One Industry, Built Whole

Somewhere in India there are silos holding thorium hydroxide. It was separated out of monazite dug from beach sand, much of it off the coasts of Keralam and Tamil Nadu, and the government’s own mineral yearbook records that the agency which produced it “has built a large stockpile of impure thorium hydroxide”.¹ The plant that separates it is at Aluva, in Keralam. The material has been mined, refined, and put away against a day that has not arrived. Thorium has fuelled power reactors in Germany and the United States, but no country runs its power programme on it.

That is not a scandal. It is a position. The thorium is already out of the ground, and the only open question is what anyone intends to do with it.

This series has argued for talent, for institutions, for clusters, for research, and for a different kind of capital. None of it decides what to build. Keralam has not stayed with one answer long enough to build around it.

The wrong inference

The obvious conclusion is that a stockpile like that is worth building an industry around. It is the wrong one, and Keralam has reached for it before.

In June 1947, with independence weeks away, the Dewan of Travancore declared that the state would stand alone, and Travancore at that moment held beach sands said to be exceptionally rich in thorium and a standing arrangement to ship nine thousand tonnes of monazite to Britain over three years.¹⁰ The arithmetic seemed to work: the new nuclear powers wanted monazite, and Travancore had it. Within weeks an attempt on the Dewan’s life ended the experiment and Travancore acceded.

The inference was that sitting on something scarce is itself a position. It was wrong then, and the eighty years since have not improved it. A deposit is worth rent to whoever processes it and a position to whoever can use it, and those are different parties.

The industry worth building is not thorium extraction, reactor construction, or a local claim on a national security resource. It is advanced materials, qualification and fuel-cycle engineering, capabilities whose customers begin in India and need not end there. Nothing here depends on thorium arriving on schedule, or at all.

What a choice costs

The gap shows in how the state presents itself. An investment summit that collects 499 letters of intent across every sector at once is not making a bet.² It is keeping 499 doors open, which is what a government does when it has not decided which room it wants to be in.

Breadth looks like ambition and costs nothing to announce. A choice is expensive, because everything that follows one has to be built, and most of it takes longer than the government that starts it. That is the real barrier, and it is not money. A state can fund a scheme in a budget speech. It cannot fund a fifteen-year institution in a budget speech, because the thing being bought is continuity, and continuity is paid for in elections rather than rupees.

What choosing looked like somewhere else

Taiwan chose. The story is usually told backwards, as though the island had a latent advantage waiting to be released.

It had none. In the mid-1980s, the podcast Acquired reckons, the average Taiwanese company ran gross margins of four to five per cent, which is the margin of a place that assembles what other people design. Morris Chang, instructed by a minister to build a semiconductor company, set out the starting position without flattering it: no strength in research and development, none in circuit design, little in sales and marketing, almost none in intellectual property. One possible strength, and by his own account a potential rather than an obvious one, in manufacturing wafers.³

His question follows from that inventory. What kind of company would you create to fit that strength and avoid all the other weaknesses?

Note what the question does not ask. It does not ask what the island was sitting on. Taiwan had no ore, no oil, no rare earths, and no deposit of anything a foreign buyer wanted.

The tidy version of what happened next makes it look like one clever decision, and it was not. Chang had spent thirty years in semiconductor manufacturing before anyone in Taipei rang him. His first success at Texas Instruments, in 1958, was a yield problem on a transistor line being built for IBM, and he went on to run TI’s worldwide semiconductor business before being passed over for the top job.³ Asked what Taiwan should build, he named the thing he had spent his working life doing.

Nor were semiconductors a fresh choice in 1985. Taiwan had committed to them a decade earlier and spent the intervening years building the capability to match. Chang inherited that decade rather than starting one.

Which makes the better question not what he chose but why a minister went looking for him at all. The state had built the capability and understood that it could not turn that into a globally competitive firm by promoting from inside its own bureaucracy. So it recruited one of the most experienced semiconductor manufacturing executives anywhere: fifty-four, recently passed over at TI, and unhappy in the job he took next.

