Pranay Kotasthane on navigating an uncertain world
Between the powerplays of modern geopolitics, innovation, and, well, Bangalore potholes.
Hi folks, Pranav (Agarwal) here.
With each podcast we do, here’s something we are starting to understand. You bring different guests for different reasons. There are guests that you bring on to understand a very specific subject, and then there are guests that you bring because they’ve spent years thinking seriously about a whole range of questions, all of which shape the world.
On that note, we recently interviewed one such guest: Pranay Kotasthane. Pranay is a deputy director at the Takshashila Institution where his work covers everything from public policy to international relations to high tech geopolitics. Just as important as his research, however, is how he takes the time to make his ideas known. He co-hosts Puliyabaazi and co-writes Anticipating the Unintended, both of which come with that mix of being deep and nuanced, yet accessible.
One of the first things that you realize about Pranay when you start following him is the sheer range of things that he has an interesting perspective about. He thinks about subjects that don’t obviously sit together, but the way he talks about them, they all feel connected. There’s a common thread running through them, an interest in systems, institutions, incentives, and the unintended consequences that hide behind the surface of big projects. You’re going to see a lot of that in this episode as well.
Below are the moments from this chat that, in my opinion, deserve their own space:
Resource leverage is a card whose value falls the moment you play it
The whole “China has us by the throat on rare earths” panic gets reframed in one sentence. The metaphor is good in itself, but its real payoff is the corollary that follows — that the only actor capable of triggering serious China Plus One diversification is China, and it is now doing that to itself. This is a useful inversion of the prevailing fatalism on critical minerals.
“But the one framework to keep in mind is that resource leverage is an exhaustible resource. It’s like a card you play, but its value reduces the moment you play it. […] Resource leverage — the moment China uses it — is making China Plus One happen. I always say the only country that can make China Plus One happen is China, and it is doing that now.”
Synthetic rubber existed since 1908 — geopolitics just productised it
Specific historical detail that makes the substitution argument land in a way pure theory wouldn’t. The fact that the technology already existed for three decades and only got built out when access to natural rubber was cut is the strongest single piece of evidence for why critical minerals doomerism is overdone. Worth listening for the cobalt-to-molybdenum analogue from the 1980s that immediately follows it as well.
“In the 1930s, the dual-use material of the world was natural rubber — used in military applications, for tanks, bullets, tyres in the wider economy, etc. Most of it came from Southeast Asia. When Southeast Asia was occupied during World War II by Japan, the West’s entire access to this dual-use material stopped. But we didn’t have a rubber crisis. Why? Because there was something called synthetic rubber — which already existed since 1908. The technique was invented in Germany, but no one had bothered with it because natural rubber was coming cheaply from Southeast Asia. Why even bother? Because of the geopolitics, synthetic rubber technology was quickly productised — it already existed, but by the end of the Second World War, we had put synthetic rubber in place as a substitute, very quickly.”
India’s entire rare earth import bill is smaller than what Americans spend on avocados
This is the kind of stat that should be pinned to every critical minerals briefing in the country. It explains, in one number, why no private sector activity exists in the space — and why the policy response is mostly about freeing prices to do their job rather than building cathedrals of industrial policy.
“I go back to prices. A lot of activity in rare earths was not happening because prices were so low. Total rare earth imports last year were around $186 million — less than what the U.S. spends on avocados. So if prices are so low, why would anyone do it? But now because of China’s restrictions, prices are rising.”
The Reasi lithium block has still not been sold — and the difference between exploration and mining is why
The exploration-vs-mining distinction is one of those policy points that is technically not hidden, but almost no one in the public conversation surfaces it. The contrast between global first-come-first-served licensing and India’s auction-everything reflex is genuinely underappreciated, and the Reasi example makes it concrete.
“Globally, exploration is on a first-come, first-served basis. You can apply for an exploration. Within a few days, you can get approval. You do the exploration and then you give a report, and then see whether someone wants to mine. In India, it is run through auctions — a heavily political instrument. They take a lot of time, and eventually many of our blocks don’t even reach the stage where mining can begin because exploration hasn’t happened. […] An example I always give: a few years ago, there was news that India had found large lithium reserves in Reasi district, Jammu and Kashmir. And we thought our lithium problem was solved. As of now, that block has not even been sold.”
Ingredients, not recipes
A clean way of phrasing why “let’s just copy East Asia” never works. The conventional manufacturing discourse in India veers between two unhelpful poles — either reverence for some specific Korean or Taiwanese policy package, or fatalism about how nothing here will ever work. The framing of an emergent system that needs the right ingredients but produces its own recipe avoids both. The Tamil Nadu non-leather footwear example that follows is a good demonstration.
“The way I think of it is: you have to think of ingredients, not recipes. You can’t borrow South Korea’s recipe, but you can try to identify what were the ingredients in South Korea’s policies — so that emergence can happen and you can have a new recipe in India.”
Two 12-hour shifts vs three 8-hour shifts is why Taiwanese chip assembly hasn’t come
This is the single most concrete binding-constraint observation in the manufacturing section. It cuts through the abstract “labour reform” debate and surfaces a specific, addressable parameter. It also implicitly makes a federalism point — if one or two states were willing to experiment, the entire national logjam could move.
“In semiconductors, there are no Taiwanese assembly firms that have come to India, even though Taiwan is a very big player on that side. Why? Two reasons they have given: labour law reforms. Most Taiwanese firms operate on two 12-hour shifts. India mandates three 8-hour shifts — and for assembly, profit margins are wafer thin. They say we can’t operate with three shifts; we’d have to employ extra people. None of our states allow 12-hour shifts. If you want that industry from Taiwan, you might have to give that option. Let the workers decide. Let the state decide. One state might take a chance — UP might take a chance, where there are no employment opportunities and people have to go outside.”
NVIDIA’s real moat is CUDA, and that is where India’s leverage might lie
Most of the AI-and-India discourse is split between people who want sovereign LLMs and people who want a local fab. Pranay points to a third option — the firmware translation layer between models and silicon — that almost nobody is talking about and that maps cleanly onto India’s existing comparative advantage in software and chip design. This is the most actionable AI-strategy idea in the conversation.
“NVIDIA’s dominance exists because of CUDA — the CUDA libraries work best with NVIDIA chips and not with AMD GPUs or other GPUs. That’s the real moat NVIDIA has. Because we already have some capability one layer up, in Python, can we actually finance open source communities that are building alternatives to CUDA? Alternatives and models — essentially firmware that works with many chips. Whether you have an AMD GPU, an NVIDIA GPU, or a TPU, it’ll just work as well because you have a translation layer on top. Can that be our focus?”
51% vs 3%
The single statistic that should reframe how Indians talk about Chinese governance. The reflexive “they can do it because they’re authoritarian” line collapses once you absorb that China is, in fiscal terms, dramatically more decentralised than India.
“There are stats to prove this: 51% of Chinese government expenditure happens below the provincial level. In India, that number is 3%. We are not decentralised. We say we are federal. We are federal in that sense. But federalism and decentralisation are two different things, and we are not truly decentralised.”
The freedom struggle was led by mayors
Striking enough on its own as a piece of historical recall, but its real value is what it implies — that local political careers were once how serious people built national stature in India, and that the disappearance of that pipeline is both a symptom and a cause of how stunted urban government has become.
“We don’t have meaningful, functional mayors — which is very ironic, because all our major leaders of the freedom struggle were mayors of their cities. Jawaharlal Nehru was mayor of Allahabad, Subhas Chandra Bose had a long career in the Kolkata city government. Sardar Vallabhbhai Patel — many of these leaders had a strong local presence and emerged from their cities. Today, do we know the mayors of our cities? All our cities are run by IAS officers.”
