India’s largest IT firm just put hard numbers behind its AI ambition. The company will incorporate a new entity focused on AI infrastructure and services and it plans a gigawatt-class AI data center in India. On the same day, it printed a steady Q2 and reiterated cash returns, signaling that the balance sheet can fund the build while servicing shareholders.
Why 1 GW is a statement
Training-class clusters are power hungry. A gigawatt target places the project in the tier of hyperscale builds designed for frontier models, large inference farms, and sovereign workloads. The move positions the company to sell not only AI consulting and apps but also managed compute for clients that cannot or will not put sensitive workloads on public clouds.
The operating model
A wholly owned subsidiary gives flexibility on partnerships, power procurement, and real estate. Expect multiple sites over time rather than a single campus, with mix-and-match accelerators and liquid cooling in high-density rooms. The company also flagged sovereign data center capabilities, which matters for public-sector and regulated industry deals that require data residency and auditability.
Why do this now
Customer demand has shifted from pilots to programs. Large clients want predictable capacity, sustained cost per token, and SLAs for safety tooling. Owning part of the compute stack can compress deal cycles and protect margin. It also de-risks supply chain shocks in accelerators and power.
Execution risks
Power availability and cost, chip supply allocation, and talent for HPC fabrics challenge every builder. The advantage here is scale, cash flow, and an existing bench that runs complex private clouds at global clients. The board also approved a tuck-in acquisition to deepen marketing automation and Salesforce capabilities, suggesting a full-stack go-to-market where AI sits beside CRM and data pipelines.
Investor read
The services engine remains the cash cow. The AI unit and data-center plan are optionality with teeth. Track capex cadence, PPA disclosures, and first anchor customers. On the P&L, watch whether the new unit drags margins before it scales or lands as accretive via bundled deals.
Follow Tech Moves on Instagram and Facebook for deal trackers, capex breadcrumbs, and a simple glossary for 1 GW builds that your CFO will actually read.
