# Beyond Literacy, a Research Republic

Author: Pramod Prasanth
Published: 2026-08-17
Canonical: https://pramod.ch/essays/beyond-literacy-a-research-republic
Description: Keralam produces more papers each year. That is a starting point, not a destination. From a series asking what Keralam 2040 would have to be.

---

On its own fellowship, the University of Kerala was still paying a doctoral researcher ₹13,000 a month in 2022, a rate set in 2016. One who wins a national fellowship is paid ₹37,000 a month in the first two years and ₹42,000 from the third, at a rate set centrally, in Delhi, that Keralam cannot change.[^9]

The average salaried employee in Keralam earned about ₹22,000 to ₹23,000 a month in 2023-24.[^9] A capable Keralam graduate holding a master's degree is therefore not choosing between research and unemployment. On the University of Kerala's own fellowship, they are choosing between research on little more than half an ordinary salary and a job in the same city, available immediately, with no thesis at the end of it. This is the same graduate the second essay in this series tracked at the tea stall: over-qualified for the local job market, under-rewarded for the research career that could feed the precision clusters this series has been describing.

Most take the job. Each of those decisions is sensible on its own, and the sum of them is a state that cannot staff a research base.

None of this appears in the number Keralam publishes about its research.

A few years ago I came across a news item reporting Keralam's research output statistics. Publication growth between 2018 and 2023 was strong, the headline said. The original study, an Elsevier report released through the Kerala State Higher Education Council, was not available online. It still is not. The question that did not get asked, in that news item or in any of the news items like it I have seen since, is what those publications do once they exist. Where they are cited. By whom. In what fields. And whether the institutions producing them are growing into research bases that compound, or producing more papers each year without building the compounding institutional base those papers should feed.

Keralam's publication numbers grew. This essay is about the distance between that and a research base whose work other serious institutions regard as essential reading.

This series has argued that Keralam's half-finished miracle has a talent-supply gap, a demand-side institutional gap, and a cluster-architecture gap. The research base is the fourth gap. Keralam has substantial talent and a broad institutional base; what it has not built is the layer that turns research into compounding capacity. Without that layer, the precision clusters this series has been describing have no research base to feed them.

### The applause trap, applied to Keralam

Publication growth makes the deeper question harder to ask. Nobody wants to be the one asking what a number counts while everybody else is congratulating the state on it going up.

Keralam's publication growth, in its current form, measures volume. It does not yet measure depth. Citation impact in priority domains. International researchers who choose Keralam as a sabbatical destination because the work is serious. Industrial firms that fund Keralam research because the output feeds their R&D pipelines. Doctoral graduates who stay because the work here can hold a career. None of these track with publication count, and three specific substitutions show up in Keralam's current research policy debates.

**Publication growth read as research depth.** Publications by Keralam's state universities grew 11% a year between 2018 and 2023 on Elsevier's count.[^1] A count of this kind can include mandatory course papers, derivative work, and conference proceedings. The volume is real. In that total, a required coursework paper counts for exactly as much as a paper other institutions treat as essential reading.

**MoU announcements read as industrial linkage.** A memorandum between a university and a firm is easy to sign and easy to announce. An MoU is not a research partnership. A research partnership involves shared problems, shared data, shared staff, and shared intellectual property. A count of MoUs measures signatures. It does not show how many reach the four-shared structure.

**Joint degree programmes read as internationalisation.** The UGC's regulations on twinning, joint and dual degrees, notified in May 2022, are precise about how coursework is divided. A dual-degree candidate must earn at least 30% of total credits from the Indian institution, a twinning student may take no more than 30% abroad, and in a dual degree both universities award a degree. The framework requires each agreement to cover intellectual property rights, but says nothing about how the research produced inside such an arrangement is attributed or whose affiliation appears on the resulting paper.[^2] Nothing in the rules stops the foreign partner from taking the affiliation and the citation for work done in India. For the individual candidate that is often a good deal. It does not internationalise Keralam's research base; it internationalises the foreign partner's, using Keralam's labour.