The full conversation is available on YouTube, Spotify and Apple Podcasts. If you prefer reading instead, you can find the full transcript of the podcast below. Pleae note: this transcript has been generated using AI, and may contain minor errors or inaccuracies.
Pranav Agarwal: So Pranay, thank you so much for joining us.
Pranay Kotasthane: Glad to be here. Pranav, I do keep reading your daily brief, so it’s great to be here.
Pranav Agarwal: That is extremely flattering to hear, partly because I have been reading you at least since 2018. So to know the possibility that you’re reading is an incredible thing.
So I think it was around when I started reading you, this is before COVID, that you started talking about a world that was turning towards something called the Matsya Nyaya, where it was no longer important that countries felt like they had to be seen to be right. That power was all that really mattered, I think.
I don’t know what I felt back then, but seeing how things have been over the last year, we have most definitely entered into a Matsya Nyaya state of the world. Now, if this is the reality that India has to live in, then how do we look at things? What is the chess board in front of us? What are the possibilities that we can play for? How likely are all those possibilities?
Pranay Kotasthane: So I mean, I will put it this way, that international relations was always Matsya Nyaya. So Matsya Nyaya is a concept in the Arthashastra, which really talks about international relations space as an amoral space. Unlike domestic affairs, where there is a morality most commonly affirmed by something called a constitution, even the prime minister or the president has to abide by certain rules and procedures which are laid down in the Constitution. That is, there’s a morality, and we can call it the constitutional morality. International relations is not like that.
I think it’ll be good to start with a framework in this idea that great powers seek not just power, but they seek authority. And authority is a combination of two things. Power and legitimacy. Okay. The way we have to think is because it is an amoral space, the big powers always have more things that they can do, but that’s not their endpoint. They want authority because you don’t want to exercise power every time. The exercise of power is time consuming. It drains your energy. So you want to also exercise legitimacy. Your exercise of power should be deemed legitimate, like you said, by most other people.
And how do you do that? The way you do that is most great powers will create certain kinds of institutions, and they will also follow some of the rules of the institutions. It’ll always be unequal because it is an amoral space. It’s not as if US has the same power as Maldives has in the international space. There’s definitely a difference, and that’s the way it is. Even though we have these constructs — the Westphalian construct — every nation has sovereignty. There are certain ideas that we have agreed to, but in terms of power, US can do a lot more things and India can do fewer things and Maldives can do even fewer things. That’s how it is.
So, in that sense, what great powers are trying to do is build their own institutions and also themselves follow some of these institutional norms so that power is deemed to be legitimate by others. Legitimacy disarms others and it makes them follow your rules without you having to coerce them explicitly. So that was the world which, after the World War II, we entered a phase where there were two big powers and both of them tried to create their own institutions, tried to affirm… follow some of those rules themselves. And we were in that phase till the nineties.
When USSR collapsed, we entered that unipolar moment where only the U.S. power and legitimacy was high, and it was able to do a lot of things. And slowly starting with — I would say, Obama’s period in the U.S. already — we started a period where U.S. power was still high, but its legitimacy was going down because it was exercising power in ways which were thought to be illegitimate in the eyes of other, nation states. So it started with the botched up Afghanistan invasion. Even before that, in George Bush’s time, the Iraq invasion, Afghanistan, finally U.S. had to go back.
And then now in this Trump era, in the U.S. it’s even more so right because, there, there is an alienation of allies. There is also the exercise of power, which is having huge negative consequences. We are speaking, even as the West Asian war is happening, no one knows what the objective of this was. Why is it still continuing? What will it imply for the world? These things… and U.S. itself backed out from many of the institutions that it upheld. So it created a lot of things — like the Bretton Woods institutions, It followed it.
But now, U.S. has quit the climate change accords. It has moved back on many of its commitments. WTO is no longer functional. So that legitimacy part is reduced now, even though power of US is very high. So it was able to do things in Venezuela and Iran, which no other nation state is able to stop them from doing. But their legitimacy has reduced.
Pranav Agarwal: So what specifically can the U.S. not do because it is bleeding legitimacy. Where do the constraints start to bite in?
Pranay Kotasthane: Yeah. So I think to understand that we have to see why the U.S. is doing these things. There, the Chinese angle also comes in. Till the unipolar moment there was this idea that we have no other equal — is how the U.S. was thinking. But slowly it was China. Also, in fact, the U.S. helped China’s rise in many ways, getting China into the WTO, a lot of U.S. companies investing in China. And that really led to China doing great things.
But slowly we saw that China’s power differential with US, at least in U.S. perception, was reducing. And that led it to do many things, which, I would’ve said, they could have handled it better. But anyways, we are in this phase where U.S. legitimacy is decreasing and there’s no other power to fill that vacuum — because even China does not have a great alternative to the U.S. rules-based orders. When people use the word rules-based order, it’s not order in, like, a normative sense, it’s U.S.’s rules-based order. The rules that applied were the U.S. rules. Now, but, there’s no China’s rules based order, which the world is willing to accept because China equally has been coercive. Its relationship with its neighbors is not great and it does not have the power that U.S. has.
Anyways, so we are in that phase, which people have termed… like former NSA Shivshankar Menon terms it, this is a world disorder. So there is… earlier there was some world order and we are in this transition phase where there is still the U.S. as a big power, but China is not yet there to fill its shoes. And there is this transition phase where we are looking towards a change in the world.
Or it might go in different ways. There might be places where, again, you have a U.S.-China G2. India hopes it will be a multipolar order where there will be multiple poles — at least that’s what the government says. And then there are other ideas. Maybe if the U.S. is able to bounce back, it can again return as a unipolar order. So there are various scenarios that we have tried to map out in our book, India’s Marathon also, where there are multiple geopolitical scenarios, there are multiple geo-economic trends and you have 20 worlds which can emerge out of it. We’ve tried to map them, but yeah, we are in that phase.
US, because of the acts that it has done through Trump and even before that, its legitimacy is decreasing. In fact, hypocrisy is very important in international relations. People always agree: Is India being hypocritical? Yeah! I mean, in an amoral space, every nation state has to be hypocritical. But you have to do hypocrisy in an artful way. You should have to be able to do hypocrisy, and yet the others should not be able to allege that you are hypocritical. U.S. was hypocritical, but it was able to still say we support democracy. And most people believed that they support democracy even though they were supporting the dictators of the world. And we never questioned that. We still had this norms: democracy is great. There used to be freedom ratings index, which people used to closely follow, Why was that happening? Because it had this. The greatest form of soft power is normative power, right? You, because you are reading US books, you are influenced by what, their, the writings come from them. You also cognitively think in those ways. That normative power no longer exists, right?
In this Matsya Nyaya age, it is very clear that they are not able to do hypocrisy artfully, the hypocrisy has been exposed. And that’s the real challenge.
Pranav Agarwal: So if I am India in a phase like this, where even if I try to say that I believe in, say, international free trade or democracy or any of that, most people would either not take me seriously or they won’t care. Because it seems like, at least the people who wanted to establish that way of doing things, they have retreated from the space. So how should I, as India, start seeing the game? Because I don’t have the hard power to do the same things as the U.S. is, what can I do?
Pranay Kotasthane: So again, when we talk about Matsya Nyaya, the currency of international relations is power. So we have to do things which increase our power. That’s why democracy, pluralism, free trade — we are not saying this because these are good for the world. We have to do this because they’re good for us. They are in our national interest. We know it from our experience.
And for example, there was more freer trade — I will not call it free trade, freer trade — between the period 2004 to 2008, before the global financial crisis. That was the golden period for the Indian economy. We grew 8% year on year. Our income tax collections ballooned. So we know that it works for us, and that’s why we want to support it. Cheaper imports mean we can do value addition and export. That’s why we want freer trade.
