By early 2026, India was the world’s third-largest producer of peer-reviewed scientific publications, behind only China and the United States. It ranked 38th in the Global Innovation Index, first among lower-middle-income economies. Patent filings rose from 58,503 in 2020-21 to 110,375 in 2024-25 [5]. On paper, this is a country arriving at the top table of global science.
Yet of 230,480 patents in force in April 2025, only 1.59 percent had been commercialized, licensed, manufactured, or sold. Counting the more generous Form 27 filings, the figure is 6.39 percent. India also runs a net outflow on intellectual property charges of roughly $14.5 billion a year. That number needs careful reading, because part of it reflects intra-group payments by multinational subsidiaries rather than purchases of technology Indian labs could have supplied. But the direction is clear: India pays far more for IP than it earns.
“Beyond publications” is not a call to publish less. It is a call for a second scorecard. Publication and filing counts measure activity inside the laboratory. A commercialization scorecard (licenses signed, royalties earned, products shipped, companies spun out) measures whether that activity ever left the building. India has spent three decades building the first. It has barely begun the second.
The Funnel That Leaks
A rupee of public research money passes four gates: publication, patent filing, patent grant, and commercialization. India clears the first better than almost anyone and the last worse than almost anyone.
Output has surged. Publications doubled between 2015 and 2024, overtaking the UK. India produced four percent of a record two million global papers, with 7.6 percent growth. The share of Indian papers co-authored abroad rose from 23 to 36 percent, and citation impact climbed from 0.7 to 0.9 of the world average.
Filing grew faster, at 17.2 percent annualized. But about 33,504 patents were granted against 110,375 applications in 2024-25. That is roughly 30 percent, a rough proxy given examination lags. Only 10,682 grants went to Indian applicants; foreign filers, mostly via the PCT route, took the majority. Part of the boom is arithmetic. Lovely Professional University filed 7,096 applications between 2020 and 2023, against 2,333 filed by all the IITs combined between 2020 and 2025. Grant rates at such high-volume filers run at just 1 to 3 percent.
Beneath this sits a spending gap. India’s gross R&D expenditure is about 0.64 percent of GDP, against a global average near 2.67 percent. The US spends 3.59 percent, China 2.56, and South Korea 5.21. Nearly three-quarters of India’s spending is public, whereas in most advanced economies industry funds the majority. A commercialization pipeline needs money entering at the industry end, from firms that intend to build products.
The Incentive Trap
Ask why, and the answer is that nobody’s career depends on it being different. Promotion in public research runs on publication and filing counts, not licenses or products. A line often repeated inside CSIR says a scientist is promoted for publishing a paper no one reads, filing a patent no one licenses and attending a conference no one remembers.
Rankings compound the problem. Under India’s institutional ranking frameworks, a university that files more applications scores higher than one holding the same number of grants, because filings rather than outcomes are counted. Meanwhile, chasing a licensee, negotiating royalties, and shepherding a prototype takes months and appears in no appraisal form. Rational scientists choose the paper and the filing. The system is not failing by accident. It is delivering what it rewards.
CSIR: Patents Without Profit
CSIR, founded in 1942, runs 38 laboratories and 39 field stations on a budget of roughly ₹6,300–7,144 crore, about 90 percent of it guaranteed whether or not any technology is licensed. In the mid-2010s it filed more than 13,000 patents, about 4,500 domestic and 8,800 abroad, at a cost of roughly ₹50 crore over three years. CSIR-Tech, its commercialization arm, ran out of money maintaining the portfolio and shut down. Then Director-General Girish Sahni later conceded that many were “bio-data patents” filed to pad scientists’ resumes.
There has been improvement. In 2017-18, licensing revenue rose about 32 percent, from ₹727.3 crore to ₹963 crore, of which ₹515 crore came from private-sector licensing. Individual laboratories, such as the Indian Institute of Chemical Technology, market their portfolios actively. But an institution with a budget above ₹6,000 crore earning ₹963 crore in a strong year is not a self-sustaining engine. Success is concentrated in a few labs with the right culture, while all 38 are funded on the same guaranteed basis. A model that pays out win or lose will keep producing scattered wins and many losses.
The Campus Gap
IIT Delhi, Madras, Bombay, and IISc run technology transfer offices offering incubation, legal support, and investor linkage. The National Science and Technology Entrepreneurship Development Board has funded Technology Business Incubators on such campuses since 2009. IIT Kanpur shows what a good office can do. It has filed over 1,000 IP applications, secured more than 400 granted patents, and achieved a technology-transfer rate of 13.76 percent, several times the national rate. Its Bhu Parikshak soil-testing device, licensed to a private aggrotech firm, reached more than a million farmers in its first year.
The problem is scale. A 2010 Prime Minister’s Office task force recommended roughly 100 incubators nationwide. Fifteen years on, that has been only partially acted upon, and the same handful of institutions still do almost all the commercialising. A programme meant to spread capacity instead concentrated it.
Money Without a Mandate: The ANRF Test
The Anusandhan National Research Foundation Act of 2023 created India’s answer to the US National Science Foundation, with a ₹50,000-crore corpus over five years: ₹14,000 crore from government and a targeted ₹36,000 crore from private and philanthropic sources. In November 2025 came a further ₹1 lakh crore Research, Development and Innovation Fund for deep-tech translation, to be operationalized through ANRF. This is exactly the patent capital India has lacked.
Then came the audit. A March 2026 parliamentary standing committee found that ANRF spent none of its ₹2,000-crore allocation in 2023-24 or 2024-25 and had used just 61 percent of its current-year allocation by January 2026. The Department of Science and Technology called the period a “preparatory phase” of building “institutional, administrative, and financial architecture.” Standing up a statutory body takes time, which is fair. Sadly, it is also the explanation that has justified thirty years of slow and laborious science bureaucracy. The RDI Fund will be judged by how many rupees reach a laboratory, pilot line, or spin-out within its window.
