Fractal Analytics: An AI Services Business Wearing an AI Platform Multiple

SumanSpeaks

Independent Capital Markets & Geopolitical Intelligence

Equity Research | Valuation Alert

Fractal Analytics: An AI Services Business Wearing an AI Platform Multiple

At ₹708 (Face Value: ₹1), the market pays about ₹12,176 crore for a company whose platform arm is 2.7% of revenue, whose dollar growth has slowed from 22% to 14%, and whose free cash flow fell 39% in FY26. SumanSpeaks puts fair value at ₹460, some 35% below the current price.

Let me put one question in front of you. What would you pay for a company that earns 97% of its revenue from AI services, but trades on the vocabulary of an AI platform?

Fractal Analytics Ltd (₹708) has slipped below the ₹732 52-week low quoted in PL Capital's note of 31 August 2026, so it is now trading at fresh lows. Yet its market capitalisation is still about 3.7 times FY26 revenue of ₹3,299.7 crore. That works out to roughly 40 times FY26 adjusted profit on the current share count and about 34 times FY27E EPS.

I do not think that price survives a hard look at the numbers. Here is why.

1	The Multiple Is Not the Problem. The Mix Is.

A sales multiple of 3.7 times is not expensive in itself. On a software company with 25% net margins it would be a bargain. Fractal's FY26 gross margin is 46.8% and its adjusted net margin is 9.2%. The multiple has to be read against that.

The red herring prospectus itself says no listed company in India or abroad follows a comparable business model. Read that again. No comparable means no valuation anchor, and when there is no anchor, the market borrows the nearest glamorous one.

Global software platforms trade at anywhere from 25 to 90 times sales. Fractal Alpha, the group's platform arm, contributed $10 million of the group's $374 million FY26 revenue. That is 2.7%. The remaining 97% comes from Fractal.ai, where direct delivery costs swallow 53% of revenue. The global IT-consulting bellwether trades near 1.8 times sales on approximate September 2026 data. Fractal's 3.7 times is more than double that.

A 2.7% platform cannot justify a 100% platform multiple.

2	Growth Is Slowing, and the Rupee Is Hiding It

Headline revenue grew 19.3% in rupees. In dollars, the currency in which Fractal actually bills, it grew about 14%, from $327 million to $374 million. Nearly five points of the rupee growth came from currency, not from clients.

The trend is the real worry. Constant-currency growth ran at 22.2% in FY25, slowed to 13.3% in FY26, and PL Capital itself projects 11.3% for FY27E. Fractal.ai's dollar revenue compounded at just 14% between FY23 and FY26, from $245 million to $364 million.

Two verticals decided the year. Healthcare and life sciences, about 19% of revenue, grew 59%. TMT, about 25% of revenue, shrank around 5%. PL says the TMT overhang has largely stabilised, which is a hope, not a delivered fact.

The latest quarter sharpens the point. Q1FY27 revenue grew 20% in rupees, but management's own figures put constant-currency growth at 9% year on year and −0.4% quarter on quarter. That 9% is below the roughly 11.3% PL assumes for the full year, so the remaining three quarters must run faster just to get there.

TMT fell 22% in Q1FY27, while healthcare grew 69%. Concentration is easing slowly, with the top 10 clients at 51.8% of revenue against 55.9% a year earlier. Even so, more than half of revenue still rests on ten accounts.

Constant-currency growth of 9% is not a platform growth rate.

Client stickiness is also thinning. Net revenue retention was 151% in FY23, then 110%, 121% and 117.4%. The top 10 accounts grew 11% in dollars in FY26 and the top 20 grew 15%.

3	The Platform Story Is a Rounding Error

Bulls love the 63% revenue CAGR at Fractal Alpha. Growth from $2 million to $10 million looks spectacular only until you place it next to $374 million. And the businesses inside Alpha are still loss-making.