So all of this, from our point of view, is not a normative idea — that we should do this because it’s the right thing, because the U.S. told us so or something. We do this because it is in our national interest. Pluralism? Absolutely. For a big country like India, pluralism is absolutely necessary. So is democracy.
India’s principle has never been that because we are doing things in a certain way, others should also adopt it — which has been the U.S. idea. The grand strategy of the U.S. is very Alexandrian in nature. India has never been so, even when it was a bigger global power in Asia. India is what India. So I think India’s biggest value lies in showing that the Indian model can succeed. That itself becomes a big thing that other nation states observe. Other nation states follow your lead, not because of what you say, but because of what you do and what you represent.
So if India is succeeding and doing well in itself, that becomes a lesson for many countries. Even when India was a much smaller power, it had enormous normative legitimacy immediately after the Second World War. When India was independent, lots of newly independent nation states were looking up to India — many of their constitutions drew from what India was doing. So in that sense, India’s goal should be to become the most successful power using its model, as quickly as possible. That’s what we want to do. If other nation states find resonance in it, they will follow.
Pranav Agarwal: One of the places where Matsya Nyaya is very clearly playing out is the space of critical minerals. And you have written quite extensively about critical minerals.
But here’s the thing: when I started to actually read about critical minerals, I realized that it was such a vast space that I couldn’t even begin to imagine how to use this idea intellectually.
As soon as you start reading about critical minerals, you realize that there is no one thing happening. There are actually dozens of stories playing out together. There’s a separate critical minerals race for batteries, a separate one for magnets, a separate one for chips. There are hegemons in this space. There are power brokers, bit players, has-beens. There’s a story touching everything from salty lakes in Bolivia to sludge-filled, toxic areas in Inner Mongolia. There is just an infinity of things all happening together, all neatly labelled as critical minerals — primarily because different countries have just put a bunch of things in a list.
As a scholar who is actually trying to understand this extremely vast space, what are the frameworks you apply? How do you make sense of all this chaos?
Pranay Kotasthane: So again, I will try to talk about two frameworks here. One critical idea is this: what’s happening is called the weaponisation of resources. And it’s not new — it has happened before. I’ll give you some examples. But the one framework to keep in mind is that resource leverage is an exhaustible resource. It’s like a card you play, but its value reduces the moment you play it.
So now, these kinds of restrictions on rare earths, or even forget critical minerals; even chips. When the U.S. plays this card of ‘I won’t give you chips,’ it immediately makes other nation states realise: we have to do something on our own, or between ourselves, because this can happen again. And similarly, when China does this with critical minerals, other nation states know they need to get their act together.
There is this idea called creative insecurity, which Mark Taylor talks about in his book “The Politics of Innovation”. The idea is that whenever the sum total of external threats is greater than the sum total of internal divisions, nation states tend to spend more on R&D and innovation. So all these things which are happening are fuelling creative insecurity. There’s empirical analysis behind this — for example, Taylor writes that Sweden does a lot more innovative things than Denmark, even though the two countries are broadly equal on many parameters: GDP per capita, geography, and so on. The difference is that Sweden faces the Russia threat more than Denmark does. It faces greater creative insecurity and has tended to invest more in R&D and innovation.
That is one angle. Resource leverage — the moment China uses it — is making China Plus One happen. I always say the only country that can make China Plus One happen is China, and it is doing that now.
The other lens is that prices are wonderful. Prices are really a decentralising coordination mechanism. Alex Tabarrok talks about how prices are a signal wrapped in an incentive. That’s what Hayek wrote about in his wonderful paper ‘The Use of Knowledge in Society.’
We didn’t need to know about critical minerals until now — even though those critical minerals have been part of our lives since the 18th century. Cerium was in your cigarette lighter. It’s been with us 250 years or more. Why are we discussing them now? Because we are not allowing markets to operate — we have put up these barriers. The way prices normally work is as a coordination mechanism: only the people who needed to know these materials knew about them. We didn’t need to know. People didn’t even know what TSMC was, because you didn’t need to know. We operate at different levels of abstraction, right? We operate in different kinds of abstractions. We had the luxury that we didn’t need to know which is the metal in my phone. Others in that industry knew, because prices were the signal coordinating them: as the price of this increases, I will probably procure less of it, I will procure something more, I will find an alternative: all that.
Now, with these two frameworks in mind, we can understand the critical minerals risk. Critical minerals are largely non-fuel minerals which are essential in use and subject to supply risks — this is a Congressional Research Service definition. The supply risks can be geologic, technical, economic, environmental. We’re not talking about oil and gas — these are non-fuel. And they were essentially ubiquitous. They were everywhere, and we were fine with it. We didn’t even need to know what they were because they were being coordinated.
China became the dominant supplier of one category of critical minerals: rare earths. Rare earths are a group of 17 elements that are not actually rare in terms of their presence in the earth’s crust. Even the rarest of rare earths are around 200 times more common in the Earth’s crust than gold. But their discovery, extraction, and refining is difficult, and they’re chemically so similar that separating them from each other is hard. That’s why the term ‘rare’ is generally used.
China became dominant in the refining of rare earths, and we were fine with it because China was supplying them cheaply to everyone in the world. It is very environmentally polluting — both the extraction process and the refining process, where you increase the concentration of rare earths, generally using acid leaching. So the West was very happy to let it move to China. These technologies were invented in the West — America was the largest producer of rare earths until the fifties and sixties — and then production moved to China. China did it cheaply, and because their investment-led model tends toward structural overcapacity, it was much cheaper than the market rate would have set. It was exported cheaply, just like they’re now doing for electric vehicles and solar panels — same story.
Because it was largely flowing cheaply to the rest of the world, the world was okay to accept it. But what that meant is we were foregoing resilience for cheapness. That was good until we hit this period where everything has become weaponised again. So now, the U.S. restricted chips to China. China said: we also have cards to play. We will restrict these critical minerals — not just to the U.S., but to the entire world. And that has led to the situation we are in today.
Pranav Agarwal: It’s almost as though, if you want to preserve your points of leverage, you have to have legitimacy — which goes back to our previous question.
Pranay Kotasthane: Yeah. So now what are other countries doing? That’s the interesting bit here. First thing: is it really difficult for other countries to produce rare earths? It is difficult, but it’s not that difficult to learn refining — it’s an old process. You can do that in other countries too. And do you have mines in other places? As I said, it is abundant. Sure, you might have to increase output from some mines. You might have to even make some new mines, which takes a lot of time.
But there is also the recycling dimension. And this is where oil is very different from many of these minerals. You can recover lithium from end-of-life lithium batteries. Lithium doesn’t get combusted by its use. Oil, obviously, is a fuel mineral — we derive energy by combusting it. So you can’t recover oil from your petrol two-wheeler, but you can recover lithium from your electric two-wheeler. A lot can be done in recycling as well.
So many countries have now woken up to the fact that China will use this as leverage. The next time you have bad relations with China, it’s very clear — their Commerce Ministry now has a department that issues export licences not just for high-end materials, but even for diamond cutters, wire saws, and a whole lot of things. In fact, under the October 2025 rules, they also said that if a Chinese national is trying to build rare earth capabilities outside China, they can be jailed or punished. So these are ridiculous restrictions where they tried to weaponise everything.
And whenever you do that, there is creative insecurity. Other countries are trying to find ways out. The most interesting part of this is what I call substitution. It’s not as if you can’t build without rare earths. Tesla’s two- and four-wheelers — their first motors didn’t have rare earth magnets. Why are rare earth magnets used? Because they are very powerful. You can achieve a lot of power and torque in the motor at a small form factor, and you can achieve high temperatures without losing magnetic power, and things like that. That’s where permanent magnets are used. But you can also use induction motors, or other kinds of motors where efficiency will be less — and you can find other ways to compensate for that efficiency.