ANRF’s PAIR programme (Partnerships for Accelerated Innovation and Research) points to the right way. It pairs a hub institution with spoke institutions, allocating up to ₹100 crore per network, about 30 percent to the hub and 70 percent to the spokes, to spread capacity beyond elite campuses. Whether “research capacity” is defined to include commercialization will decide whether it succeeds.
What Other Systems Do Differently
The US Bayh-Dole Act of 1980 gave universities ownership of federally funded inventions on condition that they commercialize them. In fiscal 2024, American institutions logged 26,196 invention disclosures, filed 14,432 patent applications, executed 9,507 licenses and options, launched 775 products, and formed 951 startups, on $109.7 billion of research spending. Licensing income runs to roughly $3.8 billion a year, and nearly 7,000 university-born startups are operating. A few blockbusters, like Northwestern’s $1.4 billion Lyrica deal, skew the averages, but the base rate of activity dwarfs India’s.
China took a different route. Its 2015 Law on Promoting the Transformation of Scientific and Technological Achievements lets universities make independent technology-transfer decisions and guarantees inventors at least 50 percent of the profit as royalty or equity. Universities also run wholly owned ‘University Run Enterprises’ that capture technology income directly.
The common thread is legal certainty about who profits and by how much. India’s 2016 National IPR Policy encouraged technology transfer offices but neither mandated them nor fixed a profit-sharing formula. Ownership and inventor share still vary by institution. A statutory approach has been attempted before, in the Protection and Utilisation of Public Funded Intellectual Property Bill of 2008, which lapsed. Any new effort must focus on ‘why.’
The Private Capital Signal and Its Limits
Indian tech startups raised $9.1 billion in 2025, up 23 percent. Deep tech raised $2.3 billion, up 37 percent, and India hosts more than 4,200 deep-tech startups, over 550 of them founded in 2025. Deep tech’s share of all VC-PE funding rose from 4 percent in 2016 to 15 percent in 2025, and the RDI Fund has reportedly catalysed roughly $2 billion in fresh commitments.
But 84 percent of deep-tech startups, and 91 percent of their funding, are AI-related. That is largely software built on globally available models, not hardware or materials emerging from Indian laboratories. Funding for semiconductor and hardware deep-tech startups since 2021 totals only $5.5 billion, “big promises, small cheques”. It is also volatile: startup funding in the first half of 2025 fell 25 percent year-on-year. Private capital is betting on India’s future output. It is not yet monetizing the patents already filed.
The Defence Mirror: iDEX and DRDO
Inside the defence establishment, India is running a smaller experiment with more encouraging results. Innovations for Defence Excellence (iDEX) has a five-year budget of just ₹498.78 crore, approved for 2021-26. It funds startups, MSMEs, and individual innovators through roughly 20 partner incubators and is designed around production-ready prototypes rather than publications. Sixty-six contracts were signed with winners in its early phase. It has since scaled to over 2,000 defence startups and 300 space-sector startups.
DRDO’s Technology Development Fund raised its per-project ceiling from ₹10 crore to ₹50 crore in 2022. DRDO has also made its patents available to industry at nil cost and opened its test facilities to outside firms. Barriers to scaling from prototype to procurement remain. Analysts compare the trajectory to Israel’s Unit 8200 in the 1990s. That is aspirational, not settled fact, but the model is the one the civilian system lacks.
A Six-Point Agenda
Every fix below already exists somewhere in the world or in miniature within India. The task is replication at system scale.
- Rewrite the incentives. Tie promotions, grant renewals, and rankings to licensing revenue and deployed technologies. Cap the weight of filing volume in frameworks like NIRF and score commercialization outcomes.
- Codify IP-sharing by statute. Guarantee inventors and institutions a fixed share of revenue, with China’s 50 percent floor as a reference point, and study why the 2008 bill failed.
- Professionalize technology transfer offices. Fund fewer, better-staffed regional offices serving clusters of institutions, instead of token IP cells at every university.
- Make ANRF spend. Introduce disbursal fast-tracks, sunset clauses on unused allocations, and a public quarterly utilization dashboard so slow spending is visible in real time rather than in a committee report eighteen months later.
- Fund the valley of death. Most grants stop at prototype. Dedicated translational grants for pilot production, certification, and first commercial units close the gap where most inventions die.
- Copy what works next door. iDEX mandates a product outcome up front, removes the cost of accessing institutional IP, and measures success in contracts signed. CSIR, the IITs, and ANRF-funded research can adopt the same logic, tying procurement offtake to laboratory outputs.
From Prestige to Profit
India has assembled most of the ingredients: a publication engine at world-third scale, a fast-growing patent apparatus, a ₹50,000-crore research foundation, a ₹1 lakh crore deep-tech fund, and a private capital market that has begun to notice deep tech. What is missing is the plumbing that connects them: incentives that reward outcomes, a legal framework that settles who profits, and institutions willing to spend the money they already have.
Until then, products will pile up in repositories, patents will collect renewal fees instead of paying dividends, and India will keep paying for technologies its own laboratories may already have invented and forgotten. iDEX has shown, on a fraction of ANRF’s budget, that funding with a product mandate produces different behaviour. The cheapest and fastest fix is not a new ministry or a new plan, but a decision to measure every rupee of research spending by what it sells, not by what it publishes.
For decades, we have measured our research by how many papers it produces. Our laboratories are full of publications; our markets are still full of imported products. It is time to change that and to build a nation that does not just publish papers but produces products from meaningful research and patents.