Case File: Fractal Alpha, FY26

Alpha revenue	$10 million, 2.7% of group revenue.
Adjusted segment margin	−16.1% (FY25: −39.9%).
Segment margin loss (PL text)	−35.7% (FY25: −40.0%).
Asper.AI	Revenue ₹67.8 crore, loss after tax ₹38.4 crore, 20 global clients.
Analytics Vidhya	Revenue ₹43.7 crore, profit after tax ₹3.2 crore, 9 global clients.

Twenty clients for Asper.AI and nine for Analytics Vidhya is a pilot-scale footprint, not a platform franchise. PL expects Alpha to be near breakeven in FY27E and FY28E. Until it is, the market is paying for a forecast.

Management's Q1FY27 commentary adds to the concern: Asper.AI revenue was flat year on year, while Analytics Vidhya grew 57%. PL's own text also puts the Alpha segment margin loss at −35.7% in FY26, against −40.0% in FY25. That is a far less flattering reading than the adjusted −16.1% above.

4	Earnings Quality: Tax, Capitalised R&D and One Good Quarter

FY26 profit looks better than the business earned. The effective tax rate was just 6.3%, ₹19.9 crore on a pre-tax profit of ₹317.7 crore. PL's FY27E assumes 25.5%.

Tax FY26 at 25% and adjusted profit drops from ₹303 crore to about ₹244 crore. On that normalised base, ₹708 is roughly 50 times earnings, not 40.

Q3FY26 tells the same story. The company booked a tax credit of ₹35.3 crore against a pre-tax profit of only ₹64.7 crore, which is how it reported a profit after tax of ₹100 crore.

Then there is capitalised R&D. It jumped from ₹25 crore in FY25 to ₹78 crore in FY26, more than triple. Capitalised spend rose from 17% to 34% of total R&D of ₹231 crore. Expense it, and FY26 EBITDA falls from ₹508 crore to about ₹430 crore, a margin of 13.0% instead of 15.4%.

Case File: Reported vs Underlying, FY26

Adjusted PAT, as reported	₹303 crore, at a 6.3% tax rate.
Adjusted PAT at 25% tax	About ₹244 crore.
P/E at ₹708	About 40x as reported, about 50x normalised.
EBITDA margin	15.4% reported, 13.0% after expensing capitalised R&D.
EV/EBITDA	About 20.5x reported, about 24x after expensing capitalised R&D.

Two more points. PL credits lower ESOP charges, alongside operating leverage, for part of Fractal.ai's margin recovery, and a lower non-cash charge is not an operating win. And FY27E leans on other income of ₹50.3 crore, about 10.6% of pre-tax profit, up from ₹15.5 crore.

Finally, dilution. Shares outstanding rose from 154 million in FY25 to 172 million. That is why FY27E EPS grows only 11.9% while adjusted profit grows 19.3%.

Add the associate. Fractal's share of loss in Qure.ai was ₹22 crore in Q1FY26 and ₹23 crore in Q1FY27. Management itself says Q1FY27 net income would have been ₹95.7 crore without it, against ₹72.3 crore reported. That is roughly a quarter of underlying profit, lost to an associate whose own quarterly revenue is ₹24 crore, even though that revenue grew 160%.

5	Cash Flow Does Not Back the Premium

Free cash flow fell from ₹314 crore in FY25 to ₹190 crore in FY26, a drop of 39%. Capital expenditure rose 2.6 times, from ₹82.8 crore to ₹218.7 crore. FCF was just 62.8% of profit after tax, and PL's own estimates keep conversion near 72% in FY27E.

At ₹708, that is a free-cash-flow yield of about 1.6%. A fixed deposit pays several times that, with no execution risk attached.

The balance sheet looks strong because of IPO money. The company raised about ₹1,078 crore in FY26 through share issuance. Strip out net cash of roughly ₹1,753 crore and the market is still valuing the operating business at around ₹10,400 crore.