A second constraint sits inside the same rules. Only Indian institutions with a National Assessment and Accreditation Council (NAAC) accreditation score of at least 3.01, a place in the NIRF university top 100, or a place in the top 1,000 of the Times Higher Education or QS rankings may enter these partnerships at all. Restricting entry to the accredited and the ranked protects standards, and it shuts out the colleges below that line, which are the places that would gain most.

> An MoU is not a research partnership. It is the press release for one.

The landscape Keralam would be entering is itself moving. In 2024 China extended its lead over the United States on the Nature Index, which tracks contributions to 145 high-quality science journals. China's adjusted share rose 17.4% in a single year while America's fell 10.1%, ending the year at 32,122 against the United States' 22,083. Georgetown's Center for Security and Emerging Technology projected China would graduate more than 77,000 STEM doctorates in 2025 against America's 40,000.[^12] The bar for what counts as a globally legible research base is rising, not holding still. A Keralam research republic built for 2040 cannot be designed against the standards of a decade ago.

### What a serious research base looks like

Two countries built serious research bases at small national scale. Neither was a historical accident. Both were institutional choices held over decades.

Singapore's research engine is A*STAR (Agency for Science, Technology and Research), together with the National University of Singapore. A*STAR funds and runs research institutes in deliberately chosen domains rather than spreading capacity across every field a Singapore academic might pursue.[^3] NUS runs joint laboratories with industry firms: actual joint laboratories, with shared problems and embedded researchers, not memoranda. Doctoral and postdoctoral fellows are paid at rates that make staying globally competitive. The signal is that the state values research talent enough to pay for it. The institutions are globally legible: a researcher in Seoul, Boston, or Stockholm regards them as serious peer destinations.

Switzerland's research engine works differently. ETH Zurich and EPFL Lausanne are federal technical universities with a mandate to produce research that compounds into firms. ETH has spun out more than 580 companies since 1973, a record 43 of them in 2023 alone, which puts it among the strongest performers in Europe on that measure.[^4] Switzerland spends around 3.2% of GDP on research and development, among the highest rates in Europe, and about two-thirds of that research is carried out by private enterprises.[^5]

Basel's life sciences cluster, Roche and Novartis and the biotechs around them, sits beside that research base; Roche was founded in the city in 1896.[^5]

Both countries express the same structural argument in two institutional forms. Research becomes infrastructure for the productive economy when three conditions hold together. The work itself, where individual researchers do frontier research. The institutions that fund, teach, and house them. And the distribution layer, which turns the first two into compounding capacity: the international peer network, the industrial linkage, the multi-generational doctoral pipeline. Keralam has the work. Keralam has institutions that publish. The distribution layer is what Keralam has not yet built.

The precision-cluster candidates this series has described depend on exactly this missing layer. A medical-devices cluster needs serious tropical health and ageing research feeding it. A green-chemistry cluster needs serious bio-based materials work. A precision-marine cluster needs serious marine science. None of these compound from publication count alone.

A common reading of AI's impact says long-form research training is becoming irrelevant: adaptability and short, modular skills are what matter now. That reading is right for applied vocational work, where skill turnover is real and accelerating, and where the apprenticeship system this series has earlier argued for is the institutional response that fits. It is wrong for research depth. AI amplifies what trained researchers can do; it does not substitute for the institutional infrastructure that produces frontier insight. In doing so, it widens the advantage of institutions with genuine depth over those with only volume output, making the case for concentration in Keralam's research system stronger now than it was a decade ago. The AI-accelerated world widens the comparative advantage of places with patient research capacity. The question is whether Keralam chooses to be one of those places.

### Where Keralam could plausibly matter

Keralam does not need research capacity in every field. It needs concentrated excellence in a small number of institutions and a small number of domains. The political challenge is that every institution wants to be included. The fiscal reality is that not every institution can be. At Keralam's scale and budget envelope, concentration is the only strategy that produces globally legible results.

The obstacle is not a lack of awareness. The incentives point the other way: an institution that adds departments, subjects, memoranda and publications looks as if it is growing, whatever the depth of any of them. Reward concentration and the system optimises for something else.

