India's AI Infrastructure Faces Demand Gap

A new report by the Rhodium Group assesses India’s AI landscape, focusing on data centre development, sovereign strategy, and IT sector capabilities. The report finds India’s strengths lie in areas where its public strategy remains quiet.

India has attracted over $300 billion in data centre investment commitments in the past year, with operational capacity projected to triple by 2030. However, demand for these facilities remains underwhelming, with GPU utilization rates at around 20 percent.

India’s largest GPU-as-a-service provider, Yotta, generates 80 percent of its revenue from foreign clients. Domestic enterprises continue to use global cloud providers rather than Indian servers, indicating limited local demand for AI services.

State governments justified incentives for data centres with employment promises, but the Rhodium report highlights a gap between these claims and actual job creation. For example, Andhra Pradesh’s IT Minister cited 200,000 jobs from a single project, while the state’s own agency estimates only 200 direct positions.

India’s alignment with US tech policies is evident in its semiconductor export controls, telecom regulations excluding Chinese equipment, and positions in international spectrum allocation forums. These choices signal a strategic shift despite New Delhi’s diplomatic stance of neutrality.

China has responded to India’s alignment by recalling engineers from Foxconn plants, restricting tunnel boring machine exports, and applying rare earth controls. The Rhodium report suggests Beijing is unlikely to accept India as a passive alternative to Chinese manufacturing.

India’s AI labs, supported by the IndiaAI Mission, have focused on large language models, with Sarvam AI’s 105B parameter model as a notable example. However, the domestic market for Indic-specific AI products is too small to fund frontier model development, allowing foreign labs to close the performance gap.

The Rhodium report emphasizes the potential of India’s IT services sector, noting that firms like TCS and Infosys have seen double-digit growth in AI services revenue. The forward-deployed engineer model, where professionals work directly with clients, offers a competitive edge due to lower costs and deep industry knowledge.