6	Q1FY27 Cracks, a CFO Exit and Falling Targets

Look at Q1FY27 against the previous quarter. Net income was ₹72.3 crore, against ₹115.8 crore in Q4FY26, a fall of 37.6%. Management's adjusted EBITDA margin dropped from 22.1% to about 17%, and gross margin slipped to about 46% from 48.2%. The 92% year-on-year headline comes off a weak base of ₹37.7 crore in Q1FY26.

Q4FY26 now looks like the exception, not the run-rate. The market priced in the exception.

Chief Financial Officer Ashwath Bhat's resignation was announced on 6 July 2026, with effect from 24 July, at a company that went public during FY26. His successor, Shilpa Kedia of Johnson & Johnson, joins only on 19 October 2026. A CFO exit at a freshly listed company is never a neutral event, and investors deserve a clear explanation.

Then came the platform. On 23 September, Chief Commercial Officer and Cogentiq CEO Matthew Gennone resigned, effective 2 October, to take a CEO role elsewhere. Cogentiq is the flagship agentic platform at the centre of the platform story. Its team now reports to the group CEO on an interim basis, and commercial responsibility passes to the business units. The company expects no disruption, and the board chair also changed on 9 October.

The market voiced its doubt early. Fractal listed on 16 February 2026 at an issue price of ₹900 and opened on the NSE at ₹876, 2.7% below issue, after the offer was subscribed only about 2.7 times. PL's ₹1,260 initiation target arrived on that same listing day.

Even the sell side is retreating. PL Capital's target price has been cut from ₹1,260 on 16 February 2026 to ₹1,010 on 25 July 2026, a reduction of about 20%. Its rating has gone from Buy to Hold to Accumulate and back to Buy within six months. That is not conviction. That is chasing the price.

Goldman Sachs initiated at Neutral and named valuation as the key concern, even while projecting about 20% revenue growth a year to FY30. Ambit Capital initiated at Sell with a target of ₹710, citing AI-led disruption, high client concentration (53% of revenue from the top 10) and 37% exposure to sectors prone to generative-AI disruption. 

The stock now trades almost exactly at that Sell target, so my ₹460 goes further. It rests on estimates being cut, not on multiples alone.

7	What the Bulls Will Say

The bull case is real, and I will not pretend otherwise.

• Net revenue retention of 117.4% and an NPS of 78 point to sticky clients.
• Healthcare grew 59%, and a multi-year deal of $17 million or more with a leading US healthcare enterprise supports FY27.
• Alpha's gross margin improved by 120 basis points to 71.1%.
• Q1FY27 revenue grew 20% year on year and EBITDA grew about 39%.
• The company holds net cash and carries almost no debt.

All fair. But every one of these is already in a stock that costs 34 times FY27E earnings. The question is not whether Fractal is a good business. The question is whether it is a good business at ₹708.

8	Where Fair Value Sits

I anchor fair value on FY27E EPS of ₹21.0, using 171.97 million shares. A multiple of 22 times, which is already a premium for a business growing EPS at 12%, gives ₹460. That equals 2.4 times FY26 sales.

Price	Market cap	P/S FY26	P/E FY27E	Reading
₹708	₹12,176 cr	3.7x	33.7x	Current price.
₹500	₹8,599 cr	2.6x	23.8x	Still a premium multiple.
₹460	₹7,911 cr	2.4x	21.9x	SumanSpeaks fair value, 35% below ₹708.
₹350	₹6,019 cr	1.8x	16.7x	Stress case: FY27E EPS cut to ₹17.5 at 20x, which is also the services-peer sales multiple.

Note the honesty built into ₹460. Even there, a buyer pays 16 times FY28E EPS of ₹28.6, which assumes a 36% jump in earnings. That is generous, not harsh.

9	Where the Price Can Go in the Next Two Years

My thesis is simple. As more IT services firms move into AI, Fractal loses the scarcity premium it carries as an early mover. It was described at listing as India's first AI solutions company to go public, and that label carries a price. The label fades as rivals pivot and list.

