT operations” business sits squarely in AI’s blast radius — the CEO himself admits 20–40% of renewal work evaporates — and whose owner, holding 74% pledged against debt, is a guaranteed future seller with every reason to optimise the story over the decade-long pivot the business actually needs. Growth has collapsed from 16% to ~1%, margins are guided down, and “AI new revenue” is still a pipeline slide, not a revenue line. The cheap multiple isn’t a gift; it’s the market correctly pricing a slower, more disrupted, less-aligned business than its peers.
What the numbers actually support: A genuinely excellent set of historical economics (the cash conversion, returns and consistency are not in doubt) wrapped around two forward uncertainties the numbers can’t yet resolve — the AI trajectory and the ownership exit. The stall is, on today’s evidence, more client-specific than AI-driven, which tilts the near term bullish; the long-term predictability question is real and unresolved, which caps the conviction.
The three things to watch that tip it: (1) CY26 revenue clearly beating 7.6% with quarter-on-quarter acceleration; (2) EBIT margin troughing in H1 and exiting the year above CY25; (3) Carlyle’s stake falling via clean blocks with the pledge released. Get those three and the bull case is winning. Miss them — especially if AI-driven deflation shows up in the renewal book — and the bear case was right. No buy/sell/hold here — that call is yours.
Sources
- Screener snapshot (financials, ratios, shareholding): https://www.screener.in/company/HEXT/consolidated/ — fetched 2026-06-20.
- Hexaware Q4 CY25 earnings call transcript, 5 Feb 2026 (AI strategy, CY26 guidance, GSE/furlough commentary, $4bn pipeline) — filed with BSE/NSE 10 Feb 2026.
- Hexaware CY25 results press release (revenue $1,537.4m, attrition ~11%, margins) — hexaware.com.
- Ownership/governance: Carlyle IPO release (Feb 2025, ₹708, 100% OFS); CA Magnum $1.255bn loan & 74.55% stake pledge (Nov 2025); Baring delisting 2020 (₹475); founder Atul Nishar exit 2013. Tagged HARD/SOFT in working notes.
- AI-disruption crux: Anthropic “Cowork” launch & Nifty IT sell-off (Feb 2026, ~$50bn erased); TCS FY26 headcount −23,460; HCLTech CEO 2–3% price-deflation comment; ICICI Direct sector estimates; NASSCOM FY19 precedent. Multiple dated sources in working notes.
- Peers (Coforge, Mphasis, Persistent): screener snapshots, fetched 2026-06-20.
- Assumptions: Cost of equity 12%; forward PAT CAGR 12% (conservative vs management’s low-teens base case) for the payback calc; trailing 17% 5-yr PAT CAGR for PEG. All figures consolidated, ₹ crore unless stated.