What is now happening is many companies are doing work on substituting rare earths totally. We know of Indian companies saying their next generation of two-wheelers will not use rare earths at all. Some are working on using micrograms instead of milligrams and getting the same efficiency. So all these things are already in place. Tesla has announced its next generation of cars will be rare earth free. So as I said in the beginning — once you use resource leverage, others start modifying. You might gain brownie points the first time you play it, but others quickly start finding alternatives.
Let me give you some cool historical examples of this. In the 1930s, the dual-use material of the world was natural rubber — used in military applications, for tanks, bullets, tyres in the wider economy, etc.. Most of it came from Southeast Asia. When Southeast Asia was occupied during World War II by Japan, the West’s entire access to this dual-use material stopped. But we didn’t have a rubber crisis. Why? Because there was something called synthetic rubber — which already existed since 1908. The technique was invented in Germany, but no one had bothered with it because natural rubber was coming cheaply from Southeast Asia. Why even bother? Because of the geopolitics, synthetic rubber technology was quickly productised — it already existed, but by the end of the Second World War, we had put synthetic rubber in place as a substitute, very quickly.
The COVID-19 vaccine was another example — within that period, people found alternatives, developed a vaccine, sidestepped many regulations that would’ve taken five years, condensed them to one year and got the vaccine. Those are the kinds of things that happen, and China is making that happen now.
In the 1950s, mercury was a critical mineral used in walkie-talkies during the Korean War. China restricted it. People thought we’ll not get access to this — how will we build walkie talkies? But it was built; substitutes were built. In the 1980s, cobalt — even now, cobalt is highly concentrated in Congo, in the DRC — , and it was occupied during that time by some rebels. They were close to the USSR. There was this term called ‘materials myopia’ — that we don’t look at materials, while the USSR is looking at materials. We have to do something.
And eventually the solution that came up was not a cartel — just like oil has cartels, there have been many attempts to build a copper cartel, there have been attempts to build a cartel for cobalt, but none of them have worked because they are different. So, during this “cobalt myopia”, a bunch of things happened: nation states could build stockpiles. You can’t build huge stockpiles of oil, but you can stockpile many of these critical minerals, which are required in smaller quantities. Cobalt was heavily used in jet engine superalloys, and people started asking: can we use molybdenum as a substitute? People knew it could work but hadn’t bothered because cobalt was available. When that pressure came, molybdenum got productised, carbon fibres got productised. And then new mines opened up in places like Canada and Australia, which were substitutes.
You don’t need to reduce your reliance on China 100%. You should just know that there are substitutes. Once you have 30-40% capacity built in other places, and the demand reduces through substitutes, then the leverage reduces.
My conclusion — my claim — is that by 2028, China’s geopolitical leverage over rare earths will go away. I’m seeing a lot of interesting things happening in other countries, companies, and including in India — and it will not be the kind of thing that it is used. So this is an optimistic take. Most people think China’s leverage over rare earths is so strong that we cannot recover. But you don’t need to reduce it to zero. You just need to have alternates and substitutes — which I think are on the horizon.
Pranav Agarwal: So if I am India looking at this situation, there are two things open to me. One: if I look at how China managed to build its stranglehold over rare earths, I might think this is a time to do a lot of industrial policy — deploy IREL, incentivise capturing all the nodes I can, as quickly as possible through various subsidies. The other way I can think about this is to say: let’s hold off. Markets can find a way around them. In some cases that’s already working out — as you mentioned, a lot of Indian auto manufacturers are already finding ways to make motors without rare earth magnets. What is the ideal mix of strategy for India here?
Pranay Kotasthane: The framework I would apply here is comparative advantage. Other countries are doing many things. The U.S. is going very heavily on industrial policy in this space. The Department of War has an equity stake in MP Materials, a rare earths company. They are not just doing equity — they also have an offtake agreement with that company that even if your prices are higher than Chinese rare earth suppliers, we will buy from you for 10 years at a fixed price. So there are price guarantees, offtake guarantees, cash flow mechanisms — all of this the U.S. is doing. And they’re giving a lot of funding and equity to other companies building rare earth magnets too. Many countries are doing that.
Some of it India also has to do. But what is our comparative advantage? I would start from our comparative advantage. Our comparative advantage would be in a few segments. One is that we do have a large resource of rare earths, found in our monazite sands. The constraint there is not so much industrial policy — it is regulations. Our monazite sands have rare earths, but they are also coterminus with thorium, which is important for India’s three-stage nuclear cycle. So because of that, no private company can extract these rare earths. Only IREL — Indian Rare Earths Limited, under the Department of Atomic Energy — is mandated to do so.
I think we don’t need to have that restriction anymore. Rare earths are really important, and you can have mandates: whatever thorium you get, you return it to a stockpile for DAE. That’s fine. But why not open this to our private companies? Get the energy in because prices are rising. I go back to prices. A lot of activity in rare earths was not happening because prices were so low. Total rare earth imports last year were around $186 million — less than what the U.S. spends on avocados. So if prices are so low, why would anyone do it? But now because of China’s restrictions, prices are rising. Which means private companies would want to do this. And extraction from sand will be quicker — it won’t take the 10-15 years required for deep mining. It’ll also be environmentally less polluting. If you can identify specific areas, you don’t have to do deep mining. You can build a feedstock within two to three years. So that’s one segment for rare earths.
There are also other things like gallium, which is found coterminus with aluminium. We do have a lot of aluminium being extracted, but no one was doing gallium extraction because, again, why would you when you can buy it cheaply? But now China has also put restrictions on gallium and germanium, prices are rising. So can we enable private companies to do that? That’s another segment.
The other segment where industrial policy is required is exploration. Exploration is different from mining. In mining, you first do prospecting — you try to determine whether you actually have the mineral in high concentration in a particular area or not, right? Once that’s done, mining starts. Exploration essentially solves a market failure called information asymmetry. No miner is going to put 10 years of money and effort into a project if they’re not close to certain there’s a huge resource available at this particular concentration here.
Globally, exploration is on a first-come, first-served basis. You can apply for an exploration. Within a few days, you can get approval. You do the exploration and then you give a report, and then see whether someone wants to mine. In India, it is run through auctions — a heavily political instrument. They take a lot of time, and eventually many of our blocks don’t even reach the stage where mining can begin because exploration hasn’t happened.
Pranav Agarwal: We’re betting money on an outcome you have no idea about.
Pranay Kotasthane: That’s exactly why no investment comes and mining doesn’t start. An example I always give: a few years ago, there was news that India had found large lithium reserves in Reasi district, Jammu and Kashmir.
And we thought our lithium problem was solved. As of now, that block has not even been sold. When it was found, someone still had to do the exploration and prospecting to confirm: this much resource exists, at this concentration, and it’s economical to extract. All that has to be done at the exploration stage — and that has not been done. So I think there’s an industrial policy case for the government to fund exploration. It should be done on a first-come, first-served basis so you can quickly conclude whether a reserve is high-value or low-value. If the government is solving a market failure of information asymmetry, I’m okay with them putting money into it. The government does have exploration licences, but it’s still not moving quickly enough.
The third element is recycling. We’ve written about urban mining of rare earths because we have a huge market that is going to generate a lot of electronic components reaching end of life, and electric vehicles reaching end of life too. We have even bigger sources of rare earths and lithium through that. We should actually be importing some of this e-waste because then we can extract rare earths and lithium from it — it’s environmentally less polluting, and we’d have a bigger feedstock because we are a large market. Can we stimulate the recycling space? We have some EPR mandates that make companies look at end-of-life cycles, but those EPRs don’t prioritise critical minerals right now. If we design regulations to incentivise EPR mandates to extract rare earths and lithium, we can have a secondary stream within two to three years. Again, it won’t solve all your problems, but you don’t need it to — if you have a credible alternative, that’s enough.
