HCL Tech’s Rs 3,500 crore AI data centre foray: A new growth engine or capital-intensive diversion?

Noida-based IT services major HCL Technologies is putting as much as Rs 3,500 crore behind an ambitious push into AI data centres, betting that owning everything from infrastructure and compute to models and managed services will produce more value than merely renting space and power.

But the company is also putting guardrails around one of its most capital-intensive strategic moves. The initial investment will create only a fraction of its proposed 50-megawatt capacity, while subsequent spending will depend on the scale of the business, cash flow generation and returns.

That calibrated approach will be central to how investors assess the foray: as a new growth engine that strengthens HCL Tech’s position in the AI value chain, or as a capital allocation experiment whose economics are yet to be established.


“The biggest opportunity is not to rent AI, but to own the full stack,” Chief Executive Officer and Managing Director C. Vijayakumar said during an earnings conference call. “As demand grows and supply catches up, the players who win will be the ones who will offer full-stack offerings.”

HCL Tech plans to combine data centre infrastructure, compute, models, software and managed services to offer sovereign cloud, secure AI and managed AI infrastructure. The company believes such an integrated model can generate more enterprise value per megawatt than a conventional colocation business, where operators primarily monetise physical space and power.

The opportunity rests on a sharp expansion in demand. Vijayakumar expects global data centre demand to triple by 2030, with AI accounting for roughly 70% of that growth. India, he said, could expand even faster because it remains one of the world’s most supply-constrained data centre markets and faces rising sovereign-data requirements.

“The convergence of AI-led demand, supply constraints, and sovereignty needs represent a very compelling opportunity,” Vijayakumar said. “This is a business which is shifting from physical infrastructure to higher value AI-ready solutions. One we believe will be a new growth vector for HCLTech.”

Also Read | HCLTech FY guidance stays muted despite $2.4 billion deal momentum

HCL’s investment in data centres

HCLTech’s long-term plan envisages scaling capacity to 50 MW, but management cautioned investors against treating the Rs 3,500 crore commitment as the cost of building the entire capacity.

“Our 50 MW is a long-term plan,” Vijayakumar said. “The Rs 3,500 crore is actually representing only a fraction of that capacity, and that is the initial investment that should get us started.”

The company has yet to specify how much capacity the initial investment will create. That leaves the eventual capital requirement for the 50 MW build-out—and the returns it can generate—as key variables for investors.

Management, however, made clear that it is not committing to a large-scale investment upfront. Further capital deployment will be paced according to demand and the free cash flow generated by the business.

“We will have a very disciplined approach to increasing investments based on the free cash flow that is getting generated from the business,” Vijayakumar said.

The company is also seeking to reduce the risk of building capacity without corresponding demand. It is in advanced discussions with clients and wants to begin operations with some committed consumption from the first day. HCL expects to use part of the capacity internally for managed-services and outcome-based contracts with global clients.

Initial capacity could also be absorbed relatively quickly in a market constrained by the availability of graphics processing units, according to management.

The IT firm does not intend to fund the entire investment through its balance sheet or deploy the full amount in a single year.

Potential financing structures include partnerships with silicon and original-equipment manufacturers, committed-capacity agreements with clients, and consumption models under which hardware is financed against contracted demand. The investment could also be funded through a combination of equity and debt.

“There are potential possibilities of funding this through a mix of partners, arrangements with silicon and OEM vendors, and committed capacity from clients,” Vijayakumar said. “We are not putting this entire investment in the balance sheet in a single year.”

Management also said the data centre investment would not come at the cost of HCLTech’s stated shareholder payout policy.

These safeguards of committed demand, partner financing and phased capital deployment could help contain the financial risk. Still, the long-term investment case will depend on whether HCLTech can monetise its broader technology capabilities rather than compete as a conventional infrastructure provider.

“This is not a colo business,” Vijayakumar said. “This is going to be a full stack play.”

Also Read | HCL Tech Q1 Results: Profit rises 20% YoY to Rs 4,624 crore; co declares Rs 12 dividend

The full-stack differentiator

HCL Tech’s strategy is centred on developing cost-effective, industry-specific small language models rather than competing directly in frontier or general-purpose models. The company wants to combine those models with its capabilities in AI data centre design, DevOps, cloud operations and software.

The thesis is that the firm can earn more from each megawatt by selling a complete enterprise AI solution instead of only providing the underlying infrastructure. If successful, the data centres could deepen client relationships and support the company’s managed-services and outcome-based contracts.

Motilal Oswal analyst Abhishek Pathak believes HCL is moving early to create the infrastructure and capabilities required for the next phase of AI adoption.

“We believe HCLT is investing ahead of the market to build the next-generation AI stack,” Pathak said.

He pointed to the company’s five-pillar strategy of transforming services, developing differentiated intellectual property, expanding AI-led services, strengthening partnerships and scaling AI talent.

“This should help the company defend against, and eventually benefit from, AI disruption, and HCLT seems ahead of the curve here,” Pathak said.

Motilal Oswal retained its buy rating on HCL Tech and raised its price target to Rs 1,450, valuing the stock at 18 times estimated FY28 earnings per share, compared with 16 times previously. It kept its earnings estimates unchanged and maintained HCL Tech as its preferred large-cap pick.

The unchanged estimates suggest that the brokerage’s increased optimism reflects HCL Tech’s stronger long-term AI positioning rather than any immediate earnings contribution from the data centre venture.

(Disclaimer: Recommendations, suggestions, views and opinions given by the experts are their own. These do not represent the views of The Economic Times)