According to Bernstein's Nikhil Nigania, artificial intelligence-driven data centers could significantly boost India's power demand growth from the current 5.5% compounded annual growth rate (CAGR) to 6% annually. As reported by CNBC TV18, Nigania explained that this represents a 0.5% bump up to the base case power demand growth, though he noted the shift would not change the sector on its own. The analyst compared this potential impact to the United States, where AI-driven demand has already reversed years of flat power use. Recent developments show similar trends globally, with Wood Mackenzie reporting that China's data center electricity consumption is projected to quadruple to 774 terawatt-hours (TWh) by 2030 as rapid growth in artificial intelligence workloads drives structural increases in computing-related power demand.
A significant development in India's power sector occurred with a renewable energy tender that came in below the cost of a new thermal plant, marking a first for the sector. According to Bernstein's analysis, the winning bid mimicked a thermal plant's round-the-clock supply profile at ₹5.25 per unit, below the ₹5.5-6 range that new thermal capacity commands. Nigania called this a big event after discussions with market participants in Delhi last week, highlighting the competitive advantage of renewable energy in India's power market. This trend aligns with global patterns, as Wood Mackenzie notes that data centres could account for 6% of China's total electricity consumption by 2030, with their share potentially rising to 17% by 2060. The surge in global electricity consumption reached a record high, climbing by nearly 1,100 terawatt-hours (TWh) in 2024, more than double the prior-decade annual average, according to the IEA.
India holds three key advantages that could attract data center investment, as noted by Bernstein's analysis. The country benefits from having some of the cheapest solar power in the world, thermal power costs that remain in line with other markets, and new high-voltage transmission lines that take approximately three years to build compared with eight to ten years in the United States. According to the report, India already ranks among the largest markets for companies such as OpenAI, behind only the United States. These advantages position India well for the growing AI infrastructure demand, with Wood Mackenzie noting that AI training is driving demand for more energy-intensive infrastructure, while the rapid expansion of AI inference adds sustained electricity demand. The country's projected power generation mix for FY 2025 targets 255 Billion Units from renewables, per the Central Electricity Authority, creating high-volume AC-DC and DC-DC converter deployments that directly drive SMPS transformer procurement.
The pricing gap between solar and non-solar hours remains substantial, with spot power falling to ₹1-2 per unit when the sun is out and climbing to the government-set ceiling of ₹10 per unit once the sun sets. As reported by Bernstein, demand for non-solar hours will continue to rise, benefiting storage technologies. Nigania emphasized that "no one wants just solar. People want non-solar supply more than anything," highlighting the growing market demand for diverse power sources. This trend is reflected globally, with Wood Mackenzie observing that data centres have historically clustered around major cities and technology hubs to access customers, talent and network infrastructure, but the growth of AI is increasingly encouraging a "compute follows power" model that could attract power-intensive AI training to renewable-rich regions. The global SMPS transformer market is valued at US$ 5.5 Billion in 2026 and projected to reach US$ 8.4 billion by 2033, growing at a CAGR of 6.2% between 2026 and 2033.
State-run lenders PFC and REC are experiencing significant challenges in their loan portfolios. According to Bernstein's analysis, loan book growth has slowed to mid-single digits from the teens seen earlier, with PFC's loan book actually shrinking quarter-on-quarter between March and June. The report notes that banks are now lending to renewable projects at rates near 8%, close to home loan levels, taking business away from these traditional power sector lenders. Both companies also carry foreign exchange exposure hedged through options rather than swaps, adding to earnings pressure. This shift reflects broader market dynamics where battery storage and intelligent workload scheduling could allow non-urgent computing tasks to be shifted to periods of abundant renewable power or lower electricity prices.