> Better two world-class labs in tropical health than twenty mediocre departments in everything.

Four domains have a genuine basis for world-class Keralam research. They are the four where an existing institutional anchor, a natural endowment, and growing global research demand intersect cleanly enough to justify concentrated investment.

**Tropical health and ageing.** Keralam has one of India's most sophisticated public health systems and the oldest population of any Indian state: 14.4% were aged sixty or over in 2021, on the Sample Registration System's count. Neglected tropical diseases affect more than a billion people worldwide and have historically sat almost outside the global health policy agenda, until the Sustainable Development Goals took them up in 2015. The Sree Chitra Tirunal Institute for Medical Sciences and Technology in Thiruvananthapuram and the Amrita Institute of Medical Sciences in Kochi are existing anchors.[^6] The research feeds directly into a medical devices and diagnostics cluster of the kind this series has been describing.

**Climate and coastal systems.** Keralam's geography (Western Ghats, coastal plain, backwaters, monsoon intensity) is a natural laboratory for tropical climate adaptation research. The 2018 floods exposed gaps in what the state knew about its own flood risk: three years later, the Comptroller and Auditor General found that its flood plains had still not been demarcated and that no large-scale flood hazard map existed.[^15] Climate adaptation research for tropical coastal systems is a globally significant gap. The output translates into climate-risk tools, water-management systems, and sustainable land-use design.

**Marine science and blue economy.** A coastline of about 590 kilometres, a deep traditional fishing economy, and an existing marine biology base at the Central Marine Fisheries Research Institute (CMFRI) in Kochi describe a credible research domain.[^7] The output feeds sustainable aquaculture, marine biotech, and precision fishing, which is the precision marine cluster.

**Applied AI for public systems.** Keralam has one of India's most extensive public service-delivery networks: health, education, local governance. The Kudumbashree programme organises 43 lakh women into 2.77 lakh neighbourhood groups, yet in a 2018 audit the Comptroller and Auditor General found fewer than one per cent of its members registered on the web portal of one of its programmes, Sreesakthi.[^8] The network is real. The data it could produce mostly does not exist yet, and building it well is itself a research problem. Applied AI research that started there, at this scale and with this institutional reach, would be globally legible. Concrete examples are not hard to find: AI-assisted epidemic surveillance over Keralam's health network data, decision-support for local-government service delivery, applied-language tools for Malayalam public services. The specific industrial direction is a question for later in this series; the research depth that any of these would require is the same.

### The PhD economics

Set even the national rate against what the comparators pay. A*STAR's graduate scholarship in Singapore pays a doctoral scholar SGD 3,600 to SGD 4,500 a month on its standard science track, depending on citizenship, and more on specialised tracks in semiconductors, quantum and computing. Switzerland's national science funder sets a band of CHF 50,000 to CHF 55,000 a year.[^10] Converted at market rates those are six to fifteen times the Keralam figure, which is the number usually quoted and the least useful one. Rent in Zurich is not rent in Thiruvananthapuram. Adjusted for what the money actually buys, a Keralam doctoral researcher on the national rate receives 1,900 to 2,100 international dollars a month, against 3,500 to 4,400 on Singapore's standard track and 3,900 to 4,300 in Switzerland.

So the honest gap is about twice, not tenfold, and once Swiss prices are taken into account Singapore's standard track and Switzerland end up close together. Twice is still the difference between a career and a sacrifice, and the nominal gap is not irrelevant either: a researcher who intends to save, support a family at home, or come back with capital is paid in the currency that converts, and there the gap is closer to tenfold. Both numbers push the same way. Neither makes staying the obvious choice.

The pattern is national before it is local. Among the countries the 2026 Stanford AI Index charts, India had the largest net outflow of AI authors and inventors in 2025, a net flow of -16.9, more than double Canada's -7.1 and Germany's -2.4.[^13] Keralam's doctoral economics are a state-level instance of that.

> The researchers who can do frontier work are precisely the ones who leave. This is not a criticism of individuals. It is a statement about a system that has not yet made staying the rational choice.

