01 — At a Glance
The AI Company That’s Learning to Walk Before It Runs
- 52-Week High / Low₹900 / ₹430
- CY25 Revenue₹13,430 Cr (USD $1,595M approx)
- CY25 PAT₹1,368 Cr
- Q4 Revenue (USD)$389M
- Q4 EPS (₹)₹4.79
- Book Value₹103
- Price to Book4.41x
- Dividend Yield2.52%
- Debt / Equity0.11x
- IPO Price (Feb 2025)₹600
The Real Drama: Hexaware listed 14 months ago at ₹600/share and is now trading at ₹456 (down 24%). Yet Q4 delivered 24.4% profit growth YoY, 12.2% revenue growth, and pipeline crossed $4 billion for the first time. The stock rewards this by declining 40% in 12 months. Welcome to IT services valuation purgatory, where good results meet AI-induced margin compression anxiety.
02 — Introduction
Once Delisted. Then IPO’d. Now Unlisted in Investor Affection.
Here’s the absolutely bizarre journey of Hexaware Technologies. In 1992, it was founded as an IT services company. Twenty-eight years later, in 2020, the promoters took it private in a delisting. In February 2025, it came back to the market via IPO at ₹600/share, raising ₹8,750 crore. Exactly one year later, the stock trades at ₹456. Down 24%. So either IPO investors are geniuses who got out early, or they’re very confused about what this company actually does.
Hexaware Technologies is a mid-tier IT services player with a peculiar strategic bet: it’s pivoting everything toward AI-powered delivery. Not AI for selling, mind you. AI for doing the actual IT work better and faster. Think of it as: we hire engineers, but then we teach them to collaborate with AI agents that do 20-40% of the grunt work automatically. The margin impact is obvious — you need fewer engineers per dollar of revenue. But the ramp-up period is brutal: retraining, productivity losses, deal repricing to clients. Q4 CY25 proved exactly that.
The company delivered its freshly-listed results with a $4 billion pipeline, 10th straight quarter of headcount additions, consolidation deals that grant “right to hunt” across massive spend pools, and a management team saying “AI is currently dampening our revenue growth, but this is temporary.” Translation: we’re sacrificing current earnings for future optionality. That’s how you get a stock down 41% in a year while generating 24% profit growth.
The Question Everyone Is Asking: Is Hexaware a cheap AI-powered IT services story being punished for delisting-era management changes and IPO-era overhype? Or is it a company that won’t achieve its growth targets because client budgets are contracting and AI productivity just means fewer billable hours? Let’s find out with data.
03 — Business Model: Code Farms With AI Jockeys
What We Do: Write Software. What We Want To Do: Train AI To Write It Faster.
Hexaware operates a classic IT services model: 33,844 employees sit in delivery centers across 39 offices globally. 74% of revenue comes from the USA, 19% from Europe, 6% from APAC. Clients pay for software development, legacy modernization, cloud migration, IT ops outsourcing, and business process automation.
But here’s the twist. In 2024–2025, Hexaware went all-in on an “AI as a delivery mechanism” thesis. Instead of just selling “we’ll build your platform,” it’s now selling “we’ll build your platform, but our engineers collaborate with AI agents, so we do it faster and you get it cheaper.” The platforms are catchy: RapidX for software engineering, Tensai for IT operations, Amaze for cloud adoption. Management positioned this as the “first-off-the-block” approach in the industry. Whether it’s true or defensible is TBD.
Revenue breakup (Q4 CY25): IT Services ~88%, Business Process Services ~12%. That 12% BPS is where industrial growth is expected, but it shrank because a GSE (Government Sponsored Enterprise) client cut headcount and several manufacturing clients faced macro headwinds. So the “growth engine” tanked. The IT Services side grew 10.3% QoQ, which is decent, but management admitted this is net of significant “productivity dampening” from AI reskilling and deal repricing.
Client concentration is manageable: top 5 customers = 25.9% of revenue, top 10 = 36.4%. So not completely hostage to one buyer. But the GSE cuts and a “large client dispute” in the US (mediation failed; legal proceedings pending) show vulnerability.
Top 5 Client Share25.9%Concentrated
USA Revenue74%Highly dependent
Employee Attrition11%IT industry: low
Headcount33,844+10 qtrs adds
The Real Bet: Management is betting that “AI-native” delivery is defensible and margins will compress for 2–3 quarters before rebounding as they “right-shore” deals and lock in AI productivity gains. That’s a bet on execution against a macro backdrop of client budget freezes and a stock market that hates uncertainty.
💬 If you were a CIO at a Fortune 500 company, would you trust an AI-powered delivery model from a mid-tier services firm that’s still learning to walk, or would you stick with TCS/Infosys?
04 — Financials Overview
Q4 CY25: The Numbers That Make You Scratch Your Head
Result type: Quarterly Results | Q4 CY25 EPS: ₹4.79 | Annualised EPS (Q4×4): ₹19.16 | Full-year CY25 EPS: ₹22.41
| Metric (₹ Cr) |
Q4 CY25
Dec 2025 |
Q4 CY24
Dec 2024 |
Q3 CY25
Sep 2025 |
YoY % |
QoQ % |
| Revenue | 3,478 | 3,154 | 3,484 | +10.3% | -0.2% |
| Operating Profit | 378 | 490 | 601 | -22.9% | -37.1% |
| OPM % | 11% | 16% | 17% | -500 bps | -600 bps |
| PAT | 292 | 321 | 370 | -9.0% | -21.1% |
| EPS (₹) | 4.79 | 5.25 | 6.08 | -8.8% | -21.2% |
What Just Happened: Revenue grew 10.3% YoY, but operating profit collapsed 22.9%. Margin compression of 500 bps YoY and 600 bps QoQ. The PAT actually declined 9% despite revenue growth. This is not normal. Management blamed: (a) labour code cost impact (~₹80 crore annualised), (b) earnout reversal on Softcrylic acquisition that freed up ₹160 crore (one-timer), (c) impairments on client relationships (₹100 crore), (d) lower license pass-through revenues (third-party licenses), and (e) AI-driven productivity dampening margins. Stripping one-timers, normalised EBITDA margin was 15.4%, down 210 bps QoQ. P/E at 19.2x is at industry median (TCS 17.9x, Infosys 18.5x), but you’re paying market rates for a company deleveraging margins in the near term.
05 — Valuation: Fair Value Range
What’s This AI Productivity Experiment Worth?
Join 10,000+ investors who read this every week.
