Cognizant: Priced for an AI Apocalypse That Isn't in the Bookings

Cognizant ($CTSH) is down 48% this year and just got thrown out of the Nasdaq-100. The index funds had no choice, they had to sell, none of it was about the business. Bookings are at a record, the buyback just doubled, and the AI deals the market is terrified of are the same ones the company is signing.

$CTSH closed at $42.59 on Friday July 10th, it started the year at $82. The stock has been cut nearly in half in seven months while revenue grew, bookings hit records and full-year margin guidance was raised. The market is pricing an AI extinction event. The filings, and the reason for the June selling, tell a different story.

What Cognizant Does

Cognizant is an IT services firm, it sells the people and systems that run enterprise technology. Application development and maintenance, cloud migration, data work, and business process operations, delivered largely out of India at an offshore cost base. It has around 358K employees and reports across four segments: Financial Services; Health Sciences; Products and Resources; and Communications, Media and Technology.

The model, historically, is people. You sell hours, you staff a client's technology function more cheaply and more flexibly than they could in-house, and you embed so deeply that switching you out is painful. That embedding is the moat, alongside scale and two decades of domain knowledge in regulated industries like banking and healthcare. FY2025 revenue was $21.4B with net income of $2.2B and free cash flow of $2.7B. It is a large, profitable, cash-generative business.

It is also the model the market now thinks AI breaks.

The Q1 2026 Numbers

Cognizant reported its Q1 results on April 29th, and they were good. Revenue was $5.41B (+5.8% YoY), landing in the upper half of guidance. Adjusted EPS was $1.40, up from $1.23 a year earlier, a 13.8% increase, GAAP EPS was $1.39, and net income was $662M, roughly flat YoY. GAAP operating margin was 15.6%, down 110 basis points on last year's quarter, which included a one-off gain on a property sale in India, on an adjusted basis the margin ticked up 10 basis points.

The bookings are the number that matters. Quarterly bookings grew 21% and trailing-twelve-month bookings grew 11%, with seven large deals signed in the quarter including one mega deal above $500M and over 70% growth in the total contract value of large deals. Financial Services, the largest segment, grew 12% YoY. The company launched Project Leap, a restructuring targeting $200M to $300M in in-year savings against $230M to $320M in charges, and it raised full-year adjusted operating margin guidance to 16.0%-16.2% and full-year adjusted EPS guidance to $5.63-$5.77.

Then there is the one number the whole thesis hinges on, Q2 constant-currency revenue growth is guided to 3.2% to 4.7%, and it reports on July 29th. Hold that date, everything comes back to it.

The AI Question

Here is the bear case, if AI writes code, automates testing, and runs process, why does any enterprise need 358K consultants billing by the hour? The unit Cognizant sells is exactly the unit AI is getting cheaper at producing. Accenture made it worse in June when it guided softer and flagged weaker bookings, the read-through hit the entire IT-services group, and Cognizant sits right in the middle of it. It is the key question hanging over the sector.

But it is worth knowing the mix before you accept the whole book is at risk. Only about 41% of Cognizant's revenue is time-and-materials (the part where AI compressing hours actually costs the company money). Half of it, 50%, is fixed-price, where doing the same work faster with AI widens the margin rather than shrinking the bill, and the small remainder is transaction-based. So the bear case bites hardest on two-fifths of the revenue, and on the other half a productivity gain is a good thing.

The rest of the counter is in the bookings. If AI were gutting demand you would see it in the order book first, and the order book is at a record, up 11% over the year and up 21% in the quarter, with clients signing bigger not smaller. Demand is not collapsing, it is growing. But the honest response to that is that bookings lag revenue, a signed deal takes quarters to convert, and the thing the bears actually want to see is revenue growth holding up as AI compresses the work inside each deal. Which is why July 29th matters, it is the first clean read on whether the bookings are converting or quietly eroding.