Keralam's budget is tight. What it adds on top of the central rate is its own decision, and three moves fit inside what it can afford. A small number of state-level doctoral fellowships in the four priority domains, paid at competitive rates: not hundreds, perhaps fifty per year, well-paid enough to send the signal that the state values the work. A returnee researcher programme: Keralam-origin researchers abroad who would come back for a credible appointment with proper compensation. International research-fellow appointments: bringing twenty to thirty senior researchers from abroad each year in priority domains. The numbers are small enough to be fiscally manageable and large enough to change the institutional culture of the receiving departments.

*Figure: Converted at market rates, Keralam looks a tenth of Switzerland. Adjusted for what the money buys where it is spent, it is closer to a half. The second comparison is the fairer one, and the Keralam bar is the national fellowship: the University of Kerala's own fellowship paid about a third of it in 2022. Note which way Switzerland moves: it is the only bar here that shrinks when purchasing power is taken into account, which is what an expensive country does to a salary.*

#### What Keralam can do now

Keralam can fund the fellowships and the international appointments described above on its own authority. It can reform university intellectual-property policy to allow Keralam institutions to retain and commercialise research outputs from publicly funded work. State procurement can act as the first customer for locally developed research outputs in applied AI for public systems and in health diagnostics. None of these requires central permission.

#### What requires federal change

Topping up has a cost of its own: it creates two tiers of researcher inside the same department, doing the same work for different money. Rules on hiring foreign faculty and running joint degrees are written centrally too.

High-skill visa policy is a central matter too, and it is the basis of any international researcher programme. The returnee researcher and international research-fellow appointments named earlier all run through that constraint, and without movement on it, topping up only goes so far.

Central research money is not allocated to states in the first place. It goes to central mission agencies. In 2020-21 the central government accounted for 43.7% of all Indian R&D spending, and 84% of that flowed through twelve scientific agencies, with defence, space, agriculture and atomic energy taking close to three-quarters of the agency total between them. The higher education sector, where a state's universities sit, accounted for 8.8% of national R&D spending.[^11] What a state gets therefore depends on which central establishments sit inside it and what those establishments were built to do. Keralam hosts a large one: the Vikram Sarabhai Space Centre in Thiruvananthapuram is ISRO's lead centre for launch vehicle development, and the Department of Space took 18.4% of central agency R&D spending that year. That is real research capacity inside Keralam. Its core mission, launch vehicles, sits outside the four domains named above.

The instrument built to correct exactly this has its own record. India's Anusandhan National Research Foundation was legislated in 2023 with a five-year corpus of ₹50,000 crore, of which 72% was expected to come from private sources rather than from the exchequer. Its allocation for the first year was ₹2,000 crore. Within the same financial year that figure was revised down to ₹258.6 crore. A separate Research, Development and Innovation scheme worth ₹1 lakh crore followed in 2025, and, in the Department of Science and Technology's words, it "aims to promote a private sector–driven research and development ecosystem".[^14] The gap between announced and disbursed national research funding is the applause trap operating at the federal level. Keralam's state-level doctoral fellowships and international appointments are not supplementary to a functioning national system. They are what a sub-national research strategy has to look like when the national system disburses a fraction of what it announces.

A research base that compounds also requires patient capital: multi-decade funding commitments that a fiscally stretched state cannot provide alone. Where that capital comes from, and what kind of investor will hold a fifteen-year horizon for Keralam research and Keralam firms, is the question the next essay opens.

Keralam's publication count will keep rising. It is the one number in this essay that requires no decision from anyone. The concentration, the salaries, the laboratories that industry pays to sit next to: each of those waits on a decision the state has not made.

---

### Sources

[^1]: Elsevier, report on the research output of Kerala's state universities, 2018 to 2023, released through the Kerala State Higher Education Council. Summarised in [The New Indian Express, "Research output of varsities in Kerala rose significantly from 2018 to '23: Report" (15 March 2024)](https://www.newindianexpress.com/states/kerala/2024/Mar/15/research-output-of-varsities-in-kerala-rose-significantly-from-2018-to-23-report): an 11% compound annual growth rate in publications by state universities, and academic-corporate collaboration at 0.6%. The study itself has never been published online, and the Council's own publications index at kshec.kerala.gov.in did not respond at the time of writing, so no link is given here. That unavailability is the point the opening paragraph makes: the growth figure circulated widely, the study behind it did not.

