Key Takeaways

  • HCLTech will invest 142.57 billion rupees ($1.48 billion) in its first AI data center, developed in Bhubaneswar with Sarvam AI and support from the Odisha government.
  • The project combines computing infrastructure, GPUs, foundation models, and sector-specific applications in a full-stack AI offering.
  • A separate 5,000-seat HCLTech technology center in Bhubaneswar is expected to begin operations by 2028.

HCLTech is making a sizeable move beyond its traditional IT services base, committing 142.57 billion rupees ($1.48 billion) to an AI data center in Odisha. The Bhubaneswar project, announced Friday, will be HCLTech's first dedicated AI data center and will be developed in partnership with homegrown foundation-model startup Sarvam AI.

Financial support from the Odisha government will be included in the investment, although HCLTech did not disclose its structure or the planned computing capacity of the facility. The companies also did not provide a construction timetable for the data center.

Still, the strategic direction is clear. HCLTech wants to control more of the AI delivery chain rather than limiting its role to consulting, systems integration, or application management. That means bringing infrastructure, accelerators, models, and enterprise applications into one commercial offering.

HCLTech CEO C Vijayakumar described that approach during a post-earnings press conference last week. "It's the data centre, it's the GPUs, it's the models, it's the applications that we will deliver on top of it," he said. "The overall value creation is significantly of a very different magnitude when you really look at this as a full stack, and that's really what we want to play."

Sarvam AI is central to the plan. HCLTech acquired a 10.5% stake in Sarvam AI for $150 million last month, giving the two sides a financial connection in addition to their infrastructure partnership. The Odisha facility will combine HCLTech's full-stack AI capabilities with Sarvam AI's foundation models to deliver sector-specific applications for government-owned and private enterprises.

Owning or operating AI infrastructure changes the economics of an IT services relationship. Instead of earning revenue primarily from implementation and labor-based contracts, HCLTech can potentially participate in hosting, model deployment, inference, application operations, and ongoing optimization. It also places greater capital, energy, and utilization risk on HCLTech.

The investment comes as enterprise AI spending is moving toward infrastructure-intensive production workloads. McKinsey estimated in 2023 that generative AI could contribute $2.6 trillion to $4.4 trillion in annual economic value globally. Gartner estimated in 2024 that annual spending on AI-centric data center infrastructure and related IT services would exceed $200 billion by 2027, while IDC projected that India's AI spending would approach $5 billion by the same year.

India offers a compelling location for that buildout because of its large internet population, expanding digital economy, and comparatively lower operating and power costs. NASSCOM reported in 2023 that Indian data center capacity was growing at more than a 15% compound annual rate, with hyperscale and AI-oriented facilities becoming a larger part of the country's infrastructure roadmap.

Competition is not standing still. Tata Consultancy Services announced in November that it planned to invest $1 billion in AI data centers through a partnership with private equity firm TPG. Amazon Web Services and Microsoft Azure are also expanding AI-optimized infrastructure in India and other high-growth markets. Can IT services providers differentiate against hyperscalers while carrying the costs of specialized facilities? Their industry expertise and existing enterprise relationships may help, but utilization will matter.

Operational discipline will matter too. AI centers have demanding power, cooling, availability, and security requirements. ISO positions ISO/IEC 27001 as a framework for managing information-security risks, while the IEEE 3006 series addresses reliability considerations for industrial and commercial power systems. Those areas become especially important when infrastructure supports government workloads or regulated industries.

Alongside the data center, HCLTech plans to establish a 5,000-seat technology center in Bhubaneswar, with operations expected to begin by 2028. The paired investments could give HCLTech local engineering capacity close to the computing environment, supporting model customization, application development, and managed services.

For enterprise buyers, the real test will be what HCLTech and Sarvam AI deliver on top of the hardware. GPUs alone are increasingly available from multiple providers. Sector-specific models, governance controls, data residency, and measurable operating improvements are harder to replicate. Odisha now gives HCLTech a physical base from which to make that broader pitch.