India will need to fix its energy backbone before it can scale data centres and fully participate in the AI economy, said Ray Wang, principal analyst and founder at Constellation Research, adding that compute capacity ultimately rests on reliable power.
“The number one thing India has to do is shore up its energy infrastructure,” Wang said in an interaction. “Once that’s in place, you can install data centres, then LLMs will come. You can’t play this game without data centres as power, chips and data go together.”
Wang framed the point in the context of rising enterprise AI adoption, where demand for compute is accelerating faster than physical infrastructure can keep up. In his view, sequencing matters: without stable and scalable power, investments in data centres risk running into utilisation and cost constraints.
On enterprise software, Wang pushed back against the view that AI will render traditional vendors obsolete. “The SaaSpocalypse narrative is a mess,” he said, adding that valuation models have shortened terminal assumptions incorrectly. Core software vendors, he added, retain structural advantages. “They have four moats – data from systems of record, hard-to-replicate distribution, deep customer relationships, and efficiency and energy costs. If you win on these, you’re not going away.”
For IT services firms, Wang said the buyer is shifting up the organisation. “The CIO is the wrong buyer. Transformation spans business strategy to deployment, so the conversation has to be at the CEO level,” he said. As deals become outcome-led, contract sizes could expand alongside risk-sharing mechanisms. “A $1 million deal can become $10 million if a client buys an insurance layer to guarantee outcomes,” he said, adding that mid-tier firms are competing effectively despite assumptions that only large vendors can underwrite such risk.
He expects market structure to evolve with AI. “We’ll see AI-native players at the small end and large incumbents that get leaner. It will be hard to sustain a mid-sized layer,” he said.
From Pyramids to Diamonds
On jobs, Wang said the impact will be uneven but not necessarily negative for countries such as India. “It’s not simply about not hiring freshers and firing seniors. Firms will pay more for people who can use new technology,” he said. The middle layer will change as managers begin overseeing both people and AI systems. “This is the last generation of managers who only manage people,” he said, adding that workforce structures could shift from pyramids to more diamond-shaped organisations.
Concerns that AI agents will erode labour arbitrage advantages may be overstated, he said, pointing to a shortage of skilled AI talent. “The more AI we use, the more the human becomes critical because decisions still need humans,” he said, citing data indicating continued demand for software engineers.
For India, he reiterated that infrastructure sequencing will determine how quickly it can capture AI-led growth. “Fix power first, then scale data centres; the rest will follow.”