Capgemini Up, Cognizant Down, and Fujitsu Falls 16%

Three of the world's largest IT service providers released numbers in less than 48 hours, with each telling a different story about the same market. Capgemini raised its targets for 2026 on Friday (30) after generating 6.073 billion euros in revenue for the third quarter, an increase of 11.7% at constant currency. Cognizant on Thursday (29) narrowed its growth forecast to 4.0% to 4.4%, down from a previous range of 4.0% to 5.5%. And Fujitsu's shares plummeted 16.2% in Tokyo on Friday, hitting the daily limit for decline, following a disappointing six-month performance.

Capgemini: Growth through Acquisition

According to Capgemini's statement, revenue from North America increased by 18.3% at constant currency for the quarter, and CEO Aiman Ezzat described a strong pace in the region and the UK, with a "clear acceleration" in France.

This number needs context. Much of the growth stems from WNS, the Indian business process operator acquired by Capgemini and consolidated since Q4 2025, along with Cloud4C. Organic growth is lower than the headline 11.7%. New contract bookings of 5.401 billion euros rose only 3.4%, a pace well below that of revenue. For Ezzat, the underlying growth was close to that recorded in Q2 and above the industry average.

Cognizant: Margin Rises, Discretionary Demand Cools

Cognizant generated $5.7 billion in revenue for the quarter, an increase of 4.6%, with an adjusted operating margin of 16.3%. The company cut the top end of its revenue forecast for 2026 and set adjusted earnings per share between $5.72 and $5.79. Management attributed the adjustment to macroeconomic uncertainty, which weighs on discretionary spending, especially in healthcare and consumer-focused sectors. Shares fell following the announcement.

Contract backlog over the past 12 months totaled $28.7 billion, up 4%, with a ratio of new contracts to revenue of about 1.3 times. Ravi Kumar, CEO of Cognizant, stated that over 50% of the code produced by the company is already assisted by AI. This data is central to the model: when half of the code is machine-assisted, project revenue may decrease even with the same number of contracts.

Fujitsu: The Cost of Loss-Making Projects

In the Japanese case, the problem was execution. Fujitsu attributed the weak six-month performance to the emergence of loss-making projects, lower engineer utilization in a quarter with fewer business days, and the rise in memory chip costs, which compressed hardware margins. The company maintained its annual projection, requiring a much stronger second half. The market did not buy it: a 16.2% plunge in a single trading session.

The Misguided Interpretation

The temptation is to treat the three figures as a scoreboard on AI, with Capgemini winning and Fujitsu losing. The axis is wrong. Capgemini is growing largely because it acquired scale in business operations in India, where AI reduces process costs, and clients pay based on results. Cognizant relies more on discretionary engineering projects, which are the first category customers delay when conditions worsen. Fujitsu suffers from fixed-price contracts in Japan, where a poorly estimated project results in direct losses.

The counter-argument deserves acknowledgment. Cognizant argues that AI increases demand by opening new projects and launched the AI Nexus platform during the quarter to govern cost and quality of agents for clients. The backlog of 1.3 times Cognizant's revenue supports part of this thesis. The opposite risk is represented by those who only see cannibalization: Cognizant's adjusted margin of 16.3%, up 0.3 points from a year earlier, shows that automation also protects profitability.

The Implication

In India, the effect is direct. Cognizant maintains most of its workforce in the country, and WNS added tens of thousands of operations professionals to Capgemini in cities like Mumbai and Pune. A contract backlog growing at 3.4% versus revenue at 11.7% indicates that hiring in these centers is likely to follow the already signed volume, rather than new expansion.

In Japan, Fujitsu's decline reignites the discussion around fixed-price contracts in legacy systems for government and banks, a model that Japanese integrators have used for decades.

In Brazil, where Capgemini and Cognizant serve banks and insurers, the impact is felt through pricing. With half of the code assisted by AI, Brazilian and European financial clients are beginning to demand that productivity gains are reflected in delivery fees, not just in the supplier's margin. The next round of renegotiating IT service contracts will precisely measure this division.