In today's FirstScroll, we break down why more than $250 billion pouring into Indian data centres might add almost nothing to the country's growth, and why the answer is the same reason the AI boom is quietly an import boom.
With that out of the way, let's dive into today's story.
The Story
If you have been reading the headlines, India is in the middle of a data-centre gold rush. Google and Adani announced a $15 billion project in Visakhapatnam. Reliance, Tata, and a parade of global tech giants have joined in. Add it all up and India has attracted announced investments of more than $250 billion in the buildings that will house the AI age.
A quarter of a trillion dollars is the kind of number that sounds like it must reshape an economy. So it is a genuine surprise to read what one of the world's big ratings agencies thinks it will actually do for India's growth.
Moody's ran the numbers, and its estimate is startling. All that investment, it reckons, will add roughly 0.13% to India's GDP by 2030. Not 13%. Zero point one three. A rounding error on the size of the Indian economy.
So the question is, how can a quarter of a trillion dollars of investment translate into almost no measurable growth?
The answer is one of the most important and least understood ideas in economics, and it explains far more than data centres. It is the difference between money that is spent in a country and value that is captured by it.
Here is the trap in that $250 billion figure. When you hear it, you naturally picture the money staying in India, paying Indian workers, buying Indian goods, circulating through the Indian economy. But a data centre is a very particular kind of project. Strip away the building and the land, and most of the cost is the expensive stuff inside: the servers, the semiconductors, the specialised cooling systems, the high-end networking gear. And India makes almost none of that. It buys it from abroad.
So when a company "invests $15 billion in an Indian data centre," a large chunk of that money does not stay in India at all. It flows straight back out to the American, Taiwanese and Korean firms that make the chips and servers. The rupees touch down in India just long enough to be converted and sent overseas to pay for imports. What stays behind is the concrete shell, the land, and the electricity bill.
This is the gap between spending and value-added, and it is what GDP actually measures. GDP does not count money that passes through a country on its way somewhere else. It counts the value a country itself adds. And a project whose most valuable components are all imported adds surprisingly little of that value locally, no matter how enormous the headline cheque.
Now, why should you care about a ratings agency's growth arithmetic? Because it punctures one of the most common ways big investment numbers are used to dazzle. Politicians and press releases love a giant "investment announced" figure, because it sounds like jobs and prosperity landing on the doorstep. But the number that matters is not how much is spent. It is how much of that spending sticks, how much is made here rather than merely assembled here. Learn to ask "how much of this stays in the country?" and a lot of triumphant announcements shrink to their real size.
The jobs picture tells the same story. Data centres are capital-intensive, not labour-intensive, meaning they are giant halls of humming machines that need very few humans to run. Moody's expects the entire boom to lift employment by around 0.02%, with the permanent jobs concentrated in a handful of specialised technical roles. A $250 billion wave that barely moves the jobs needle is not the mass-employment engine the country actually needs.
But here's the twist, and this is where two FirstScroll stories collide into one. We wrote recently about how Tata is playing the smart hand in the AI gold rush by selling the "picks and shovels," the buildings and power, rather than mining the AI itself. That is a great trade for Tata. But zoom out to the whole nation, and India is in a subtly weaker version of that position. It is hosting the picks and shovels, but it is not making them. It is providing the land and the electricity, the lowest-value parts of the chain, while importing the high-value chips and servers that the whole thing actually runs on. The country is renting out its floor space to the AI revolution, not manufacturing its tools.
And this connects to an even older FirstScroll thread. India's great economic strength has been selling services, brainpower delivered down a wire, while never quite building the manufacturing muscle to make physical high-tech goods. The data-centre boom lands right on that fault line. The demand for computing is exploding, but the value flows to whoever makes the machines, and India, for now, is a buyer, not a builder.
Now, none of this makes the data centres pointless. Moody's is careful to say they are strategically and locally important, just not big enough yet to move the national growth needle. Having your data stored on Indian soil matters for security and sovereignty. And the real prize is what could grow around these centres. If the boom eventually pulls chipmakers and server assembly onto Indian soil, if it feeds a wider ecosystem of cloud services and digital exports, the value captured could climb sharply. The 0.13% is a snapshot of today's shallow footprint, not a life sentence.
That is the actual stakes buried in a dry GDP estimate. The buildings are the easy part, and India is winning those. The question is whether it can climb from hosting the machines to making them, because that is where the money, and the jobs, truly live.
So a quarter-trillion-dollar boom adding a rounding error to growth is not a contradiction. It is a precise measure of how much of the AI economy India currently owns, versus how much it merely houses. The investment is real. The value, for now, mostly lands somewhere else.
Whether India turns its floor space into a factory floor is the difference between renting the future and building it.
Until then…
If this changed how you read a giant "investment announced" headline, share it with a friend on WhatsApp, LinkedIn or X.
