HCLTech's AI Infrastructure Play Can Redefine the Limits of Indian IT

For a company that spent decades selling engineering hours, HCLTech's June-quarter earnings commentary wasn't dominated by revenue, margins, or deal wins. Instead, executives preferred discussing AI data centres, sovereign compute, token costs, and AI infrastructure. 

Despite reporting record deal bookings of $2.4 billion and strong growth in its advanced AI business, the IT services giant believes compute infrastructure will be critical to delivering AI capabilities. 

HCLTech’s ₹3,500-crore investment in AI-ready data centres and sovereign AI ambitions reflect that wider shift in how India's IT services firms expect to compete in the AI era.

The country's third-largest IT services company reported revenue of ₹34,579 crore ($4.14 billion), up 13.9% year-on-year, and an EBIT margin of 16.9%, up 39 basis points quarter-on-quarter. Net profit rose 20.3% year-on-year to ₹4,626 crore ($554 million). 

HCLTech maintained its FY27 revenue growth guidance of 1–4% in constant currency.

Its Advanced AI business, a category the company has carved out and now reports separately, brought in $171 million in the quarter, up 10.6% sequentially and 62.1% year-on-year, taking its annualised run-rate to $688 million. While it is still a fraction of the company’s overall revenue stream, the management is optimistic that this sliver will decide who wins the next decade of enterprise technology spending.

AI Infrastructure Play

HCLTech’s strategic investment plan of up to ₹3,500 crore ($420 million) to build an AI-focused data centre business, with the potential to scale to 50 megawatts of capacity, is seen as a departure from the company’s asset-light model.

“The company’s continued investments in AI factories, physical AI, AI Force, engineering-led services, and strategic ecosystem partnerships are helping it transition from a traditional IT services provider to a higher-value AI-led transformation partner,” says Biswajit Maity, Senior Principal Analyst, Gartner.

The move marks a reversal from October last year, when CEO and MD C Vijayakumar had ruled out entering the data centre business. Since then, larger rival TCS has unveiled plans to build up to 1 GW of AI-ready sovereign data centre capacity in India, with investments exceeding $6.5 billion.

"We have not really departed from our core business model of remaining asset-light. We see this as a logical extension," Vijayakumar affirmed in the post-earnings call. 

He added that AI has made compute and data centre capacity very scarce, which has also become the strategic bottleneck in the entire value chain. So, owning a foothold in data centres lets HCLTech sell a higher-margin, full stack of services layered on top. The company is close to signing its first anchor client for the facility and has framed the investment as a hybrid of debt and equity, with other partners likely to be announced later.

That sovereignty pitch was reinforced by HCLTech’s $151 million investment in Sarvam AI. Vijayakumar said the logic was to pair Sarvam's model and research capability with the IT major’s global enterprise relationships, engineering expertise, and software IP to build an AI stack spanning models, platforms, applications, and managed services.

No Longer Someone Else's Problem

Compute has become a cost line item that Indian IT companies no longer want to be exposed to without a hedge. Vijayakumar said that owning capacity is partly about protecting margins on the managed-services and fixed-price contracts HCLTech runs for large clients.

"Token costs are a very important component of the overall delivery," he said, adding that owning capacity and models gives the company an edge in pricing managed services and fixed-price contracts for global clients.

That is a notable admission from an industry that has traditionally treated compute as someone else's problem. HCLTech's bet is that as AI inference becomes a bigger share of delivery cost, controlling that layer becomes as important to margins as controlling headcount pyramids once was.

AI Deflation is Real

The other side of HCLTech's AI story is less comfortable. The same technology driving growth in its AI business is also putting pressure on its traditional services portfolio.

The company said it expects AI deflation of 2–3% on its overall portfolio this year, as clients also demand to benefit from AI-driven efficiencies.

On whether HCLTech absorbs any of that saving into its own margin rather than passing all of it to clients, the company declined to give details, calling it "very situational" and dependent on the nature of each relationship, leaving open exactly how much of the deflation is offset by HCLTech's own take.

The company wants to grow the AI-native and AI-amplified business faster than the deflation shrinks the legacy one.

Headcount Story

The workforce numbers told a similar story. HCLTech's headcount fell by 3,292 during the quarter to 223,889 even as revenue grew, while revenue per employee rose 3.3% year-on-year—a metric that has now improved for five consecutive quarters.

The company is also reshaping the kind of talent it wants. Chief People Officer Ramachandran Sundararajan said HCLTech is building a dedicated pipeline of Forward Deployed Engineers (FDEs), adopting a "build and buy" strategy that relies primarily on internal AI skilling alongside selective lateral hiring for immediate needs.

AI is forcing Indian IT to rethink its decades-old playbook. The competitive advantage is no longer defined solely by the size of the workforce or the offshore delivery engine. Increasingly, it will hinge on who owns the compute, builds the platforms, develops AI-native talent, and captures a larger share of the AI stack.