The fourth element — which is outside the comparative advantage space — is diplomacy. At no point will we be able to have all the critical minerals we need. It’s like playing a whack-a-mole game. We’ll solve rare earths today. Tomorrow scandium will be important, yttrium will be important. The way you have to think is: can we build diplomatic arrangements with friendly nations for joint stockpiles? Some of these materials are required in very small quantities — terbium, for instance, is not needed in huge quantities. So you can build national stockpiles when China’s prices go low. Just like we have strategic petroleum reserves, it’s okay to have a strategic mineral reserve — or better still, a joint reserve. Maybe a Quad reserve, maybe with Japan. And depending on your contribution, you can draw from it.
We can also think of offtake guarantees across countries. The U.S. approach with MP Materials alone is very inefficient — you’re dependent on only one company. Maybe U.S. companies can have offtake mechanisms with an Indian company, or India can have something with a Japanese company. Those kinds of arrangements are required. You can’t become Atmanirbhar in this space. You will have to think of the Atma Shakti perspective — where is your power — build your leverage so others don’t try to coerce you, and then you can do things with your friends as well.
Pranav Agarwal: So I want to move a little bit now to manufacturing. I think manufacturing is one of our personal obsessions at the Daily Brief. And we have read enormous numbers of papers on what is holding back India’s manufacturing. Here’s how we are beginning to look at it. There’s one way of talking about the problems with Indian manufacturing where you just start listing: oh, Indian manufacturing is overregulated. We have labour issues, we don’t have clear land titles, electricity is cross-subsidised, and so on and so forth. You can have a long laundry list of problems. Naturally, all of them need to be solved. But solving so much is a multi-decade project.
Right now, imagine I am a secretary in some heavy industries department. At this moment I do not have the will or the political capital to build a wide consensus and change everything. But I have seen some successes. I have seen India successfully create a smartphone manufacturing ecosystem — it has had its own agglomeration effects, and lots of things have come around it. Or I have seen India, in just a few years, build the beginnings of a semiconductor assembly and testing cluster. And that is the thing I want to replicate. My political masters have already bought into the idea that we should do industrial policy. But now I want to answer the question: if I have to do industrial policy, how do I replicate what worked? What can we learn from those experiences?
Pranay Kotasthane: The lens I would apply here is the framework of complexity. These are complex adaptive systems — they are non-linear. There is no simple action-and-direct-result chain that will be linearly followed. The system has many emergent things happening in it. Bangalore traffic is a complex adaptive system. One vehicle stopping can have non-linear effects on the outer ring road. Similarly, manufacturing has second, third, fourth-order effects. You have to approach it from the lens of complexity.
This means the conventional way of thinking — see what East Asia has done, learn best practices, apply to India — is not what works in a complex adaptive system, because many nodes are connected in different ways. A great policy which works in South Korea might not work in India because the initial conditions are very different. For example, South Korea always had a small market, so they were always focused on export. If you want to do manufacturing, the only way is through export — you can’t succeed without that. And once you’re exporting, you automatically bring in competitiveness because you’re competing with the world. You can give a lot of subsidies to your companies, but finally you’re measured by how globally competitive you are. That discipline is built in through exports. We don’t have that, unfortunately, because big countries always face this problem: we can just close off foreign players, we have a reasonably okay domestic market which is growing, and we can just serve that market. Why do we have to look at export?
So what South Korea does, we can’t simply follow. The way I think of it is: you have to think of ingredients, not recipes. You can’t borrow South Korea’s recipe, but you can try to identify what were the ingredients in South Korea’s policies — so that emergence can happen and you can have a new recipe in India. Those ingredients are, unfortunately, all the things you mentioned. Better trade policies, better tax policies, better labour reforms, low electricity prices. These are the ingredients which, put together, allow interesting and emergent things to happen. I don’t necessarily think of it from a sectoral approach — identify a sector, do industrial policy there. You just put these ingredients in place, and our entrepreneurs are good enough to find what’s next.
An example of this in recent times is non-leather footwear. Tamil Nadu has become, very quickly, a major player in this segment. The ingredients: cheap labour, reasonably good electricity, some capital equipment. There was also a government angle. Non-leather footwear is manufactured very interestingly, much like semiconductors — a lot of it is contracted out. Nike doesn’t manufacture your shoes; they contract it out to a Taiwanese firm, just like in semiconductors. So there are contract manufacturers in Taiwan who used to manufacture these. Tamil Nadu was able to get many of those contract manufacturers to set up their bases there. A lot of Nike, Crocs, etc. are now manufactured in smaller towns in India.
That happened because the ingredients were in place — reasonably priced electricity, good infrastructure. Things have improved in India on many of these fronts. Infrastructure is better because you are now connected to global supply chains. With GST, the earlier problem went away — every company used to need godowns in every state because we weren’t one market. People used to sometimes export from Tamil Nadu to a port in Kerala by going through Colombo because it was cheaper. Literally. With GST, even though tax rates are still high, that problem is gone. You don’t have huge trucks waiting at border checkposts. Some tax rationalisation has happened. Electricity prices are still high, though — we haven’t solved our DisCom problem. That ingredient is still missing.
In fact, in 2004 we had the National Manufacturing Competitiveness Council, formed under the Manmohan Singh government, headed by V. Krishnamurthy, who had set up Maruti in India. That report already has the things we need to do. We need to go back to it and see what constraints remain. Some have been removed. But labour and land policy are still constraints, and there’s no shortcut. This deregulation commission idea is really important. The four labour codes are important. We’re not yet where we wanted to be on that, but we’re in the right direction. We have to keep accelerating.
Let me give you an example. In semiconductors, there are no Taiwanese assembly firms that have come to India, even though Taiwan is a very big player on that side. Why? Two reasons they have given: labour law reforms. Most Taiwanese firms operate on two 12-hour shifts. India mandates three 8-hour shifts — and for assembly, profit margins are wafer thin. They say we can’t operate with three shifts; we’d have to employ extra people. None of our states allow 12-hour shifts. If you want that industry from Taiwan, you might have to give that option. Let the workers decide. Let the state decide. One state might take a chance — UP might take a chance, where there are no employment opportunities and people have to go outside. Maybe Tamil Nadu or Karnataka doesn’t want to — that’s fine. But right now none of our states are allowing that. This is a binding constraint.
So PLI, et cetera — these are one way we’ve found, but PLI is a bandaid, not a solution. PLI might work in one area where we already have comparative advantage, giving it a boost. But as we’ve seen from the PLI experience, of 14-15 PLIs, only one or two are showing positive results. For the others, you need to solve the binding constraint. Deregulation, electricity reforms, labour reforms, land policy — these are crucial.
China succeeded because of three basic ingredients. First, education. There’s a nice paper by Nitin Bharti and others looking at a hundred years of education trajectory in China, from 1900 to 2000. Even before communism, China invested a lot in human capital from the 1920s. We had good universities, but not a basic education system. So by the time China opened in 1978, they had a workforce that could quickly learn and adapt. Second, bureaucratic decentralisation. A lot of things in China don’t happen because Xi Jinping says so — there’s a lot of competition between local governments. China is actually more decentralised than we are, even though it’s an authoritarian regime.
The third ingredient is land policy. Land was easy for companies to acquire in China. Right now, land is a really big constraint in India. Land acquisition policy is supposed to be an exception — a company should be able to go and buy land from a person directly. But it has become the rule: for all big projects, the government has to acquire the land and hand it to the company. That means our land markets are completely broken. Why would a company invest when they need the government to acquire land for them? China was able to do it through other means — they can coerce people. But land remains a very big constraint here. There are ideas floating around, but we need to solve them if we want to see manufacturing as a percentage of GDP rise consistently to 25%, which was the goal of the National Manufacturing Competitiveness Council.
