Fractal Analytics IPO

It’s had a pretty clear comeback year. In FY25, it reported a profit of ₹220.6 crore, versus a loss of ₹54.7 crore in FY24. Revenue from operations also jumped 25.9% to ₹2,765.4 crore over the same period.

A big positive here is how well it holds on to customers and grows business with the same set of clients. You can see that in its Net Revenue Retention (NRR - how much revenue you keep and grow from the same clients) of 114% for the six months ended September 2025. Also, its top 10 clients have stayed with it for an average of more than eight years, which is a solid sign of stickiness.

It’s intentionally going after large, high-value companies it calls “Must Win Clients” (MWCs - basically, priority accounts it really wants to win and keep). As of September 2025, it served 122 MWCs, which it defines as companies with over $10 billion in revenue, $20 billion market cap, or 30 million customers. The simple benefit: this kind of client list usually makes revenues feel steadier and more predictable.

Profitability at the operating level is improving, too. Its adjusted EBITDA margin expanded to 17.4% in FY25 from 10.6% in FY24. In rupee terms, adjusted EBITDA more than doubled to ₹482.1 crore in FY25 from ₹232.1 crore in the previous year.

It also keeps putting money into innovation to stay ahead, with ₹143.6 crore or 5.2% of operating revenue spent on R&D (research and development - building new tech and improving products) in FY25. That steady investment has helped it build a base of 28 registered patents as of January 2026, which is basically its legally protected “ownership” over certain inventions and methods.

It depends a lot on a small set of customers for most of its income. In the six months ended September 2025, its top 10 clients, which include Philips, Mars, Mondelez, and C3 AI, made up 54.2% of revenue in the Fractal.ai segment. So if even one or two of these key clients cut back or leave, it could hit the company’s finances in a big way.

This is a people-heavy business, meaning talent is the main engine. Employee benefits costs were 72.2% of revenue in the six months ended September 2025, or ₹1,125.2 crore. On top of that, it reported an attrition rate of 15.7% (nearly 16 out of every 100 employees are leaving the company), which makes it harder to keep experienced people and deliver projects smoothly.

Even though it’s profitable recently, it has a track record of losses, including a net loss of ₹54.7 crore in FY24. There’s no certainty it can stay profitable as it scales up and keeps spending on new bets and initiatives.

A large chunk of its business is tied to the US. In the six months ended September 2025, the United States contributed 64.9% of revenue, or ₹1,012.5 crore. That means a US slowdown, tighter budgets, or even regulatory changes could affect it more than you’d expect.

While the main company is in profit, some subsidiaries are still losing money. For instance, Senseforth AI posted a loss of ₹6.8 crore in FY25. If these losses continue, the group may need to keep funding them, which can pressure consolidated cash flows (the overall cash moving in and out across the whole group).