AI Is Not Killing Cognizant and That Should Make You Rethink Your Whole Strategy

February 26, 2026

Sometime last year - I think it was a Tuesday, we were eating lunch - Derek told me that AI was going to destroy the entire IT outsourcing industry within eighteen months. He said it with real confidence. He's usually that confident right before he's wrong. He also once told me The Phantom Menace was the best-reviewed film of 1999, which I did not look up because I didn't want to have that conversation again.

I mention Derek because the thing he said is basically the dominant narrative in business media right now. AI is coming for the outsourcing firms. The headcount model is dead. You don't need 300,000 engineers in India if a coding agent can do it at the cost of electricity. It's a compelling story. It's also currently being disproved in real time by a company called Cognizant, and I think most people running businesses are missing what that actually means.

What Actually Happened

Here's the news, just straight. Cognizant posted full-year 2025 revenue of $21.1 billion, up from $19.7 billion in 2024. That's a 7% increase, year over year, for a company that the conventional wisdom had basically written off as AI roadkill. Total headcount as of December 31, 2025 was 351,600 - an increase of 14,800 from December 31, 2024. They added nearly 15,000 employees in a single year. While AI was supposedly abolishing the need for human IT workers.

CEO Ravi Kumar called 2025 "a defining year" in which the company "put its AI builder strategy in motion," invested in talent, strengthened its partnership ecosystem, and advanced AI platforms - which helped them sign 28 large deals with large deal total contract value growth of nearly 50% year-over-year.

Fifty percent deal growth. In a year when AI was supposed to be eating their lunch.

Operating margin came in at 16.1%, up from 14.7% in 2024. And full-year 2026 revenue is expected to hit $22.14 to $22.66 billion, growth of 4.9% to 7.4%.

This isn't a company in survival mode. This is a company accelerating.

The Narrative That Won't Die

To be fair to the doom camp, there's a real version of this story that has legs. India's IT services giants - TCS, Infosys, and Wipro - face mounting contract cancellations as AI erodes the cost advantage that underpinned the country's outsourcing boom, because the model was built on providing skilled engineers at lower wages than their Western counterparts, which is now threatened by AI coding agents performing similar work at marginal cost.

As one Citrini Research report bluntly put it: the entire model was built on one value proposition - Indian developers cost a fraction of their American counterparts. But now "the marginal cost of an AI coding agent had collapsed to, essentially, the cost of electricity."

And the hiring numbers at some of these firms back that up. Wipro increased its payroll by 6,500 people and Infosys hired 5,000 more - muted growth by their standards - while TCS and HCL went backward by 11,000 and 261 people respectively. Over the last year, those four companies added just 3,910 staff, an unusually slow rate of hiring.

So the doom story is real. It's just not the whole story. And Cognizant is the evidence.

Why Cognizant Is the Outlier That Changes Everything

Here's my actual take: Cognizant didn't survive AI disruption. They repositioned into it. There's a significant difference between those two things, and that difference is what most business owners are failing to understand about their own situations.

As of the end of 2024, Cognizant had 1,200 AI projects underway, focusing on cloud migration, data modernization, and agentification. They didn't wait to see what AI would do to their core business. They decided they were going to be the people who build the AI infrastructure for everyone else.

Cognizant is betting it can embed agentic AI across enterprises, expand its total addressable market and drive "hyper-productivity" for its customers - a strategy driven by the ability to design enterprise AI agents that work across systems, create industry-specific large language models, and build its own platforms like Cognizant Neuro AI, built on Nvidia's stack.

The acquisition strategy is part of this too. Cognizant's $1.3 billion acquisition of Belcan, completed in August 2024, brought 6,500 engineers and expertise in aerospace and defense engineering research and development, which directly aligns with Cognizant's AI and digital engineering strengths. They bought domain expertise. In an AI world, domain expertise is the moat. Code generation is commoditizing. Knowing which code to generate, for which business problem, in which industry context - that's still human.

While IT services companies could be disrupted by agentic AI, there's a strong case that these integrators are actually best positioned to do well building AI agent systems - because they work horizontally across business functions and have deep domain expertise and industry-specific knowledge. That insight is worth sitting with for a minute, because it applies to almost every sector facing AI pressure right now.

Watercolor illustration of a lush green plant growing vigorously out of an old metal toolbox on a workbench, while smaller wilted toolboxes with dying leaves sit around it, bathed in warm golden light — Wanted something that felt like unexpected growth out of a thing everyone had written off. Derek looked at it and said the upside-down wrench was a mistake but I think it actually works.

The Thing Nobody Is Saying Out Loud

The AI-kills-outsourcing narrative assumes that AI makes human expertise worthless. What's actually happening is that AI makes undifferentiated human expertise worthless. Bug fixes, documentation, routine testing - yes, those are going. Traditionally, Indian IT companies thrived by offering cost-efficient services, large workforces, and long-term outsourcing contracts - but automation and generative AI could reduce the need for large teams handling repetitive coding, testing, and support work.