The business model came from that career rather than from anything in Hsinchu. In his final months at General Instrument a founder called Gordie Campbell asked Chang for fifty million dollars to start a chip company, and when Chang called him three weeks later, said he needed only five, because he had decided against building a fab.³ Chang saw what that implied before the industry did. A foundry that makes other companies’ designs and competes with none of its customers is what you build if you think a generation of designers is about to be freed from owning factories.

Note the order. The institute came first, and the company fourteen years later.

The institute that measured the industry instead of itself

The company is the part everyone remembers. The part worth copying came first.

The Industrial Technology Research Institute was founded in 1973 out of three merged government laboratories, with a budget of NT$210 million and four hundred staff, of whom ten held doctorates.⁴ Three years later it signed a technology transfer with RCA for about ten million dollars. The technology was deliberately old: a 7.5-micron process, several generations behind what the American and Japanese leaders were running. Engineers went to RCA sites in New Jersey, Ohio, California and Florida for a year of training in design, process and quality control.

In 1977 the demonstration fab opened and reached a seventy per cent yield. RCA’s own American plants were running at fifty.⁴ The trainees beat the licensor inside a year, on a process several generations behind the leaders.

Then the institute did the thing that makes it worth studying. In 1980 it gave the demonstration plant away, to a new company called UMC, and over the next twenty years it spun out nine more.⁴ It set itself a target of fifteen per cent staff turnover a year, deliberately, and funded employees through doctorates and MBAs at Stanford and MIT at an average cost of four hundred thousand dollars each. By the late 1990s it held nine hundred doctorates among six thousand employees. Around eighty-five per cent of the people who left went to Taiwanese firms, and roughly forty per cent of the managers in the Hsinchu Science Park had worked at ITRI at some point.⁴

A research institute normally measures retention and treats departures as a failure of compensation. ITRI measured the industry, and counted the people it lost as its output. In 1993, after the firms it had created objected that it was no longer needed, the ministry cancelled the annual budget its founding terms had guaranteed.⁴

The place was built to match. Hsinchu covers six and a half square kilometres, with two universities beside it; more than fifty of its tenant companies began as spinoffs from those universities or from ITRI.⁵ Park sales went from NT$10.5 billion in 1985 to NT$1.6 trillion in 2005. The Hsinchu campus now holds about 420 companies and some 148,000 workers on that same small footprint.⁵

None of that is a subsidy schedule. It is an institute whose product was trained people, a park where the firms and the toolmakers and the universities were close enough to walk between, and twenty years of not changing the plan.

The usual dismissal is that Taiwan did this under authoritarian rule, so none of it transfers to a place with elections.

Martial law in Taiwan ran from May 1949 to July 1987, thirty-eight years.⁶ TSMC was incorporated that February, in the final months of it. The company listed in Taipei in 1994. The first direct presidential election was in 1996 and the first transfer of power between parties in 2000.⁶ The founding decisions, from ITRI in 1973 to TSMC’s incorporation, were all made under martial law. Most of the compounding, including the bulk of the park’s growth in sales, came after it ended, while Taiwan democratised. The institutions outlasted that transfer.

What survived the transition was not a strongman’s preference but a set of institutions: an institute with a budget line, a park beside two universities, a company with published accounts and a board. Keralam does not need more executive authority than it has. It needs the thing Taiwan turned out to have, which is continuity, a fifteen-year commitment that survives a five-year cycle and a front that alternates.

This is not a free lesson. Singapore, the other comparator this series keeps returning to, also bought coordination by narrowing the range of things that could be argued about. Both places got their industrial policy partly by constraining contestation, and contestation is precisely what Keralam’s social contract exists to protect. The honest version of the lesson is narrower than the usual one: the institutions can outlive the conditions that created them, but somebody has to build institutions rather than schemes.

Taiwan had no deposit of anything, and built the industry anyway. It bought old-generation technology, trained about thirty engineers on it for a year, and beat the seller’s own factories within twelve months. The endowment was never the point. The people were the point, and the institutions that kept turning out more of them were the point.