[^2]: [University Grants Commission (Academic Collaboration between Indian and Foreign Higher Educational Institutions to offer Twinning, Joint Degree and Dual Degree Programmes) Regulations, 2022 (PDF)](https://www.ugc.ac.in/pdfnews/4555806_UGC-Acad-Collab-Regulations.pdf). Notified in the Gazette on 2 May 2022. Dual-degree candidates must earn at least 30% of total credits from the Indian institution, and twinning arrangements cap foreign-earned credit at the same 30%. The regulations set the terms on which credit for jointly supervised work is divided, require each collaboration agreement to cover intellectual property rights (clause 7.5), and say nothing about how the resulting citations are attributed.

[^3]: [Agency for Science, Technology and Research (A*STAR), Annual Report](https://www.a-star.edu.sg/about/annualreport): the research institute structure, the choice to concentrate on named domains, and the industry partnership model. On joint laboratories, see [National University of Singapore research and industry partnerships](https://www.nus.edu.sg/research).

[^4]: [ETH Zurich, "ETH spin-offs: an annual record set in 2023"](https://ethz.ch/en/news-and-events/eth-news/news/2024/01/press-release-eth-spin-offs-an-annual-record-set-in-2023.html): 43 spin-offs founded in 2023 and 583 since 1973. See also [ETH Entrepreneurship](https://ethz.ch/en/industry/entrepreneurship.html) and [EPFL Innovation Park](https://epfl-innovationpark.ch/). ETH publishes no per-franc density ranking against other European institutions, so the comparison here is limited to what the counts support.

[^5]: [World Bank and OECD, research and development expenditure as a share of GDP, Switzerland](https://data.worldbank.org/indicator/GB.XPD.RSDV.GD.ZS?locations=CH): 3.22% in 2023, against Sweden's 3.60%, Belgium's 3.27% and Austria's 3.26%. Switzerland is among the highest in Europe on this measure, not the highest. On who carries the research out: [Federal Statistical Office, "Almost CHF 25 billion dedicated to research in Switzerland in 2021" (25 May 2023)](https://www.bfs.admin.ch/asset/en/24368697): CHF 24.6 billion in 2021, carried out mainly within private enterprises (68%) and higher education institutions (28%). On the Basel cluster, see [Basel Area](https://www.baselarea.swiss/); on Roche's founding, [Roche, "History"](https://www.roche.com/about/history) ("F. Hoffmann-La Roche & Co was founded on 1 October, 1896, in Basel, Switzerland").

[^6]: Kerala's elderly share: Sample Registration System 2021 data, reported in ["Grey wave rising as South India's population ages fast"](https://thesouthfirst.com/health/grey-wave-rising-as-south-indias-population-ages-fast/). At 14.4% aged sixty or over, Keralam has the oldest population of any Indian state, ahead of Himachal Pradesh at 12.3%. On the institutional anchors, see [Sree Chitra Tirunal Institute for Medical Sciences and Technology, Thiruvananthapuram](https://www.sctimst.ac.in/) and [Amrita Institute of Medical Sciences, Kochi](https://www.amritahospitals.org/). On neglected tropical diseases, [World Health Organization, "Neglected tropical diseases"](https://www.who.int/health-topics/neglected-tropical-diseases) ("It is estimated that NTDs affect more than 1 billion people"; they "have historically ranked very low and almost absent from the global health policy agenda", gaining recognition in 2015 with the Sustainable Development Goals).

[^7]: [ICAR-Central Marine Fisheries Research Institute, Annual Report 2024](https://eprints.cmfri.org.in/19030/): the institute's research programme and its national resource assessment, which put India's marine fish landings at 3.45 million tonnes in 2024, down 2% on the year before. Earlier edition: [Annual Report 2023](https://eprints.cmfri.org.in/18531/). Coastline: [Ammini et al., "Marine fisheries in Kerala: an overview," Marine Fisheries Information Service T&E Series 204 (2010): 1–10](https://eprints.cmfri.org.in/6491/) ("Kerala coast runs for about 590 km").

