01 — At a Glance
The AI Productivity Play That Investors Love to Hate Right Now
- 52-Week High / Low₹6,599 / ₹4,149
- FY25 Revenue (Full Year)₹11,939 Cr
- FY25 PAT (Full Year)₹1,400 Cr
- Full-Year EPS (FY25)₹90.54
- Q3 FY26 EPS₹27.86
- Book Value₹453
- Price to Book10.5x
- Dividend Yield0.73%
- Debt / Equity0.06x
- Return 3-Month-24.7%
The Setup: Persistent Systems just reported Q3 FY26 with ₹3,778 Cr quarterly revenue (+23.4% YoY), ₹439 Cr PAT (+17.8% YoY), and an EBIT margin of 14.4% — down 190 basis points, because India’s new Labour Code demanded an unexpected ₹86 crore provisioning for gratuity and leave encashment. Adjust for that one-timer, and margins were actually 16.7%, up 40 bps sequentially. The stock traded down 25% in 3 months anyway. Welcome to mid-cap IT services investing — where fundamentals and price discovery play by different rules.
02 — Introduction
When Your AI Strategy Actually Works (But Investors Don’t Care Yet)
Persistent Systems is what happens when a mid-cap IT services company decides to bet their entire future on engineering productivity through AI tools. Not “AI consultants” who tell you things. Not “AI evangelists” who charge you to talk about AI. Actual AI tools — SASVA, iAURA, GenAI Hub — embedded into service delivery, scaled across 200+ agent deployments, and driving a commercial construct where clients pay less money for more output. That’s the plot.
Now here’s where it gets interesting: this is actually working. Revenue is growing at 23–28% annually. Operating profit is growing faster. TCV bookings are hitting record levels. Client concentration is dropping (Top 10 was 45% in FY22, now 39%). And the company is expanding into spaces like private equity, healthcare R&D, semiconductor R&D, where they hadn’t really played before. “23 sequential quarters of revenue growth” — that’s literally a six-year streak without a down quarter.
So naturally, the stock crashed 25% in three months.
The reason? Labour Code provisioning hit operating margins unexpectedly. Wage hikes added structural cost headwinds. A few senior exits signalled transition risk. And the market — ever the patient investor — immediately concluded that Persistent must be a value trap. This is the intellectual rigour on display when mid-caps stumble on margin expansion. One provision hits, and suddenly nobody cares that organic demand is the strongest it’s been in years.
Concall Insight (Jan 2026): CEO on AI tool-driven margin benefit (+150 bps QoQ): “Not on account of one deal. It reflects multiple deals using SASVA, iAURA, GenAI Hub that are now scaling.” Translation: it’s structural, it’s repeatable, and the market is choosing not to believe it yet.
03 — Business Model: WTF Do They Even Do?
Digital Engineering for Companies That Want to Not Die (It’s a Growing Market)
Persistent sells “digital engineering” — a fancy term for “we’ll help you modernize your applications, data, and infrastructure so your decade-old monolith doesn’t explode in your face.” The company operates across three verticals: BFSI (31% of revenue), Healthcare & Life Sciences (22%), and Software/Hi-Tech/Emerging Industries (47%). This vertical mix matters because all three are drowning in technical debt and absolutely desperate to modernize before their competitors do.
The pitch has evolved. In 2020–2022, Persistent won deals like everyone else — “let’s move your app to cloud,” “let’s build you a modern data lake,” “let’s teach your engineers Kubernetes.” Standard. By 2024–2025, the message shifted: “We’ll use AI tools to do it 60–75% faster. You pay less. We make more. Your problem solved in half the time.” That’s the SASVA platform. The iAURA platform handles enterprise data readiness with similar time-to-market acceleration.
Revenue mix: technology services (70%) are still the bread and butter. Partnership operations — Salesforce, AWS, Google Cloud, Microsoft integrations — add another 26%. Accelerite, their product accelerator, contributes 4%. Geographic concentration remains a weakness: 80.8% from North America in FY25. But growth is broad-based. BFSI +35.7% YoY. Healthcare +13.4% YoY. Hi-Tech +19.6% YoY. That’s not concentration risk. That’s demand strength across the board.
Top 10 Clients39%Down from 45% (FY22)
Clients >$1M195Up from 36 (FY15)
Headcount26,711+487 QoQ
Utilization Rate88.4%Q3 FY26 (incl. trainees)
The SASVA Moment: In Q3, they closed a European global bank using SASVA to “decode millions of lines of legacy code.” Outcome: ~22% productivity/time-to-market improvement. They’ve filed 105 patents in the last 18 months. This is not “advisory work.” This is actual engineering IP. If it scales, margins expand. If it doesn’t, they’re just another body-shop.
💬 Do you think AI tools embedded into service delivery will actually change IT services unit economics, or is it just another consultant’s fairy tale? Drop your hot take!
04 — Financials Overview
Q3 FY26: The Raw Numbers
Result type: Quarterly Results | Q3 FY26 EPS: ₹27.86 | Annualised EPS (Q3×4): ₹111.44 | FY25 full-year EPS: ₹90.54
| Metric (₹ Cr / USD Mn) |
Q3 FY26
Dec 2025 |
Q3 FY25
Dec 2024 |
Q2 FY26
Sep 2025 |
YoY % |
QoQ % |
| Revenue (₹ Cr) | 3,778 | 3,062 | 3,581 | +23.4% | +5.5% |
| Revenue (USD Mn) | $422.5 | $361.4 | $406.0 | +17.0% | +4.0% |
| Operating Profit (EBIT) | 733 | 538 | 683 | +36.2% | +7.3% |
| EBIT Margin % | 19.4% | 17.6% | 19.1% | +180 bps | +30 bps |
| PAT | 439 | 373 | 471 | +17.8% | -6.8% |
| EPS (₹) | 27.86 | 23.55 | 30.15 | +18.3% | -7.6% |
The Labour Code Curveball: Reported EBIT margin was 14.4% (down 190 bps QoQ) due to one-time Labour Code provisioning of ₹86 crore (~2.3% of EBIT, ~1.8% of PAT). Exclude that, and EBIT margin would have been 16.7% — actually up 40 bps sequentially. This matters because operating leverage is real; management is simply paying off regulatory debt in one hit rather than spreading it over time. Meanwhile, wage hikes (effective Oct 1, 2025) added -180 bps structural cost pressure. Net? Margin expansion story interrupted by bad timing, not bad business.
05 — Valuation: Fair Value Range
What’s This Company Worth When Margins Actually Stabilize?
Join 10,000+ investors who read this every week.