What Keralam holds

The Chavara deposit runs about twenty-two kilometres along the Kollam coast from Neendakara to Kayamkulam. Kerala’s own Department of Mining and Geology puts it at 127 million tonnes of heavy minerals, including 80 million tonnes of ilmenite, out of a raw sand reserve of around 1,400 million tonnes.⁷ Monazite, which carries the thorium, is a small fraction of the black sand by weight. India holds roughly an eighth of the world’s identified thorium.¹

The deposit is worked by two companies. The Kerala Minerals and Metals Limited is a state undertaking making titanium dioxide pigment at Chavara, with revenue on the order of nine hundred crore rupees. IREL is a Government of India enterprise. It has sat under the Department of Atomic Energy since 1963 and reported turnover of more than ₹2,024 crore in 2023-24 across its mineral-sands and rare-earth units.⁸ Kerala’s own mining department describes the barrier beach blocks as apportioned between the two.⁷

So Keralam’s position is that it hosts a world-class deposit, captures the titanium end of it through a state company, and watches the atomic end go to a central undertaking that refines a small part of the thorium for the Bhabha Atomic Research Centre and the Nuclear Fuel Complex and holds a large stockpile of it unrefined.

And then, this April, the future stopped being hypothetical. On 6 April 2026 the 500 megawatt Prototype Fast Breeder Reactor at Kalpakkam attained first criticality, having satisfied the Atomic Energy Regulatory Board. It was designed by the Indira Gandhi Centre for Atomic Research and built and commissioned by Bharatiya Nabhikiya Vidyut Nigam Limited.⁹ India has formally entered the second stage of the three-stage programme Homi Bhabha sketched in the 1950s, the stage whose reactors breed the fuel that makes the third, thorium, possible. On full operation India becomes the second country after Russia to run a commercial fast breeder.⁹

The geography of that sentence invites a wrong reading. The reactor is at Kalpakkam in Tamil Nadu, the centre that designed it is at Kalpakkam, and another of IREL’s beach sand plants is at Manavalakurichi, also in Tamil Nadu. Tamil Nadu built none of it. IGCAR, BHAVINI and IREL are central bodies, and where the Union puts its atomic establishment is settled in Delhi rather than in Chennai.⁹

That closes off the complaint, and what is left is worse. Keralam has hosted a central scientific institution of the first rank since 1962. The Vikram Sarabhai Space Centre at Thumba designs India’s launch vehicles, and around it in Thiruvananthapuram sit the Liquid Propulsion Systems Centre, the ISRO Inertial Systems Unit, the Indian Institute of Space Science and Technology and BrahMos Aerospace.¹³ Nothing was withheld from Keralam.

Sixty years on, the state’s technology press counts more than fifty spacetech firms, and reports that between them they handle less than a tenth of the work India’s space agencies contract out and that none has grown into a leading national player.¹³

Set that beside ITRI, which over a comparable span gave away a fab to UMC and spun out nine more companies, TSMC and Vanguard among them, with UMC in turn producing MediaTek.⁴ The institutions here are not weaker, and neither are the people in them. One of those bodies was built to put companies into the world and was measured on whether it had. The other was not asked to.

Keralam has spent this series being told it has the input that matters and no structure pointed at it. That is the same finding again, in a new sector.

So what is the capability worth building

If the state does not mine the sand, it has to build a capability instead. So where are the hard unsolved problems? Two of them face every nuclear programme, not only India’s.

The first is what neutrons do to materials. A reactor that burns thorium subjects its inner wall to intense neutron bombardment. So does any fusion machine that might one day breed the fuel faster. A first wall would take upwards of ten displacements per atom every operating year. Neutrons knock each atom out of its lattice position ten times a year, and the structure has to hold anyway.¹¹ Engineers have to test materials that survive that before anyone builds with them, and here is the gap: no facility anywhere in the world can currently irradiate candidate structural materials under a genuinely fusion-like neutron spectrum. The proposed answer, IFMIF-DONES in Spain, received its building permit only in June 2026.¹¹ An unsolved problem with no test rig is an opening.

