The headline was neat. Clean. Almost too clean. Hexaware's CEO, R. Srikrishna, told the world that AI could cut IT project costs by 25%. The market nodded. The press moved on. But as someone who spent 2017 auditing whitepapers for projects that promised revolutionary change and delivered only empty token contracts, I've learned to read the subtext before the text. This isn't a story about efficiency. It's a confession. A public acknowledgment that the oldest arbitrage in the digital economy—selling human hours at a markup—is finally hitting its liquidity wall.
Liquidity doesn't disappear from a market. It just migrates. And the liquidity that once flowed into the veins of the 1.2 trillion-dollar IT services industry is now looking for a new host. AI isn't the disruptor here. It's the catalyst that revealed the underlying fragility of a business model built on time, not value.
The 25% figure isn't the story. The story is what that number represents: a repricing of human capital at an industrial scale. Let me break down the macro mechanics.
The Context: A Global Labor Arbitrage Under Siege
To understand what's happening, you need to map the global liquidity flows that built this industry. The IT services boom was never about technology. It was about geography. For three decades, capital flowed from high-wage Western economies to low-wage talent pools in India, the Philippines, and Eastern Europe. The arbitrage was simple: a developer in Bangalore cost 20% of a developer in San Francisco. The margin was the difference. TCS, Infosys, Wipro—these aren't tech companies. They're labor aggregation machines with a technology veneer.
I watched this model work in 2020 during DeFi Summer. The composability of Aave and Uniswap created a 4,000% TVL surge in six months. But the underlying logic was the same: capital finding the most efficient yield. The IT services model was the traditional finance equivalent—capital finding the cheapest labor. And just like algorithmic stablecoins in 2022, the model worked until it didn't.
The 25% cost reduction Hexaware's CEO mentioned aligns with McKinsey and Gartner projections of 20-40% efficiency gains from AI integration. But here's what those projections don't tell you: they're measuring the death of the unit economics, not just the cost line. IT services' cost structure is 60-80% human capital. When AI compresses the human component, the revenue base—priced per person-day—compresses with it.
The Core: AI Leverage vs. Human Arbitrage
The dialectical tension in Hexaware's statement is the giveaway. Srikrishna said the company needs to balance productivity gains with pricing strategy to maintain profit margins. That sentence is the smoking gun. It reveals a structural contradiction that can't be resolved within the existing framework.
Let's model this. Suppose a traditional IT project costs $1 million. Human costs account for $750,000. AI reduces the human input by 25%. The cost drops to $562,500. But here's the problem: if the client knows AI cut the cost, they'll demand a price cut. The new price might be $800,000. The company saves $187,500 on costs but loses $200,000 on revenue. Net result: margin compression.
Skepticism isn't about doubting the efficiency gains. It's about questioning who captures them. In the current pricing model, the gains are split between the client (lower prices) and the IT vendor (higher margins). But clients have all the leverage now. They know the game has changed. They can go to a competitor, or worse, adopt AI tools themselves. The power dynamic has shifted from a seller's market to a buyer's market. This is the liquidity migration I mentioned earlier.
Based on my experience auditing 50+ whitepapers during the ICO boom, I can tell you when an industry starts talking about "transformation" publicly, it's usually a defensive move. In 2017, every project claimed to be "revolutionary." They weren't. They were chasing liquidity. Hexaware's announcement is similar. It's a signal to clients: "We know the game has changed, and we're willing to cut prices to keep your business." It's a signal to investors: "We're adapting, so don't sell our stock." But it's also a signal of weakness—an admission that the core value proposition (cheap, scalable human labor) is no longer defensible.
The deeper issue is the transition from Time & Material (T&M) pricing to Outcome-Based pricing. AI makes the former obsolete. If a machine can do the work in 15 minutes, charging by the hour is absurd. But Outcome-Based pricing requires IT companies to take on more risk. They need to deliver specific business results, not just code. This is a fundamental business model change, not a technical upgrade. Most IT services firms—even the giants—aren't ready for this transition.
