|
India's biggest opportunity from the global artificial intelligence (AI) investment boom may lie not in building frontier AI models but in meeting the massive power and infrastructure needs of data centres, according to a thematic report by Shriram Mutual Fund, which argues that the AI spending cycle is "first and foremost a power-demand story" for India.
In its report titled "The AI Bubble Debate: A Unit-Economics Lens," the fund house said India has limited direct exposure to the global race among AI model developers and hyperscalers, but stands to benefit through companies linked to electricity generation, transmission and power infrastructure. "For India, the AI capex cycle is first and foremost a power-demand story," the report said, adding that "the binding constraint on the entire build-out is not GPU but electricity. Whatever the resolution of the unit-economics debate, whichever lab wins, however fast tokens deflate, electricity will get consumed." The report said this forms the basis of Shriram Mutual Fund's continued overweight position on the power sector, with exposure across generation, transmission, transmission EPC, power financiers, diesel genset manufacturers, and the wires and cables ecosystem. According to the report, India has "no listed frontier lab and no hyperscaler," meaning its direct exposure to the AI race is limited largely to the impact on Indian IT services demand. Instead, it said investors should focus on the physical infrastructure required to support the AI build-out globally. The report added that even within capital goods, it prefers companies linked to AI infrastructure rather than the broader sector. It said the fund is selectively positive on "the data-centre power and electrical-infrastructure chain (switchgear, transformers, grid, cooling and cabling)," describing them as "the 'picks and shovels' of the physical build-out" that benefit from gigawatt-scale AI infrastructure irrespective of which AI company eventually dominates. While the report is framed around the debate over whether AI spending represents a bubble, it argues that the central issue is not whether companies are spending too much, but whether those investments generate adequate returns before the underlying infrastructure becomes obsolete. "The entire 'bubble or not' debate collapses into one question of unit economics. Does the revenue thrown off by the compute earn an adequate return on the capital sunk into building it, before that capital depreciates? Everything else is just headline numbers," the report said. The report also said the current AI investment cycle differs from the dot-com and telecom booms because the spending is largely being funded by some of the world's most profitable technology companies. "On the evidence today, it is not a solvency bubble since the spend is overwhelmingly self-funded by the most cash-generative companies on earth... It is instead a return-on-capital question, and that question is still genuinely open," the report said. The report concluded that regardless of whether AI demand ultimately exceeds expectations or disappoints, India can still benefit by supplying the physical infrastructure required for the global expansion of AI. (ANI)
|
