Compounding · Investor Behaviour · Discipline
The Decade Clock Behind India's Research Push
India's innovation push runs on a clock no reporting calendar can read, and the evidence that counts is disbursal, not announcement.
A research budget is spent long before anyone can tell whether it worked. Money approved this year buys equipment next year, operated by people hired the year after. The first sellable product arrives late in the decade, if it arrives. That sequence fits no reporting calendar, which is why national research spending is easy for markets to ignore.
India spent 0.65% of GDP on research and development in 2020 and has raised that to 0.84% by FY24. The stated ambition is at least 2% within four to five years. Patent filings crossed one lakh for the first time in FY25, growing near 20% a year, with domestic filers now around 60% of the total.
| Country | R&D spending as a share of GDP |
|---|---|
| India | 0.65% |
| China | 2.36% |
| United States | 3.42% |
| South Korea | 4.52% |
| Israel | 5.83% |
The comparison sets the size of the task. Closing a gap that wide takes sustained spending from government and private companies alike, held steady across several election cycles and several profit cycles.
Buying down the risk premium
The main instrument is a Rs 1 lakh crore fund lending over six years, unsecured and collateral free, at 3 to 4 percent interest, repayable over 12 to 15 years, covering up to half a project's cost. It is aimed at artificial intelligence, quantum, defence, clean energy and semiconductors, and a young company can take equity financing instead of debt.
The pricing is the design. A private lender looking at deep research charges heavily for the chance that the science simply fails, and that premium is why domestic capital has largely stayed out of long-horizon technical work. Lending below it moves the risk onto a balance sheet willing to hold it.
Cheap money for research transfers risk rather than cash. Someone has agreed to be wrong for a decade so that someone else can try.
Announcements are cheap, disbursals are not
Intent gets announced often. The useful evidence is slower and duller. This fund was approved in July 2025, launched that November, and sent its first loans to private companies in May 2026. The semiconductor programme built on similar lines has three plants in commercial production. Under an earlier manufacturing incentive scheme, the smartphone was absent from India's top 100 exported commodities in 2014 and is now the largest single one, with 99.2 percent of phones used in India made here.
That shift took roughly a decade from policy to result, and it happened in assembly, the easier half of the problem. Original research runs slower. Anyone using the phone example as a template should carry the timeline along with it.
What is worth watching
Three markers show whether intent is turning into capability. Whether lending keeps pace with its six-year schedule rather than bunching at the end. Whether private research spending rises, since government money alone cannot reach the 2 percent target. Whether patent filings keep growing with domestic filers holding their share. India has more than 130 unicorns and over 2.4 lakh recognised startups, so demand exists.
None of these reads well month to month, and that is the honest difficulty. A speech is absorbed in a morning. The capability it describes takes the better part of a working life to show up, and the only edge available is a willingness to measure it on that clock.