The Contrarian Angle: The Decoupling Myth
The mainstream narrative is that AI will make IT services more efficient, lower costs, and create new value. The contrarian view—the one I've arrived at after watching the Terra-Luna collapse in 2022—is that AI will decimate the IT services industry's revenue base faster than it creates new opportunities. The 25% cost reduction is real, but it's a liquidation event, not a growth story.
Consider the industry's scale. 25 million IT professionals globally. If AI reduces the need for entry-level developers by 15-20% by 2028, that's 3.75 to 5 million jobs lost. The structural impact on India's economy, which relies on IT services for a significant portion of its GDP, would be profound. I analyzed the Terra-Luna withdrawal rates in 2022 and saw how a death spiral accelerates when the underlying collateral is fake. The IT services industry is facing a similar dynamic: the collateral is human labor, and AI is making it less valuable by the day.
The real question isn't whether AI will cut costs. It's whether the industry can pivot fast enough to avoid a margin collapse. The answer, based on my 2024 analysis of Bitcoin ETF flows, is that institutional adaptation is slower than market disruption. When the Spot Bitcoin ETFs launched, I modeled the daily inflows against traditional equity fund flows. The dampening effect on volatility was real, but it took months to materialize. Institutional capital is risk-averse and slow. The same will happen in IT services: the giants will adapt, but the mid-tier players like Hexaware will be caught in the transition.
The competitive landscape is a pincer movement. From above, Accenture has announced $3 billion in AI investment. Infosys has built the Topaz platform. Wipro is spending $1 billion on AI transformation. From below, AI-native tools like Cursor and Devin are eating away at the lower end of the market—the simple coding tasks that were the entry point for IT services. Hexaware, with roughly $1 billion in annual revenue, is caught in the middle. It doesn't have the scale to out-invest the giants or the agility to out-innovate the AI-native startups.
But there's an even darker implication. The 25% cost reduction is a floor, not a ceiling. AI capabilities are improving exponentially. As models become more capable, the cost reduction will accelerate. The industry could see 40-50% cost reductions within 3-5 years. That would mean the IT services market shrinks from $1.2 trillion to $600-700 billion in constant dollars. The value isn't destroyed—it's transferred. Some goes to clients (lower prices). Some goes to AI infrastructure providers (OpenAI, Anthropic, Microsoft). Some goes to shareholders (if companies can maintain margins). But the aggregate revenue pool for human-centric IT services will shrink.
The investment implications are stark. Traditional IT services stocks trade at 15-25x P/E based on the assumption of steady growth. But if the revenue base is shrinking, those multiples are unsustainable. The market is starting to realize this, which is why we're seeing a bifurcation. Companies that can demonstrate "AI leverage"—the ability to do more with less and capture the efficiency gains—will be rewarded. Those that can't will see their valuations compress to utility levels. Hexaware's statement is an attempt to position itself in the former category, but the market will demand proof, not promises.
The Takeaway: A New Liquidity Map
The real signal in Hexaware's announcement is about the location of value creation, not the cost structure. Liquidity is migrating from human capital to algorithmic capital. The IT services industry was built on a geographic arbitrage. That arbitrage is now obsolete. The new arbitrage is between those who can leverage AI and those who can't.
For investors, the question is whether you're holding the picks-and-shovels of the AI economy or the businesses being disrupted by it. For IT services executives, the playbook is clear: pivot to outcome-based pricing, build proprietary AI tools, and accept that the "human scale" advantage is gone. For the broader market, this is a preview of what happens when any industry built on information asymmetry meets general-purpose AI. The margin compression, the job displacement, the business model collapse—this is the pattern. I saw it in the stablecoin collapse. I'm seeing it now in IT services. And it's coming for every industry that sells time instead of outcomes.
The 25% number will be remembered as the moment the IT services industry admitted its core value proposition was no longer viable. The question is no longer "if" AI will transform the sector, but "how quickly" the transformation will unfold. And based on my analysis of liquidity flows, the answer is: faster than most companies can adapt. The industry is entering its own death spiral—not from a lack of collateral, but from a surplus of efficiency. That's the liquidity paradox of AI. It creates value while destroying the systems that produced it.