Second, you cannot measure efficiency in isolation. When the laboratories reported a fusion breakthrough, they measured the energy at the target. Measure it at the wall plug instead, against the power the building drew from the grid, and the same shot returns a figure a hundred and fifty times smaller. Both numbers are correct.¹² They differ only in where the boundary is drawn. The gap between a celebrated laboratory result and a commercial system is an inventory of engineering work still available to whoever acquires the capability to do it.

Which brings the question back to Thiruvananthapuram, where there is already more of this than the state seems to notice. The Vikram Sarabhai Space Centre works on materials that have to survive launch and re-entry. CSIR’s National Institute for Interdisciplinary Science and Technology works on advanced materials and process engineering. Sree Chitra Tirunal Institute qualifies materials that go inside human bodies, which is a regulatory discipline as demanding as any reactor code. The Indian Institute of Space Science and Technology, IISER Thiruvananthapuram, CUSAT and NIT Calicut all feed the same pool.¹³

None of that amounts to a nuclear industry. What it amounts to is the potential strength Chang described rather than the obvious one, which is close to the raw material ITRI was handed in 1973 with its four hundred staff and ten doctorates.

What has to be true by 2040

This series picked 2040 as its horizon, and a horizon is not a forecast. The question each essay asks is what would have to be true by then, and for this one the answer does not depend on any particular technology arriving on schedule.

A materials and fuel-cycle capability is worth building on four counts. It is learnable, in the way wafer yield was learnable, by acquiring processes other people have already worked out and training about thirty engineers in them. It requires nothing further from the coast, because the thorium in question has already been mined and is sitting in storage. It sits upstream of a bottleneck that every serious programme will hit, which means the customer is not only Indian. And it survives its own schedule slipping, because the same materials science, separation chemistry and qualification discipline serve fission reactors, medical isotopes and space power whatever else happens.

That last property is the one that makes it a responsible bet rather than a gamble on a technology that may take longer than anyone says.

The institution this requires

Name it whatever it ends up being called. A Keralam Institute for Materials and Qualification will do as a working label, and the name is the least important thing about it. The mandate is the thing, and five parts of it are not negotiable if the point is to copy what worked rather than the words.

A statutory mandate of fifteen years, with a board the government does not control. ITRI kept its guaranteed budget for twenty years and lost it in 1993, when the firms it had seeded argued it was no longer necessary and the ministry agreed.⁴ Fifteen years is the minimum span over which a materials programme produces anything, and it is three governments. A board that a chief minister can reconstitute is a board that will be reconstituted.

Funding by formula, argued once. The most useful thing Morris Chang did on returning to TSMC was to fix research spending at a set percentage of revenue whatever the year, so that his research director never had to make the case again.³ A state institute has no revenue to take a percentage of, so the equivalent is a statutory share of something predictable rather than a line item competing afresh in every budget. The principle survives the translation. An institute that has to re-win its existence each year will spend its best people on winning it.

A training pipeline with defined numbers, and no bond. Doctoral places and multi-year overseas placements in materials science, irradiation testing and separation chemistry, sized in hundreds rather than dozens. Tie them to opportunities worth returning for rather than to a contractual obligation to return. Keralam has fifty years of evidence that people leave anyway, and little reason to think compelling them would work; it has rather less evidence about what happens when there is something specific to come back to.

A spinout mandate, and a measure that admits it. The institute should be assessed on firms created and on where its alumni end up, not on how many staff it retained. This is the part that will be hardest to get past a state audit process, because every existing instrument treats an employee leaving as a loss. ITRI targeted fifteen per cent attrition on purpose and ended up staffing forty per cent of Hsinchu’s management.⁴

One campus, not fourteen grants. This is where the geography matters, and it argues against the obvious answer. Keralam’s materials research sits in Thiruvananthapuram. Its materials industry sits seventy-odd kilometres north in Kollam, where KMML already does pigment chemistry and process engineering next to the deposit itself.⁸ Hsinchu was built seventy to eighty kilometres from Taipei, next to the institutes and universities rather than next to the government.⁵ The equivalent here is a corridor of almost exactly that length, and the campus belongs at the Kollam end of it, close to the industry and far enough from the capital not to be absorbed into it.