[^8]: [Comptroller and Auditor General of India, Report No. 4 of 2018, General and Social Sector, Government of Kerala, Chapter III (PDF)](https://web.archive.org/web/20260116132700/https://cag.gov.in/uploads/download_audit_report/2017/Chapter_3_Local_Self_Government_Department_of_Report_No_4_of_2018_-_General_and_Social_Sector_Government_of_Kerala.pdf). The same performance audit the fourth essay in this series relies on, which describes Kudumbashree as "a mass women participation programme, encompassing 43 lakh members", federating from neighbourhood groups up through ward-level societies to a society in each local government. The programme's own current figures are roughly 46 lakh members across some 3.07 lakh neighbourhood groups.

[^9]: [University Grants Commission, revised fellowship rates for Junior and Senior Research Fellows](https://news.careers360.com/ugc-revises-fellowship-amounts-for-jrf-srf-post-doctoral-other-schemes): ₹37,000 a month for a Junior Research Fellow and ₹42,000 for a Senior Research Fellow, approved at the Commission's 572nd meeting on 20 September 2023 and effective from 1 January 2023. House rent allowance and contingency grants are paid on top. The University of Kerala's own fellowship: [order 2821/2022/UOK, 11 April 2022 (PDF)](https://fmnc.ac.in/wp-content/uploads/2020/03/Bijina-S-.pdf) ("eligible for fellowship @Rs. 13000/-(Rupees Thirteen Thousand only) per month as per U.O.No.Ac.E1/Bud.Sp.16/2016 dated 05.10.2016"). Average salaried earnings: [Ministry of Statistics and Programme Implementation, *Periodic Labour Force Survey Annual Report 2023-24*](https://mospi.gov.in/publication/annual-report-periodic-labour-force-survey-plfs-2023-24), Appendix A, Table (38): regular wage and salaried employees in Kerala earned ₹21,922 to ₹23,216 a month across the four quarters of 2023-24.

[^10]: [Swiss National Science Foundation, salary increase for doctoral students](https://www.snf.ch/en/kHunysxdKHpyUL8s/news/salary-increase-for-doctoral-students): a prescribed band of CHF 50,000 to CHF 55,000 gross a year from 1 January 2026, within which Swiss institutions set their own rates. Detail in the [salary scales, Annex XII (PDF)](https://media.snf.ch/Wg9oQo2WuIdqSIA/Annex_XII_Ausfuehrungsreglement_Beitragsreglement_E.pdf). Singapore rates from the [A*STAR Graduate Scholarship](https://www.a-star.edu.sg/scholarships/home/scholarships/ags--scholarship) (page as of 10 October 2026): on the STEM track SGD 4,500 a month for citizens, SGD 4,100 for permanent residents and SGD 3,600 for international scholars, and SGD 6,200, 5,600 and 5,000 on the specialised tracks. Two conversions are used. Market rates are August 2026: 95.42 rupees, 1.28 Singapore dollars and 0.81 Swiss francs to the US dollar. Purchasing-power figures use the [World Bank private-consumption PPP conversion factor](https://data.worldbank.org/indicator/PA.NUS.PRVT.PP) for 2025, at 19.84 rupees, 1.02 Singapore dollars and 1.07 Swiss francs to the international dollar. The private-consumption factor is used rather than the GDP one because the question here is what a stipend buys the person receiving it. On that basis a Keralam doctoral researcher receives roughly 1,900 to 2,100 international dollars a month, against 3,500 to 4,400 on Singapore's standard track and 3,900 to 4,300 in Switzerland. Switzerland is the only one of the three whose pay falls when the adjustment is applied, because Swiss prices sit well above the international average.

