The number arrived with the casual confidence of a weather forecast: 25%. Twenty-five percent of IT project costs, gone, evaporated by artificial intelligence. Hexaware's CEO offered this figure to the world, and the market barely blinked. But we should. We should always blink when a comfortable industry suddenly starts talking about its own obsolescence with a smile. Because behind that single, tidy percentage lies a confession far more significant than any efficiency metric. It is an admission that the old architecture of trust—built on billable hours and human leverage—is crumbling. And in its place, something far more interesting is struggling to be born.
Let me set the stage. For three decades, the global IT services industry ran on a remarkably simple formula: bodies times hours equals revenue. TCS, Infosys, Wipro—these firms built empires on the arbitrage of human labor, moving work to where skilled people cost less. The model worked because the unit of value was the person. You bought a developer, a tester, a project manager. The software was almost incidental. This is the industry that employs roughly 2.5 million people in India alone, generating over a trillion dollars globally. It is, in the truest sense, a human industry. And now, its leaders are publicly declaring that the human is the bottleneck.
Based on my years auditing protocols and building educational frameworks around decentralized systems, I've learned to read these public pronouncements as carefully as I read smart contract code. The 25% figure is not a technical specification; it is a communication strategy. The CEO knows that clients have already heard the AI gospel. They've seen the demos. They've read the McKinsey reports suggesting 20-40% efficiency gains are possible across software development. The question was never whether costs would fall. The question was who would capture the savings. By volunteering the number first, Hexaware is attempting to frame the narrative before clients frame it for them. This is defensive positioning disguised as aggressive optimism.
The uncomfortable truth, which the CEO's statement dances around without ever addressing, is structural. Human labor constitutes 60-80% of IT project costs. If AI genuinely removes a quarter of that—and the industry consensus suggests this is plausible—then the billable hour model loses its logical foundation. You cannot charge clients for time when you are selling them outcomes that require less time. This is the pricing paradox at the heart of the entire AI transition. Cost down 25%, revenue down 25%, margin unchanged. The only winners are the clients. And the clients know this. The power dynamic has already shifted, and every IT services CEO is now negotiating from a position of acknowledged weakness. The 25% is not a projection of future savings; it is a retroactive admission that the industry's pricing power is gone.
But let's push further into the competitive landscape, because this is where the story gets genuinely interesting. Hexaware is not TCS. It is not Accenture. It is a mid-tier player with roughly three billion in annual revenue, caught in the worst possible position: too small to out-invest the giants in AI infrastructure, too large to pivot with the agility of an AI-native startup. The giants are spending billions—Accenture has committed three billion to AI initiatives, Infosys built an entire platform around it. The startups are building tools like Devin and Cursor that threaten to commoditize the very coding work that forms the industry's core. Hexaware's announcement is a signal to both directions simultaneously. To the giants: we are relevant. To the startups: we are adapting. And to the capital markets, where the company recently returned after a private equity period, it is a simple plea: do not discount us yet.
This is where I must introduce a contrarian angle, because the industry consensus is too comfortable. Everyone assumes that AI-driven cost reduction is an unqualified good for the IT services industry. I disagree. I believe this transition represents the most significant destruction of shareholder value in the industry's history, precisely because the CEOs are framing it as an opportunity. Think about it. If costs drop 25% and that savings flows entirely to clients through price reductions—which is what the Hexaware CEO's careful language about balancing productivity gains with pricing strategy actually implies—then the industry is voluntarily walking into a margin compression event. The only way to avoid this outcome is to shift to outcome-based pricing, where the vendor captures a share of the value created rather than charging for inputs. That shift, however, requires a level of trust that the current market structure does not support. Clients do not trust vendors to price outcomes fairly when they know the vendor's costs have collapsed. This is not a technical problem. It is a trust problem. Code is law, but humans are the protocol. And the protocol is currently being rewritten under extreme pressure.
I saw this dynamic play out in DeFi during the 2020 summer, when a protocol I audited faced a similar structural challenge. OpenYield had built a flash loan module that was technically elegant but conceptually fragile. The code worked. The trust model did not. The developers were so focused on the efficiency gains of their design that they ignored the human incentive structures that would ultimately determine whether the system survived. We found the reentrancy vulnerability before launch, but the deeper vulnerability was in their assumption that users would behave rationally. They did not. They panicked. They withdrew. The protocol failed not because the code was broken, but because the humans operating it lacked confidence in the system's fairness. The IT services industry is facing the same moment now. The AI efficiency gains are real. The question is whether the human systems—pricing models, employment structures, client relationships—can adapt at the same speed.
The employment implications alone should give us pause. IDC projects that AI will eliminate 15-20% of entry-level programming positions by 2028. For India's IT sector, this is existential. The industry's entire educational pipeline is built around producing large numbers of junior developers. If those roles vanish, the industry loses its training ground for future senior talent. We are not just automating tasks; we are dismantling the career ladder that built the industry. And the replacement jobs—AI architects, prompt engineers, systems integrators—require entirely different skill sets that the current workforce does not possess. The transition is not a gentle evolution. It is a discontinuity. And the CEOs who talk about 25% cost savings without addressing this human cost are either willfully blind or strategically deceptive. Neither option inspires confidence.
Here is what the market should be watching, and here is where I believe the real opportunity lies. The firms that survive this transition will not be those with the most impressive AI tooling. They will be those that fundamentally restructure their relationship with clients. The shift from time-and-materials to outcome-based pricing requires a level of trust that the current market structure does not support. And trust, as I have learned through years of building educational platforms in this industry, is not created by efficiency. It is created by transparency, by demonstrated competence, and by a willingness to share risk. The IT services company that can say to a client, 'we will build this system for a fixed price based on the business outcome it generates, not the hours we spend on it,' and actually deliver on that promise, will capture the market. The 25% cost reduction is merely the entry ticket to this game. The real prize is a new trust architecture for software delivery.
This brings me back to the original question: what does the Hexaware announcement actually mean? I think it means the industry has finally reached the point where the old narratives no longer work. The AI wave has made the traditional model unsustainable, and the CEOs know it. They are offering these numbers—25%, 30%, 40%—as a form of self-protection, an attempt to control the narrative of their own disruption. But the numbers do not matter. What matters is the structure beneath them. We built trust in the chaos, not despite it. The chaos of AI-driven transformation is our opportunity to rebuild the industry's foundations on something more honest than billable hours. The firms that understand this—that see AI not as a cost-cutting tool but as a catalyst for a more mature, outcome-driven relationship with clients—will lead the next decade. The rest will fade, their 25% promises becoming footnotes in the story of an industry that could not let go of its past.
I have no doubt that the next two years will bring a wave of consolidation. The mid-tier firms will be squeezed between giants with capital and startups with agility. Some will be acquired. Others will simply disappear. The market will reward those who embrace the outcome-based model with genuine conviction, not those who merely talk about it in press releases. And the clients? They will ultimately benefit, but only after navigating a period of uncertainty where the quality of AI-generated code, the reliability of automated testing, and the accountability of algorithmic decision-making remain open questions. Trust is earned in drops, lost in buckets. The AI transition is testing whether this industry can earn that trust anew. The tech is ready. The question, as always, is whether we are.
Hold through the noise, build through the silence. The noise right now is deafening, filled with promises of cost reductions and efficiency gains that may or may not materialize as described. But the silence—the quiet work of restructuring business models, retraining workforces, and rebuilding client relationships on more honest foundations—is where the future is being built. I intend to be there, teaching and learning, as the industry discovers that its greatest challenge was never the technology. It was always the humans. And we are capable of remarkable things when we stop pretending that the numbers tell the whole story.