Cognizant understood this earlier than most. Their internal numbers tell the story. AI is now creating 30% of Cognizant's code. But their revenue went up 7% and they hired 15,000 more people. How? Because those people aren't writing the code AI writes. They're doing the harder work - designing the systems, integrating the agents, managing client relationships, solving problems where the context is complex and the stakes are high.

I tried to understand the AI agent platform stuff when I was messing around with some workflow automation tools last year. I had this whole process set up and it kept running on the wrong day - just Saturdays, every Saturday, for no reason I could identify. Turned out I had the scheduling parameters configured backwards, triggering on the start of the week offset instead of a fixed day. I unchecked a box and it stopped. The point is: someone has to know which box to uncheck. In a $500 million enterprise AI deployment, that "box" is an organizational problem wrapped in a technical problem wrapped in a change management problem. AI doesn't know how to uncheck that. Cognizant does.

This is why their CEO said, when talking about building the AI orchestration layer: "Every software ecosystem is building agents. We built an orchestration layer, which we believe is a gap today, and can get agents from different ecosystems to talk to each other and generate the kind of productivity clients are looking for."

That's not a job AI can do. That's exactly the kind of integration work that gets harder and more valuable as AI gets more powerful.

What This Means If You're Running a Business Right Now

I want to be direct about this, because I think there's a real strategic error happening at a lot of companies right now.

The companies getting hurt by AI are the ones whose value was purely in executing predictable, scalable, low-margin tasks. The companies growing through AI are the ones who figured out which parts of their work are hard to automate - judgment calls, context-heavy decisions, trusted relationships, deep domain knowledge - and doubled down on exactly those things. The error isn't using AI wrong. It's not asking the question at all.

Cognizant shed headcount in 2024. Their workforce shrank by 10,700 employees year-over-year in Q4 2024, bringing total headcount to 336,800. The doom headlines wrote themselves. Then they hired 15,000 people in 2025 while growing revenue 7%. The people they let go were doing the work AI can do. The people they hired are doing the work AI makes more valuable.

That's the playbook. And it applies whether you're running an IT firm or a marketing agency or a logistics company or literally anything else.

Tory - he's been going through a lot lately, the divorce, the car situation - told me last week that he's thinking about using AI to replace his entire outreach process. Just automate the whole thing. I didn't want to be discouraging because genuinely, I want him to succeed. But what Tory is good at is the conversation part. The empathy. The way he can tell someone is struggling before they say it out loud. If he automates the outreach and shows up to a canned AI intro, he's just made his worst asset cheaper and his best asset invisible. That's the wrong direction. For B2B sales in particular - and we get into the mechanics of this in pieces like The AI Panic Splitting Office Teams in Half - the panic response tends to automate the thing that was differentiating you.

The Broader Picture Is More Complicated Than Either Camp Admits

I'm not saying AI isn't disrupting outsourcing. It clearly is. The shift is leading to lower discretionary IT services spending growth, delays in project ramp-ups, and reduced outsourcing intensity, with pricing and volume headwinds persisting as global technology firms rationalize headcount to fund higher AI investments.

And some of the macro risk is real. Hyperscalers like Microsoft, Google, Meta, and Amazon are sharply increasing capital expenditure on AI infrastructure, with the top four cloud companies expected to invest up to $630 billion in capital expenditure in 2026, a more than 60% increase over the record $390 billion spent in 2025. That money is going to chips and data centers, not to outsourced headcount. Some of that demand destruction is permanent.

But Cognizant's response to all of this - position yourself as the firm that makes sense of all that AI infrastructure for the enterprise clients who can't do it themselves - is genuinely smart. The company explicitly framed its strategy around building "a resilient and durable portfolio that positions us to win in the AI era," with four straight quarters of organic year-over-year revenue growth and accelerating bookings including two $1 billion deals.

Two billion-dollar deals. In a year when AI was supposed to be ending their business model.

The Lesson Is Simpler Than It Looks

Stephanie sat in on our last strategy meeting - she does this occasionally, she's technically an advisor, she once asked if we could "just buy a better algorithm" as if that's a thing you acquire like a desk chair - and she said she thought every company should just wait and see what AI does before making any moves. Just hold. Let the picture clarify.

That's exactly wrong. Cognizant's results are evidence that the companies winning right now are the ones who made decisions early - about what they were going to be, about where human expertise still mattered, about which capabilities were worth acquiring. By the time the picture is completely clear, you're already behind the firms that moved when it wasn't.

There's a real risk that rapid advances in AI could replace traditional coding and integration services, undermining demand for some firms' core offerings. That risk is legitimate. But the firms that are growing through it aren't the ones who avoided AI. They're the ones who hired 15,000 people to build AI agent factories while their peers were freezing headcount and hoping for clarity.

The narrative that AI is killing the outsourcing industry is partially true and mostly useless. The more interesting question - the one that actually matters for your business - is whether your value lives in the part AI is replacing or the part AI is amplifying. Cognizant apparently figured that out. The numbers are pretty clear about what happened next.

Derek is going to tell me I'm wrong about this. He always does. Then he's going to explain something about Star Wars to me and I'm going to nod and think about something else. Some things AI cannot change.