[^11]: [Department of Science and Technology, "Research and Development Statistics at a Glance 2022-23" (PDF)](https://web.archive.org/web/20251206162304/https://dst.gov.in/sites/default/files/Updated%20RD%20Statistics%20at%20a%20Glance%202022-23.pdf), NSTMIS, March 2023, reporting 2020-21 figures. National R&D expenditure by sector: Central Government 43.7%, Private Sector Industry 36.4%, Higher Education 8.8%, State Governments 6.7%, Public Sector Industry 4.4%. "During the year 2020-21, 84% of the R&D expenditure incurred by Central Government sources came from 12 major scientific agencies", of which DRDO took 30.7%, the Department of Space 18.4%, ICAR 12.4%, the Department of Atomic Energy 11.4%, CSIR 8.2% and DST 6.8%. The same publication records gross expenditure on R&D at 0.64% of GDP in 2020-21. On the Vikram Sarabhai Space Centre as ISRO's lead centre for launch vehicle development, [VSSC](https://www.vssc.gov.in/VSSC.html). Note that no official series publishes central research allocation broken down by state, which is why the argument here is made about agencies and establishments rather than about state shares. On visa provisions as a constraint on any international researcher programme, [Ministry of Home Affairs](https://www.mha.gov.in/).

[^12]: [Nature Index 2025 Research Leaders](https://www.nature.com/nature-index/research-leaders/2025) (the figures here are from the edition as published in June 2025, [archived copy](https://web.archive.org/web/20250611195052/https://www.nature.com/nature-index/research-leaders/2025/); the live page has since been revised), covering 2024 output across 145 natural-science and health-science journals: China's Share at 32,122 against the United States' 22,083, with China's adjusted Share up 17.4% over the year and America's down 10.1%. Summary via [Nature Portfolio](https://www.natureasia.com/en/info/press-releases/detail/9110). On doctoral output, [Center for Security and Emerging Technology, "China is Fast Outpacing U.S. STEM PhD Growth" (PDF)](https://cset.georgetown.edu/wp-content/uploads/China-is-Fast-Outpacing-U.S.-STEM-PhD-Growth.pdf), which projected more than 77,000 Chinese STEM doctorates in 2025 against 40,000 in the United States. That figure is a projection from enrolment data rather than an observed count, and is quoted as such.

[^13]: [Stanford HAI, AI Index Report 2026](https://hai.stanford.edu/ai-index/2026-ai-index-report), p. 66 and Figure 1.8.6: net flow of AI authors and inventors, "India had the largest net outflows at -16.9 in 2025". Also reported in ["India leads in AI talent, but also brain drain and anxiety"](https://theprint.in/india/governance/india-leads-in-ai-talent-but-also-brain-drain-anxiety-says-stanfords-ai-index-report/2909479/), The Print: India's net flow at -16.9 for 2025, the largest outflow among the countries charted, against Canada's -7.1 and Germany's -2.4. The same report places India second in the world for total AI authors and inventors, and first on relative AI skill penetration at three times the global average.

[^14]: Binay Panda, ["Budget 2024-25: operationalisation of ANRF may be too little too late"](https://www.downtoearth.org.in/science-technology/budget-2024-25-operationalisation-of-anrf-may-be-too-little-too-late), Down To Earth, 23 July 2024: the Anusandhan National Research Foundation's corpus of ₹50,000 crore over five years, 72% of it expected from private rather than government sources, and the revision of its 2023-24 allocation from ₹2,000 crore down to ₹258.6 crore. On the separate later scheme, [Department of Science and Technology, "PM launches Rs 1 Lakh Crore Research, Development and Innovation Scheme"](https://dst.gov.in/pm-launches-rs-1-lakh-crore-research-development-and-innovation-scheme), launched 3 November 2025, which the Department says "aims to promote a private sector–driven research and development ecosystem".

[^15]: [Comptroller and Auditor General of India, *Report No. 6 of 2021: Preparedness and Response to Floods in Kerala*](https://cag.gov.in/en/audit-report/details/114526), presented to the Legislature on 11 November 2021. Executive Summary: "Flood plains of the State are yet to be demarcated and flood plain zoning legislation remains to be enacted"; "No large-scale flood hazard map is available in the State".