Why India cannot grow at par with China or the United States until economic activity stops sitting in ten cities.

The India we talk about and the India that exists

Having lived in India and travelled across almost every state, I keep coming back to the same observation.There is a stark difference between the India we talk about and the India that actually exists on the ground. And I believe this is the reason we are not able to grow at par with China. Until we change it, we never will, whether the comparison is China or the United States.

Leave a handful of cities. Mumbai, Pune, Bangalore, Chennai, Hyderabad and Delhi etc . Beyond these, wherever you go, it is empty. There is no business. There is no structure. There is no technology. Whatever we see as India is sitting on a very small part of India. That is not just an impression from travelling. It is measurable. Five states, Maharashtra, Tamil Nadu, Uttar Pradesh, Karnataka and Gujarat, together produced close to 48 per cent of India's GDP in FY2025, while the bottom ten states accounted for under 3 per cent. Maharashtra alone contributed 13.3 per cent.

Follow the capital and the picture is sharper still. Of roughly ₹4.22 lakh crore in FDI equity inflows in FY2025, five states captured 83.3 per cent. So output is concentrated, and the money coming in to create future output is even more concentrated than the output itself.

The load falls on the same few places

The consequence is that the entire load of the country falls on that small part. The entire load of tax. A very large chunk of what the state spends goes into running these same few places, because that is where the population keeps moving. That is where the housing, the roads, the water, the transport and the healthcare have to be built again and again, in the most expensive land in the country. And the regions people are leaving never build the tax base to fund anything of their own. So the country pays for this twice. Once at the destination, in the cost of congestion. Once at the origin, in the cost of stagnation. Neither cost shows up as a line item anywhere, but both are real, and both compound.

Why South Korea is the wrong comparison

This is not a problem unique to India. South Korea is in a far worse position on concentration than we are. Roughly half its population sits in and around Seoul. But I do not think we should be comparing ourselves with South Korea. South Korea became rich while being concentrated, and someone will use that against this argument. The honest answer is scale. Concentration is survivable in a country of fifty million people. It is not survivable in a country of 1.4 billion. What one metropolitan region can absorb in Korea is a structural impossibility here. We should be comparing ourselves with the United States and China.

What multi-engine economies actually look like

Look at what those countries have. They have multiple state-level economies, each enormous on its own. California alone is above four trillion dollars, larger than most countries in the world. But that is not the interesting part. The interesting part is that Texas at roughly $2.9 trillion and New York at roughly $2.47 trillion are themselves larger than most national economies, and six American states now clear a trillion dollars each. China has done something comparable, building several industrial and commercial centres rather than concentrating everything in one or two. That is the point. Large economies run on many engines. India is running on about five. And this is not an argument for copying either country. It is an argument about structure. You cannot become a ten or fifteen trillion dollar economy while an overwhelming share of high productivity activity stays inside the same ten or twenty cities.

The counterargument, stated honestly

There is a serious argument against everything I have said so far, and it deserves to be put properly. Concentration is not a policy failure. It is agglomeration economics working exactly as it should. Firms cluster because clustering lowers the cost of finding workers, suppliers, customers, capital and specialised services. Productivity inside a dense cluster is genuinely higher than the same activity spread thin. On that reading, Mumbai and Bangalore are not hoarding growth. They are producing it. There is also an uncomfortable historical record. Attempts to engineer industrial dispersal by administrative means have gone badly before, in India and elsewhere. Our own freight equalisation policy, which ran from the early 1950s until 1993, is widely argued to have neutralised the natural advantage of the eastern mineral belt. Backward area subsidies have often produced paper units rather than durable industry.

I accept both points. So my argument is not that activity should be pushed out of the successful cities. Nothing about this requires Mumbai or Bangalore to shrink. It requires more places to start existing economically at all.

We are asking too few people to carry too much

There is another part of this that we do not say out loud often enough. As Indians, we are putting enormous pressure on the wealthy and the formal sector to keep the economy flowing. A very small number of people and a very small number of formal businesses are being asked to carry the taxation, the investment and the employment for everybody else.