Dispersing the same money as grants across the institutions that already exist would be easier, would offend nobody, and would produce papers. The entire lesson of the last fifty years of industrial policy is that proximity is not a nice-to-have. People have to be able to walk between buildings.

What requires Delhi

The hard parts are not Keralam’s to decide. Atomic energy is a Union subject. Thorium and uranium are prescribed substances under the Atomic Energy Act of 1962, and under the Sustainable Harnessing and Advancement of Nuclear Energy for Transforming India (SHANTI) Act that Parliament passed in December 2025 to replace it they stay under central control: thorium that is mined or extracted vests in the Union, and only government bodies and government companies may mine it. IREL answers to the Department of Atomic Energy rather than to Thiruvananthapuram.¹⁴ A state cannot licence itself into this sector.

What has changed is who may apply. Under the same Act, as its provisions come into force, any government company, including one a state owns, and any other company may seek a licence to build and run a reactor or to make nuclear fuel, while reprocessing, heavy water and enrichment beyond a notified threshold stay with the Union. The government has set targets of about 22.5 gigawatts of nuclear capacity by 2031-32 and 100 gigawatts by 2047.¹⁵ Keralam already owns a government company working the Chavara sands. A state with trained people and a functioning institute is in a position to use that licence. A state with a deposit and a grievance is not.

Keralam has been holding this material for as long as it has been a state, and holding is the only verb that applies. The reactor that might one day use it went up in the next state along, built by an institution that did not exist when the sand was first dug and does now.

There is a second thing Keralam is holding, and it is not financial either. It is older than the sand, most of it is still standing, and almost none of what it is worth has ever appeared on anyone’s books. The next essay tries to put a number on it.

Sources

1. Indian Bureau of Mines, Indian Minerals Yearbook 2022, Part III, “Rare Earths” (December 2023), pp. 24-3 and 24-6: IREL’s Udyogamandal plant at Aluva separates thorium as hydroxide from monazite, a small part is refined to nuclear-grade thorium oxide for BARC and the Nuclear Fuel Complex, and “IREL has built a large stockpile of impure thorium hydroxide”. India’s thorium resources: World Nuclear Association, “Nuclear Power in India”, “about 13% of the world total”; world thorium resources by country, WNA, “Thorium”.

2. Invest Kerala Global Summit, February 2025: “A total of 499 LoIs, with a huge investment of Rs 1.80 lakh crore together, were received at the IKGS”, from Kerala Industries Department figures announced on 12 November 2025, as reported in “Kerala hits century in turning investment summit projects into reality”, Outlook. The sixth essay in this series, “Builders, Not Investors”, uses the same release.

3. Morris Chang on Taiwan’s position at the time TSMC was conceived, and on the founding question. Quoted from his remarks as reproduced in Acquired’s TSMC episodes (Season 9, Episode 3, 2021; and the interview with Dr Chang, 2025). Chang’s career at Texas Instruments from 1958, including the yield problem on the IBM transistor line and his rise to head of the worldwide semiconductor business, and the account of Gordie Campbell telling Chang, when Chang called him three weeks later, that he needed five million rather than fifty because he would not be building a fab, are from the same interviews. Note that the podcast is a narrative source. It rounds TSMC’s founding equity to half from the government and 28% from Philips; a 1990 Dataquest company backgrounder held by the Computer History Museum gives 48.3% for the Executive Yuan Development Fund, 27.5% for Philips and 24.2% for Taiwanese industrial partners.