The early signs are already in the numbers. Net revenue retention has slipped from 151% in FY23 to 117% for FY26, and 112% in Q4. Constant-currency growth is 9%. PL notes that large clients want vendors who take outcome accountability, which moves risk onto the vendor. And Fractal's own strategy leans on hyperscalers and model makers, who also sell AI tools to its clients directly. Ambit Capital, which initiated at Sell, puts 37% of revenue in sectors prone to generative-AI disruption.

To be fair, the edge has not visibly gone yet. Healthcare grew 69% in Q1FY27 and retention is still above 110%. My call is that the erosion shows up in pricing and growth over the next four to eight quarters.

Here is what each outcome is worth. I start from PL's FY28E EPS of ₹28.6 and apply my own arithmetic on revenue, margin, depreciation, other income, tax and share count, all taken from PL's model.

Scenario	What has to happen	FY28E EPS	Multiple	Price	vs ₹708
Premium holds	PL's numbers delivered (revenue +18.5%, EBITDA margin 18%) and a 30x multiple kept.	₹28.6	30x	₹858	+21%
PL delivers	Same earnings, but the multiple eases to 25x.	₹28.6	25x	₹715	+1%
Muted	Revenue +11%, EBITDA margin 16.5%.	₹25	22x	₹550	−22%
SumanSpeaks path	Revenue +9%, EBITDA margin back to about 15% as pricing pressure bites.	₹22	21x	₹462	−35%
Stress	Revenue +7%, EBITDA margin about 13%.	₹17.5	20x	₹350	−51%

Read the table honestly. The multiple alone does not take this stock below ₹500. Even at 22 times, PL's own FY28E EPS is worth ₹629. The sub-₹500 call needs FY28 earnings to be cut by roughly a quarter, to about ₹22. Note also that all five rows start from PL's FY27E revenue, which a 9% constant-currency quarter makes look generous.

This is the path I expect, if the thesis plays out.

• Next two to three quarters, starting with the Q2FY27 results: if constant-currency growth stays near 9% and the adjusted EBITDA margin near 17%, estimates get cut and the first leg is ₹550 to ₹600.
• By the FY27 results, around May 2027: FY28E EPS resets to ₹22 to ₹25, which at 21 to 22 times is ₹460 to ₹550.
• By late 2028: if commoditisation bites, ₹460 is fair value and ₹350 is the stress case. If Alpha scales and growth re-accelerates above 15%, I am wrong and ₹800 or more comes back.

10	The SumanSpeaks Verdict

Fractal Analytics is a decent AI services company priced like a platform. Dollar growth is decelerating, the platform arm is 2.7% of revenue and still loss-making, FY26 profit was flattered by a 6.3% tax rate, free cash flow fell 39%, and Q1FY27 margins slipped sharply from the previous quarter.

My fair-value view is ₹460, with a stress case of ₹350. The path I expect is ₹550 to ₹600 first, then ₹460 to ₹550 as estimates reset. At ₹708, the risk-reward is tilted against the buyer.

What would change my mind? Dollar growth sustained above 15%, Alpha reaching genuine breakeven, free-cash-flow conversion above 80% of profit, and a clean explanation for the CFO exit. Until I see those, I would not chase this stock.

For personalized stock market insights and guidance, feel free to reach out at: sumanm2007s@gmail.com | suman2005s@rediffmail.com

Disclaimer

This article is published by SumanSpeaks for general informational and educational purposes only. The author has over 25 years of capital markets experience. This is not a recommendation to buy, sell, or hold any security. Fair values shown are illustrative valuation scenarios built on published estimates, and peer multiples are approximate figures from September 2026, not a synchronised global snapshot. All data is sourced from public exchange filings, regulatory orders, and credible financial media. Readers must conduct independent due diligence before making any investment decision.

SumanSpeaks  |  Estd 2006  |  sumanspeaks.blogspot.com