And because money, power and opportunity are concentrated in the hands of so few, India is not growing properly and throughout. It is growing only at the top. Meanwhile, across large parts of Bihar, Uttar Pradesh and Uttarakhand, you have enormous numbers of people earning very little. India's average per capita income is around ₹2.58 lakh, and only sixteen states clear that bar. Bihar sits at the bottom, at roughly ₹69,321. Uttar Pradesh and Bihar together account for more than a quarter of the country's population. These are not people who lack ability. They lack anything around them to be productive inside of. The answer is not to squeeze the successful harder. We need more successful businesses and more wealthy individuals, not fewer. The answer is to increase how many people are capable of being economically productive in the first place. Which means putting three things into hands that currently do not have them. Capital. Technology. Knowledge. And doing it at scale.

The UPI lesson, and what it actually was

This brings me to what I believe the mechanism is. Look at UPI . That is the whole case study. Nothing about it was complicated for the user. A small, useful piece of technology reached everybody. A vegetable seller could accept a digital payment. A customer could pay in seconds. A tiny business could receive money without any banking infrastructure of its own. And the benefit was humongous, for everybody, not just for people who already had cards and bank accounts.

The scale is not in dispute. UPI processed 241.62 billion transactions worth about ₹314 lakh crore in FY2025-26. Banks live on the system went from 44 in its first year to over 700. India now accounts for roughly half of global real-time payment volume. But it is worth being precise about why it worked, because the usual explanation leads to bad conclusions.

UPI did not succeed because somebody built a good app. It succeeded because of a stack of deliberate decisions underneath the app: an identity layer, near universal bank accounts, mandated interoperability so that no single company could close the network around itself, a pricing decision that removed the per transaction cost of participating, and a regulator willing to enforce all of it. Rails. Interoperability. Zero cost to enter. That is the transferable lesson. Not "make the technology simple." Simple was the outcome, not the cause.

Now apply it to AI

Today AI sits mostly with large companies, technology professionals and businesses that already run sophisticated systems. But imagine a small manufacturer in Bihar using it to look at his inventory, find his inefficiencies, prepare his accounts and forecast his demand. Imagine a small business owner in Uttar Pradesh getting the kind of financial analysis that today requires hiring a consultant he cannot afford. Imagine a local entrepreneur building a website, understanding his own compliance obligations and looking properly at his competition. Imagine a farmer using technology to decide on price, timing and logistics, not merely to read information. At that point AI stops being a productivity tool for the existing economy. It becomes an economic infrastructure layer, the way UPI became infrastructure for payments. If we can productise AI and take it out to everybody in a genuinely useful way, alongside capital and knowledge, you could be looking at an India very different from the India we see today. But the UPI comparison cuts both ways, and it is worth being clear about what has to be true.

Power and connectivity. Reliable electricity and bandwidth in exactly the districts where the productivity gap is largest.

Credit. Analysis that identifies an opportunity is worthless to a business that cannot finance acting on it. Formal credit access is itself concentrated.

Language and interface. Not translation added at the end, but tools built for users who will never read documentation in English.

Trust and recourse. A small business will not act on advice it cannot verify and cannot appeal. Somebody has to be accountable when the output is wrong.

Miss those four and AI will widen the gap rather than close it, because the firms best placed to adopt it are the ones that already have all four.

What would prove this wrong

An argument that cannot fail is not worth much. This one fails if, over the next decade, the GDP share of the top five states falls meaningfully and the bottom ten rise, without any deliberate distribution effort, purely on market forces and infrastructure already committed. The early signals to watch are whether FDI dispersion broadens beyond its current five recipients, and whether mid sized states start gaining share. Uttar Pradesh, compounding at roughly 15.3 per cent nominal over five years, is the case to watch most closely, in either direction.

Where this leads

India does not have a shortage of people. It has a shortage of distributed economic opportunity. If capital, technology, education and entrepreneurship can reach smaller cities and districts, we start building hundreds of economic centres instead of continually expanding the same twenty. That changes the structure of the economy, not just its size. An India where opportunity is not decided by how close you live to Mumbai, Bangalore or Delhi. An India where a young person does not have to leave Bihar, Uttar Pradesh or Uttarakhand simply to find something productive to do. An India where the growth is not just larger, but broader and deeper. That, to me, is the real challenge and the real opportunity of the next 25 years.

Sources: Client Associates, State of Indian States: 2026; National Payments Corporation of India, UPI product statistics and decade report, FY2025-26; US Bureau of Economic Analysis, state GDP 2025; Public Policy Institute of California.