4. The Industrial Technology Research Institute: 1973 founding from three merged laboratories, NT$210 million initial budget, 400 staff of whom 10 held doctorates; the RCA technology transfer of 1976 at approximately US$10 million for a 7.5-micron process; engineer training at RCA sites in New Jersey, Ohio, California and Florida; the 1977 demonstration fab reaching 70% yield against RCA’s 50%; the gift of that plant to UMC in 1980 and nine further spinouts; the deliberate 15% annual staff turnover target; PhD and Master of Business Administration (MBA) sponsorship at Stanford and MIT averaging about US$400,000; 900 doctorates among 6,000 employees by the late 1990s; approximately 85% of departing staff joining Taiwanese firms and roughly 40% of Hsinchu managers being ITRI alumni; and the 1993 cancellation of the guaranteed annual budget. “The Institute Behind Taiwan’s Chip Dominance”, Asterisk, and ITRI’s own institutional history.

5. Hsinchu Science Park: area; 417 registered companies (register updated June 2026) and 148,270 workers (August 2026) on the Hsinchu campus alone, against 592 companies and 181,683 workers across the bureau’s six campuses (company register; employment); the two universities beside the park; more than 50 tenant firms originating as spinoffs from ITRI, National Tsing Hua University and National Chiao Tung University. Hsinchu Science Park Bureau and National Science and Technology Council statistics. Park sales of NT$10.5 billion in 1985 and NT$1.6 trillion in 2005: Office of the President, Republic of China (Taiwan), “The Vice President Attends the Global High-tech Industry Forum and the 25th Anniversary of Hsinchu Science-based Industrial Park” (15 December 2005), reporting Vice President Annette Lu’s speech.

6. Martial law in Taiwan, declared on 19 May 1949 and lifted in July 1987, after thirty-eight years: TaiwanPlus, “Taiwan Remembers the White Terror, Martial Law” (19 May 2026) and Lowy Institute, “The end of martial law: An important anniversary for Taiwan” (13 July 2017). TSMC was incorporated on 21 February 1987 and listed on the Taiwan Stock Exchange on 5 September 1994: TSMC, consolidated financial statements for 2023, note 1, Form 6-K filed with the US Securities and Exchange Commission. The first direct presidential election in 1996 and the first transfer of executive power between parties in 2000: Government Portal of the Republic of China (Taiwan), “History”.

7. Department of Mining and Geology, Government of Kerala: the Chavara deposit between Neendakara and Kayamkulam, estimated at 127 million tonnes of heavy minerals with an ilmenite content of 80 million tonnes from a raw sand reserve of the order of 1,400 million tonnes; and the apportionment of the Chavara barrier beach blocks between KMML and IREL.

8. The Kerala Minerals and Metals Limited, a Government of Kerala undertaking at Chavara producing titanium dioxide pigment: turnover ₹891.66 crore in 2022-23 (KMML Annual Report 2022-23). IREL (India) Limited, a Government of India enterprise under the Department of Atomic Energy since 1963, with mineral-sands and rare-earth units at Chavara, Manavalakurichi, Aluva and OSCOM: turnover above ₹2,024 crore in 2023-24 (IREL).

9. Department of Atomic Energy, “Prototype Fast Breeder Reactor at Kalpakkam, Tamil Nadu attains First Criticality”: the 500 MWe PFBR attained first criticality on 6 April 2026, after meeting all stipulations of the Atomic Energy Regulatory Board; designed by the Indira Gandhi Centre for Atomic Research and built and commissioned by Bharatiya Nabhikiya Vidyut Nigam Limited. India’s entry into the second stage of its three-stage programme, and its becoming “only the second country in the world after Russia to operate a commercial fast breeder reactor” once fully operational: Press Information Bureau, backgrounder of 7 April 2026 (PDF).

10. Travancore’s declaration of independence under Dewan Sir C.P. Ramaswami Aiyar in June 1947, the monazite agreement with Britain, the attack on the Dewan and the Maharaja’s accession: Sandeep Bamzai, “The plots and manoeuvrings behind the making of independent India, Part 2”, Observer Research Foundation, 16 November 2020 (excerpted from Princestan), drawing on Itty Abraham, “Rare Earths: The Cold War in the Annals of Travancore”, in Entangled Geographies. How far monazite drove the Dewan’s decision is argued in that literature, not settled.