Pranav Agarwal: So Pranay, one of the problems we have run into is that when you look at the failure cases, you have all of these lists of answers. But when you look at what worked, it seems like the answers weren’t actually in policy as much as they were in political elbow grease. When Foxconn comes in, the things that make it successful are managing lots of competing groups. The things that almost held Apple manufacturing in India back were things like worker hostels not being good, or the bathrooms not being cleaned. There is a huge amount of on-the-ground management required. So if you are embarking on this project of bringing lots of global industry to India, how much of it happens at the 40,000-foot policy level, and how much happens on the ground solving problems like ensuring worker housing is adequate?
Pranay Kotasthane: I’m not sure about that, because we have to compare: what was the Foxconn situation in China — in Zhengzhou, where it has its biggest manufacturing facility — when they started? You have to compare their starting point and India’s starting point. India’s starting point is much better in terms of worker quality and the freedoms workers have.
If Foxconn sees a problem, it’s in their interest to fix it. They will build what they need to build. They will face concerns on land acquisition — if we solve those, they will sort out the rest for their own interest. I think things like worker hostels were reported as barriers far more than they actually were. How do you establish the causality? Is Foxconn not doing this because a worker hostel wasn’t there? I don’t think so. A hostel can be built. Is the binding constraint that they were not getting land for the hostel? Then yes, we have to look at that.
Pranav Agarwal: What I meant rather was that there is a lot of solving of mundane problems on the ground. You can get your policy package right, but you still need someone on the ground going and talking to people and making things happen. Is there a last mile of government service delivery that we really need to perfect in order to get a lot of manufacturing going?
Pranay Kotasthane: Yes, there are last-mile things to do. A lot of this deregulation will only work if last-mile implementation happens — for example, how our labour inspectors work. All of these have touchpoints at the ground level. But the kinds of things governments have to do at the ground level should be about deregulation and ease of doing business, and how these translate on the ground — rather than building big facilities or hostels. Companies can build those on their own, if we can support them on the things that matter.
Pranav Agarwal: Fair. So I want to now shift to artificial intelligence — really understand what India should do, how India should think about this. We’re entering a very weird era. When I look at the space, I see two very different kinds of narratives emerge.
On the one hand, you have this notion that a lot of the dynamism in our economy came from our service exports, and a lot of that was doing simple, replicable work for rich clients. Now, AI has come in at exactly that node. It has eaten exactly that advantage, and that should spell disaster for India’s service exports. Very difficult questions are being asked of companies like Infosys or TCS: how are you doing it? How are you rethinking your models? And it isn’t clear that they have perfect answers. That is a very scary story.
And then there is a more optimistic story that you hear if you drive down to HSR or Koramangala, where there are lots and lots of people sitting at coffee tables asking questions like: okay, if I can sell the service for $60 and my API token costs $35, I can make a business. I don’t have to be as rich as Sam Altman.
And I can also imagine that there are thousands of Indians doing this, who will become successful and create a different world. What I do not know, however, is: looking 10, 20 years into the future, which of these stories succeeds? Is it the pessimistic story of our advantage being frittered away? Or is it the optimistic story of a lot of people creating new kinds of ventures? How do you see this juncture?
Pranay Kotasthane: It’s an interesting thing to debate. I think to start off with, I’m not so worried about the software services companies. I mean, they should be worried — but I’m not worried from an Indian national interest, public policy point of view. That’s an industry where India has actually done no protectionism, and those companies have survived on the basis of their own capabilities. They have seen many transitions. They withstood the Y2K crisis, the GFC, many other humps — they managed, they adjusted. So I think they will find a way out. There is a challenge in that you can’t have just their model of pricing based on the number of employees or contractors you provide — that model might not work in this era. They have to figure out a different model and use AI before their customers do.
I think some companies will go down, some will learn, and some — like you said — there will be new Infosys-es of the AI era, new HSR-based companies doing more interesting things. That will happen. I’m not so worried from a national point of view for this sector.
What I’m looking at is this: if we think of AI as a general purpose technology, there are various narratives people have used. Some people think of this as nuclear weapons — I should get it first and deny it to others. Some look at it as an innovation race. Some as an arms race. Some as a general purpose technology. I think we should look at it as a general purpose technology, just like electricity. And if you look at that frame, you have to see what happened in past technological revolutions.
There’s an excellent paper by Mike Mazarr, written for the RAND think tank, which argues that success in the AI era is a societal challenge and not a technological one. I think that framing really resonates with me. Innovation first happens in some other country, but the country which is able to deploy it in useful and interesting ways is always the one that leads. For example, the new chemical revolution happened in Germany — a lot of the processes we learned in chemistry have German names. But it was implemented and diffused in the U.S. because the U.S. established chemical engineering departments first, and the productisation of those inventions happened in the U.S. more than anywhere else. This comes from Jeffrey Ding’s great book on technological revolutions. So if we take this frame, we need to think about how AI can be deployed and how we can prepare our society to make use of it in the best way — because this will take time, just like electricity took 100 years. This won’t take 100 years, but is our society prepared for this?
From this framing, Mike Mazarr identifies seven factors based on past technological revolutions that determine which countries succeed in a technological transition era. I wrote them down. They are: one, national ambition and will; two, unified national identity; three, shared opportunity; four, an active and engaged society; five, effective government institutions; six, a learning and adaptive intellectual climate; and seven, diversity and pluralism. These factors, based on a historical analysis of technological revolutions, determine whether a country becomes the U.S. of the chemical revolution era or becomes the Ottoman Empire, which declines. Technology always creates new powers, but what creates new power is societal factors — not so much whether you have access to chips or models.
The models are now available to many countries. DeepSeek changed that on the model side, and on chips, alternatives will come. GPU is just an architecture that worked for parallel processing, but there will be newer architectures, more efficient models. One focus area is: do I have models? Do I have chips? Government should think of it. But I think these seven factors are what ultimately matter.
Pranav Agarwal: Could you make those seven factors a little more intuitive? What exactly are they pointing at? How did it work in the past? How did societal factors actually beat out hard innovation?
Pranay Kotasthane: So national ambition and will — during the Cold War era, the U.S.-USSR competition was very acute. Most of the time, people felt the USSR would outpace the U.S. because it was more centrally coordinated. They were making five-year plans, the U.S. was not, and they seemed to be pulling ahead. There was a Sputnik moment where people in the U.S. could see the satellite from their naked eyes, and it really led to creative insecurity. That national ambition and will — the U.S. put together the National Science Foundation, a new NASA, ARPA which became DARPA. Vannevar Bush’s landmark report was part of all of that — all of it signalled there is a national ambition and will to get ahead. In today’s context, India now has this idea of Viksit Bharat — whether it’ll work or not is a different question, but the fact that we have articulated a measurable national ambition, becoming a developed country at $11,500 per capita, is a good sign. Once that is a focus area, we try to align things towards that goal.
The other factor is unified national identity. That is still a problem in India — we still fall into different buckets and are often unable to come together when unity is required. Then if you look at shared opportunity — that factor can reasonably work in India, because of how people are seeing things like UPI: benefits going to many people, not just the richest. If the users of AI are able to help people across socioeconomic groups, that creates a shared opportunity sense. That’s why some of the things India is doing — language models on Indian languages, multimodal, multilingual — are useful. They can convey a sense of shared opportunity to a broader number of people.
Effective government institutions are our weak link. We still have government institutions that are not set up for a technological revolution. That is a governance challenge. These are the things which are important. Adaptive learning and intellectual climate — we have to change our education policies. Can we get out of the credentialism mindset and actually enable people to adopt these tools? That’s where governments have to focus.
