The 25% Confession: What an IT CEO Just Told Us About the End of the Human Ledger

There is a specific moment in every market cycle when a CEO stops selling a vision and starts managing a surrender. It happened this week when the head of Hexaware Technologies, a mid-tier IT services firm, told the press that AI could slash IT project costs by 25%. On the surface, this is a simple efficiency forecast. But in the context of the global liquidity map, this statement is not a prediction—it is a capitulation. It is the first formal acknowledgment from the labor-arbitrage economy that the underlying asset of its entire business model—billable human hours—has just been repriced by a machine. For those of us who moved from the froth of DeFi summer into the cold reality of institutional balance sheets, this quote lands with a familiar thud. It is the sound of a legacy protocol admitting its tokenomics are broken. The ledger remembers what the market forgets: the IT services industry is the largest white-collar labor market on earth, a $1.2 trillion machine built on the assumption that coding, testing, and maintenance are inherently human activities. If that assumption cracks, the collateral damage extends far beyond a single company's stock price. It redraws the entire map of where value is created in the digital economy. Let's unpack the context. Hexaware is not Accenture, and it is not TCS. It is a roughly $1 billion revenue player with about 30,000 employees, sitting in the uncomfortable middle of the IT services food chain. Above it sit the giants who can afford $3 billion AI war chests. Below it lurk AI-native tools that can generate boilerplate code for pennies. The CEO's 25% figure is not plucked from thin air—McKinsey and Gartner have floated similar ranges for process automation. But the number is not the story. The story is the admission embedded within the quote: the CEO specifically mentioned balancing productivity gains with pricing strategy to maintain margins. That single phrase is the tell. It reveals a company that knows its clients have already done the math, knows they will demand a share of the efficiency dividend, and is pre-emptively offering a discount to keep them from walking. This is where my technical skepticism kicks in. Based on my experience auditing DeFi protocols during the 2022 bear market, I have learned that when a project announces a "sustainable yield" or a "revolutionary cost structure," the first thing you check is the collateral. Here, the collateral is human capital. If AI reduces the cost of a project by 25%, and the client immediately demands a 25% price cut, the IT firm's margin profile does not improve—it merely shrinks its revenue base. The CEO is essentially announcing a deflationary spiral for his own industry. The cost of goods sold drops, but so does the top line. To actually benefit, these firms must transition from selling hours to selling outcomes. That is a structural change akin to a liquidity provider moving from a simple AMM to a concentrated liquidity position. It requires sophistication, capital, and a tolerance for risk that most mid-tier firms do not possess. We built the cathedral before the saints arrived. For three decades, the Indian IT sector built its cathedral on the principle of wage arbitrage. You could hire a competent engineer in Bangalore for a fraction of the cost of a counterpart in San Francisco. The entire global supply chain of software development was optimized around this labor gradient. AI does not attack the gradient; it attacks the need for the labor. When the CEO says AI cuts costs by 25%, he is implicitly saying that 25% of his workforce's current tasks are either redundant or can be automated. The industry is staring down the barrel of a jobless recovery. IDC projects that AI will displace 15-20% of junior programmer roles by 2028. For a country like India, where the IT sector employs over 5 million people directly and supports millions more indirectly, this is not a margin issue—it is a macroeconomic stability issue. The social contract that underpinned the middle class of an entire nation is being rewritten by a prompt engineer. Now, let me offer the contrarian angle, because the herd is always looking in the wrong direction. The mainstream take is that AI is a threat to IT services. The contrarian take is that the 25% figure is actually a defensive negotiation tactic that signals a massive transfer of value from the IT providers to the AI infrastructure layer. When Hexaware cuts costs by 25%, who captures that value? Not necessarily the client. The client might demand lower prices, but the real winners are the companies providing the AI tools—the GPU providers, the model API vendors, the cloud platforms. IT service firms are about to become middlemen in a transaction between enterprises and AI compute. Their margin will be squeezed from both sides: clients want cheaper outcomes, and AI providers charge premium prices for the intelligence that enables those outcomes. This is the classic "picks and shovels" dynamic. The miners of the IT world—the labor—are being replaced by the mining rigs—the AI models. And the rigs are sold by Nvidia, OpenAI, and Microsoft. The IT service firm is left holding the electricity bill. This brings me to the investment thesis. For years, the valuation of IT services companies was anchored on headcount growth. You valued TCS by multiplying its engineers by a revenue-per-employee metric. That logic is now obsolete. The market is shifting to a new metric: AI leverage. How much revenue can a company generate per unit of human input? A firm that can grow revenue by 10% while keeping headcount flat is worth more than a firm that grows revenue by 20% but needs 20% more bodies. Stability is a myth; liquidity is the only truth. In this case, the liquidity is not cash—it is the flow of high-value work. The firms that will survive are those that can prove they can deliver the same outcome with 25% less human capital, and then keep the margin. That is a brutal transition. It requires firing people, retraining others, and renegotiating contracts. Most CEOs will not have the stomach for it. The ones who do will emerge as the new giants. The ones who don't will be acquired for their client lists. The deeper issue here is trust. Code is law, but trust is the currency. The IT services industry runs on trust—trust that the vendor will deliver, trust that the code is secure, trust that the project will not fail. AI breaks that trust loop. If a client cannot verify that an AI-generated codebase is secure, they will not accept the 25% cost reduction. This is why we are seeing a rise in "verifiable compute" and on-chain integrity checks. The same way DeFi protocols needed audit layers to reassure users, IT services will need verification layers to reassure clients. The firms that can bridge the gap between AI-generated output and human-verified security will capture the lion's share of the value. The firms that simply throw AI at the wall and hope for the best will fail. Surviving the winter makes the spring inevitable. But this is not a winter—it is a season change. The IT services industry is not going to die, but its current form is going extinct. The 25% figure is the first snowflake of that extinction event. The real question is not whether AI will cut costs. It is whether the industry can reprice its own labor before the market does it for them. If they wait too long, they will be left with a cathedral full of empty desks. If they move fast, they might just build the new foundation. As I watch this unfold from my perch in Tallinn, I am reminded of the early days of DeFi. Everyone was focused on the APY, the yield, the get-rich-quick schemes. No one was looking at the collateral. And when the collateral failed, the whole house collapsed. Here, the collateral is the human workforce. The yield is the 25% cost reduction. The question is whether the industry can manage the withdrawal of that collateral without triggering a systemic bank run. The answer will determine not just the fate of Hexaware, but the shape of the global digital economy for the next decade.