11. Neutron damage to fusion first-wall and in-vessel components in excess of 10 displacements per atom per operating year, and the absence of any existing facility able to irradiate candidate structural materials under a fusion-relevant neutron spectrum. IFMIF / IFMIF-DONES, the proposed international fusion materials irradiation facility. The Escúzar Town Council granted IFMIF-DONES its building permit on 16 June 2026, with on-site construction “scheduled for August 2026” (IFMIF-DONES Spain Consortium, 25 June 2026).

12. The two boundaries, in figures. National Ignition Facility, shot of 5 December 2022: 2.05 megajoules of laser energy delivered to the target yielded 3.15 megajoules of fusion energy (Lawrence Livermore National Laboratory, 14 December 2022). The peer-reviewed account gives 3.1 megajoules and a target gain of 1.5: H. Abu-Shawareb et al., “Achievement of Target Gain Larger than Unity in an Inertial Fusion Experiment”, Physical Review Letters 132, 065102 (2024). The laser drew about 300 megajoules from the grid for the shot, so the fusion energy was about 1% of it, a gain at the wall plug of around 0.01: LLNL’s Mark Herrmann in Physics Today, 13 December 2022. LLNL puts the gain needed for an economically viable power plant at 50 to 100: Tammy Ma, “Ignition and the Path Toward an Inertial Fusion Energy Future”, HDIAC webinar, 13 February 2025.

13. The Vikram Sarabhai Space Centre at Thumba, Thiruvananthapuram, established as the Thumba Equatorial Rocket Launching Station in 1962 and operating under the Department of Space, where India’s launch vehicles are designed; and the presence in the same city of the Liquid Propulsion Systems Centre, the ISRO Inertial Systems Unit, the Indian Institute of Space Science and Technology and BrahMos Aerospace. The count of more than fifty spacetech firms in Keralam, and their handling under a tenth of the work India’s space agencies outsource, are from Hari Kumar, “New consortium aims to lift Kerala spacetech”, Kerala Technology, 11 March 2025: “There are over 50 spacetech firms in Kerala, yet they currently handle less than 10 per cent of the work outsourced by Indian space agencies”. That none has become a leading national player is from Hari Kumar, “Aerospace chamber hopes for an extra thrust”, Kerala Technology, 30 April 2024: “none of them have been able to scale up to a higher level or become leading national players”. Both are the reporter’s account, not a state policy document. Other research institutions in Thiruvananthapuram and Keralam working on materials and related disciplines: the Vikram Sarabhai Space Centre, CSIR’s National Institute for Interdisciplinary Science and Technology, the Sree Chitra Tirunal Institute for Medical Sciences and Technology, the Indian Institute of Space Science and Technology, IISER Thiruvananthapuram, Cochin University of Science and Technology and NIT Calicut, all of which are represented in the Thiruvananthapuram chapter of the Materials Research Society of India.

14. The Sustainable Harnessing and Advancement of Nuclear Energy for Transforming India (SHANTI) Act, 2025 (No. 39 of 2025), Gazette of India, 21 December 2025: thorium as a prescribed substance (s.2(33)), mining of uranium and thorium by government bodies and government companies only (s.5(2)), vesting in the Central Government (s.5(5)), and repeal of the Atomic Energy Act, 1962 (s.91). Atomic energy is a Union subject under the Seventh Schedule of the Constitution. IREL (India) Limited is under the administrative control of the Department of Atomic Energy.

15. Licensing under the SHANTI Act, 2025: eligible applicants, including government companies and other companies (s.3(1)); licensable activities, including reactors, fuel fabrication and enrichment up to a notified threshold (s.3(2)); activities reserved to the Central Government, namely enrichment beyond that threshold, spent-fuel reprocessing and heavy-water production (s.3(5)); provisions in force from dates the Central Government notifies (s.1(2)). World Nuclear Association, “Nuclear Power in India” for the passage of the Act in December 2025 and the targets of about 22.5 GWe by 2031 and at least 100 GWe by 2047.