If I have to look at this from a geopolitics mindset, I would go to the frame of: in the near term, are there domains under AI where we have a comparative advantage and can occupy one segment of the supply chain and do that really well? Instead of focusing on having everything from data centres to models. Sure, we need those, and some companies will build them. But as a government priority, what should we focus on?
I would look at our strengths. Our strengths are in software and firmware. Python is a layer where a lot of Indian contributors have built. So can we think of an open source strategy? NVIDIA’s dominance exists because of CUDA — the CUDA libraries work best with NVIDIA chips and not with AMD GPUs or other GPUs. That’s the real moat NVIDIA has. Because we already have some capability one layer up, in Python, can we actually finance open source communities that are building alternatives to CUDA? Alternatives and models — essentially firmware that works with many chips. Whether you have an AMD GPU, an NVIDIA GPU, or a TPU, it’ll just work as well because you have a translation layer on top. Can that be our focus?
Can we have more inference-level chips focused on that? These can be open source — just like you have RISC-V as an open-source alternative to ARM, can you have a RISC-V alternative for low-end inference as well? These are the things I think we can do because we have talent in design and talent in software. If I had to focus, I would prioritise these over others.
Pranav Agarwal: So one thing I did not think about until maybe two weeks ago was technological sovereignty — it sounded like a very far-off thing, until I started reading about Palantir and Claude and all of what the U.S. can do today. And now I don’t know what to think.
There are these deep learning models that have thousands of datasets enmeshed on each other until the computer itself can identify specific people, specific cars, things like that. You have Claude added on top of that, so you can actually talk to all of those data streams and gather ridiculous amounts of intelligence together. You can create systems that can take out heads of states on the first attack of a war. And if I’m India looking at that, I am probably at least a little afraid. If this is the frontier of technological capability, and we are already seeing it deployed at such a large scale, what do I do to have this capability? What does having my own models matter in this era?
Pranay Kotasthane: I think the way to look at this is: what does technological sovereignty actually mean? There are different versions of this idea. Technological sovereignty could mean a full autarkic version — I need to have everything in the AI stack. That could be one vision.
The second vision: I just need to have some base level capabilities across the value chain. In nuclear weapons, are we technologically sovereign? We still import some uranium, but we have the base capabilities. We don’t have as many nuclear weapons as Russia, the U.S., or China, but we have enough to make mutually assured destruction apply — so don’t mess with us. Similarly, in space — we don’t have the world’s best space program, but we know we have certain capabilities, and we have the ASAT test to show that we can take things down if you act funny with us. That’s the second vision: base capabilities across the value chain.
Third is an innovation vision — do we have the ingredients to innovate, regardless of what happens? That’s how many in Europe see sovereignty. It’s actually about R&D and innovation: put those ingredients in place and people will figure out what’s next after LLMs, and they’ll build that instead of chasing the previous game.
The fourth vision of sovereignty is asymmetric advantage — you might have some capabilities across, but is there a segment where you have so much dominance that others would not want to mess with you? Because they know you can use that leverage. So without using the leverage, you communicate your power just by being so dominant in that segment.
I think it’s very difficult for any country to pursue the autarkic vision. The U.S. will still be dependent on talent from outside. China will be dependent on chips from outside, or ASML machines from outside. No one is fully technologically sovereign. As we are seeing, it’s a whack-a-mole game — you try to do a lot of things at home, but there will still be one choke point you hadn’t thought of, and someone can use that leverage.
So if I am India, given my starting point — we’re thinking about many of these things right now and have to build many of them — I would look from the fourth vision. Can we, from a geopolitical point of view, develop expertise in certain segments where the world is so dependent on us that they don’t mess with us? What would those capabilities be? Since we have 20% of the chip design talent here, can we quickly convert that into having more companies here? If foreign companies are here, they know they can’t quickly move out tomorrow — even if they wanted to, they wouldn’t find that density of trained people in other countries. So they’re automatically tied to you.
For defence-related applications, sovereignty matters a lot more. For defence, we will need our own models, our own compute, and I think defence will be able to assemble that. But from a broader economic sense, we need to think about where we can have asymmetric leverage. And the government’s role is also to have base capabilities in some segments. That’s why I think having one fab is important. It won’t make you self-sufficient — even at full capacity, it won’t meet 10% of our 28-nanometer chip demand. But you have that base capability, you have the talent, you have the supply chains to build something 20 years from now. One fab is good. Ten fabs? We don’t need that. But you build base capability in certain segments, and in others you think about where you can get leverage quickly and invest there.
Pranav Agarwal: I think finally, what I wanted to talk about was the city that we’re in. It is one of the great cities of South Asia — a beautiful mix of culture, incredible weather, science and technology talent, a lot of things mixing together. It’s probably one of the most dynamic cities in this part of the world.
It is also a very angry city. It is a city where everyone feels neglected. You feel neglected on the roads. You feel neglected on the footpaths. You feel neglected when you see the bridges. There’s a constant feeling that the city is being broken down. And so many people have different answers for this. Some say there are outsiders who have taken over the city and don’t care about its wellbeing. There are people who think the population is too much for the city to sustain. There are people who think the problem is corruption.
The one thing I really loved about Takshashila’s work on urban governance is that you actually started asking the question: who is supposed to run this city? What powers do they have? Who is paying them, who ensures they’re in office next year? Could you help us understand what is happening in Bangalore, and how do you solve a problem like Bangalore using the kinds of frameworks you have?
Pranay Kotasthane: So before coming to the framework, there are many misconceptions to unpack. Is Bangalore overpopulated? Absolutely not. There’s nothing called overpopulation. Nitin Pai, co-founder of Takshashila, always says that the problem is under-governance, not overpopulation. Bangalore can extend in all directions. There’s no shortage of land. If we put together all the top 20 cities in India, the area they occupy is less than 1% of India. We have enough land.
New York in the 1800s, London during the Industrial Revolution — they grew much faster than Bangalore has, constrained by their geography as well. And yet they are some of the great cities of the world. People will always come to areas where there is opportunity. We have to insist on getting better governance instead of saying people are the problem. If you are a democracy, it’s a ridiculous statement to say: there are too many of you, so I can’t provide for you.
The second misconception — and I touched on this while talking about China — is the comparison with Chinese cities. India doesn’t really compete with China. India competes with its 300-plus cities. And what a party representative in a Chinese city has in terms of powers is much more than what a mayor in an Indian city has. There are stats to prove this: 51% of Chinese government expenditure happens below the provincial level. In India, that number is 3%. We are not decentralised. We say we are federal. We are federal in that sense. But federalism and decentralisation are two different things, and we are not truly decentralised.
So we have to understand: Bangalore will always attract more people. Across the world, high-tech centres are always clustered. Even in the U.S., the world’s biggest economy, there’s only one Silicon Valley. The biotech industry is also concentrated around Boston. They’ve tried to have five Silicon Valleys — they can’t. Even though San Francisco is many times a shitty place, things are still around Silicon Valley. Clustering always happens because employers exist here because employees exist here — it’s a virtuous cycle. The idea that we can create five Bengalurus in Karnataka won’t work. We need to find out what Belagavi is good at and enable that. Companies going to Mangaluru won’t happen just because we want it.
So now, coming to the framework for Bangalore — it’s about decentralisation. There are three dimensions: political, administrative, and fiscal. On political and administrative, we’ve tried to solve it half-heartedly through the 73rd and 74th Amendments. What the government finally said is that state governments can devolve certain powers to local governments. That was what the compact was. But state governments are the incumbents here. They don’t want cities to be as powerful — so some states didn’t even devolve tax-raising powers to local governments. There are a whole lot of responsibilities for local governments, but no money-raising powers. How can a government work where you are always dependent on the higher level for anything you do?
