$11bn of state money, matched
India is preparing to deploy about $25 billion into deep-tech start-ups, CNBC reported on Monday, as it tries to close the distance on the United States and China and reduce its reliance on foreign frontier technology.
The structure matters more than the headline figure. The government has committed $11 billion under its Research Development Infrastructure Fund. That money is to be matched by venture capital and private equity managers, with a further three or four billion added on top, producing roughly $25 billion available to deploy.
Deep tech, in the Indian government’s usage, is broad: artificial intelligence, semiconductors, advanced manufacturing, drones and space technology.
Against a base of about $3bn
The comparison that gives the number its weight is the starting point. India’s domestic deep-tech funding totalled nearly $3 billion in 2025.
A programme of $25 billion against a $3 billion annual base is not an increment; it is an attempt to create a funding market that does not currently exist at that scale. Whether private managers put up the matching half is the load-bearing assumption, and nothing in the reporting says they have committed.
Why now
Two pressures are named. The first is supply-chain exposure: Indian officials have watched US export controls reshape who can buy which chips, and the deep-tech fund is explicitly about reducing dependence on frontier technology sourced abroad.
The second is capital flow. Officials say US tariffs are steering more investment toward India, which makes a matching-fund structure easier to fill than it would have been two years ago.
What to watch
Three things will show whether this is a fund or an announcement. Whether the private half materialises, and from which managers. Whether the money goes to semiconductors — the most capital-hungry and least forgiving of the listed categories — or spreads thinly across all five. And whether any of it reaches model training, which India has so far largely left to imported systems.
The government has not published a deployment timetable.
It is also worth being precise about what $25 billion buys at current prices. Goldman Sachs expects five American companies alone to spend $1.2 trillion on AI infrastructure next year. India’s entire deep-tech programme, spread across five sectors and an unstated number of years, is about two per cent of that single annual figure. The fund is not an attempt to match the build-out; it is an attempt to own some of the technology that runs on one.