We don’t have meaningful, functional mayors — which is very ironic, because all our major leaders of the freedom struggle were mayors of their cities. Jawaharlal Nehru was mayor of Allahabad, Subhas Chandra Bose had a long career in the Kolkata city government. Sardar Vallabhbhai Patel — many of these leaders had a strong local presence and emerged from their cities. Today, do we know the mayors of our cities? All our cities are run by IAS officers. They’re just administrators. We don’t know our councillors because our councils don’t meet. The problem is fundamentally political.
On the administrative side, we’ve fixed things somewhat, but we would actually need a metropolitan service or something of that nature. City governance is a complex issue — you can’t just have anyone become a city administrator. We need a cadre for that. But the biggest problem in all this is the fiscal angle.
When we did the 73rd and 74th Amendments, we did the rural version better. For urban governance, we did it even worse. Rural governments now actually have more financial powers than our local urban governments do. In Bangalore itself, the city budget is roughly 10,000 crore. About 50% of that comes as grants from the state government. How can you plan anything when 50% of your budget is literally dependent on the state’s goodwill? The economist Raja Chelliah said: everyone wants decentralisation, but only up to their own level. No one wants decentralisation below that. That’s where our cities are stuck — because our state governments won’t give them the powers, won’t give them the fiscal powers.
And with GST, another thing happened: city governments don’t even have the revenue-raising powers that existed earlier. What you actually need is a compact — if there is 18% GST, it should be 8-8-2: eight for CGST, eight for SGST, and 2% for local government. You don’t increase the GST rate, but you ensure a fixed, definite sum of money coming to the city so it can plan for itself. Then you can start thinking: we know what our fiscal situation will be for five years, we can take up long-term projects. That is not happening.
The amount of civil society engagement in Bangalore is amazing. The last 15-20 years, we’ve seen a lot of philanthropy and activism. We have great institutions coming in outside the government. But some of that activism is going into substituting the government — doing things that the government is supposed to do. You can substitute the government for a while, but ultimately you can’t. You have to complement the government. We need to put our energy into making our city government more competent.
For example, Church Street is the best street in Bangalore — our office is there, so we obviously love it. Earlier, under the TenderSURE project, its management was offloaded to an agency outside the government. Maintenance was given to people outside BBMP. It worked well — the substitution worked well. But then when that contract ended, we had huge amounts of rubble on Church Street because the government said: it’s not my problem, it’s someone else who was allocated this. It’s now finally been resolved. That’s what I mean. You have to complement the government, not substitute it. We need to make BBMP more competent — or GBA now. We need to make all the agencies more competent, rather than saying: you don’t do the work, we’ll do it for you. The government is happy to offload things to you because they have the capability problem — and it becomes someone else’s problem.
Pranav Agarwal: Thanks a lot. I wanted to ask some questions that are less about the world and more about how you think about things. I think the reason we did this episode with you was that you have a lot of very interesting ideas about a very wide area of things. At the Daily Brief, we are also trying to develop interests in that fashion — trying to understand very large parts of the world. And I think the problem we constantly run into is that you can understand a very significant chunk of a big problem, and the few things you don’t understand might make your conclusions entirely wrong. We are almost never sure if we’ve actually nailed the problem, or if we’re just being overconfident.
How do you resolve that in your head, given that you’re looking at so many very different things?
Pranay Kotasthane: Yes, that is a very good question. There’s an arrogance of ignorance — if you are ignorant, you can be arrogant. But once you know many things, you also find the limits of your knowledge. The more you discover you don’t have all the answers, the more you recognize the factors you haven’t considered. So in that sense, I think you just have to be humble. You say: to the extent that I know, this is my view. If I get new data, I will update my priors. Bayesian updation of your thoughts, rather than taking a stance that ‘this is correct and I will stay with it.’
For example, I have put a view that China’s geopolitical leverage over rare earths will go away by 2028. If that doesn’t happen, I will update my priors and say I was wrong. Maybe I didn’t consider some factor. But it requires you to still say: based on my analysis, this is how I think, these are the lenses I have brought to this problem. You get views from others, you try to see if you have to update your Bayesian priors. If not, you stick with it. When the empirical situation on the ground changes, you try to update.
I think it’s a great skill in public policy to have a lot of breadth, because it’s a complex adaptive system. You need to have a good understanding of some core disciplines and their key frameworks. International relations, for example — how different it is from domestic affairs. You can’t apply the Arthashastra to domestic affairs; there’s no Matsya Nyaya there. But many people try to apply those things. So you need to have that cognitive flexibility: different areas have different ideas and frames. Public finance is one such field, social psychology is one such field, public policy itself — how governments work — is one such field. Arm yourself with key frameworks from all these, and when you have a problem, try to bring them all together.
Knowledge generation happens at the intersection of fields, not within one field. If you just look at it from a sociological lens, you’ll never get the entire frame — you need to bring frames from different streams. Knowing fully well that it’ll never be complete. But you don’t need to be a hundred percent complete. You can always update your view. I think Ambedkar had quoted Emerson saying that consistency is the virtue of an ass — and I don’t want to be one.
Pranav Agarwal: The second thing I want to ask you — one of the reasons we have been doing everything we’re doing, or at least the reason we started, is that we did not actually know what to read to access depth about what is currently happening in India. You would eventually find one good book that explained one thing that happened ten years ago, but very rarely would you find intelligent commentary on what is currently happening. I think one of the sources we look at is a lot of what you write. You are constantly thinking very deeply about very high-level problems that are currently going on in the world. For that, I’m sure you have some information diet where you are actually capturing all of those ideas. What does your information diet look like?
Pranay Kotasthane: Yeah, so I still follow old RSS feed readers. Feedbin is one important source for me. I’m subscribed to many newsletters and things on Feedbin. I try to check it once every day, and I have made tags — there’s a ‘morning brief’ one, which is only news items, and a ‘deep bench’ one for deeper studies. Then there are subject-specific Google Alerts that I’ve subscribed to via RSS — for semiconductor geopolitics, for example. That’s how I try to follow things. In the AI world, it might be even better, but I’m still trying to improve my workflow there.
Books are really important for me because I do a podcast as well. I have to read books — that commitment is built in because you can’t be a fool in front of your guest. You need to know the book to ask good questions. We’re a think tank, we write long-form works, so we have to read long-form work. Essentially it’s about putting yourself in situations where you will have to get these things done.
Newsletters — yours are much better than me at this — because you’re doing it on a daily basis. You have to get it out. You can’t let yourself wander too much. And writing is a form of thinking, not just reading. If you start from the point that you have to read, you’ll go in many places — which is good — but writing will make you think in a focused manner and read with an objective in mind. Writing really helps. The newsletter format and frequency helps.
Being at Takshashila, I get to interact with people from so many different disciplines — people from the armed forces, people retired from the armed forces, people from your world of lawyers, people from different disciplines all together. So you are always exposed to new ideas and new knowledge. That’s how I pick things up.
I also try to have a good note-taking habit. I use Obsidian and have a large store of notes — they could be just one line or one idea. But I try to build on that, and it becomes my input to my writing. When I’m writing, I go back to my notes to see if there was any interesting thought I had captured, and then I try to build on top of that. One newspaper, once a day — I go through it quickly.
And podcasts are a big source — we are in Bangalore, I have almost a two-hour total commute in a day. I’m a big podcast consumer. I have a very strong list, I listen daily, and they have been really helpful in building my knowledge base.
Pranav Agarwal: Thank you so much for being here, Pranay. It was an absolutely lovely conversation.
Pranay Kotasthane: Thanks, Pranav. All the best for the Daily Brief and all the cool things you’re up to.