Enjoyed reading this. I think you're right to point out the premature imitation problem but may be misapplying it to concerns about the environmental costs of AI, which seem far more salient to a nation that regularly faces shortages of water and energy than they are to Virginia. Instead, isn't there premature imitation in the rush to find uses for the shiny new tool before attempting to grapple with the mundane challenges of execution you identify? Won't those same problems of execution plague the adoption of AI as well and undermine any attempts to leapfrog?
This is profoundly stupid. I'm sorry for being so blunt, but this fake narrative haks me at my wits end. The myth that Data centers consume water has been debunked 100s of times by now, and it will not die. You want to save water, then curtail agriculture. Two outlets of McDonald's in US use more water than a data centre. Even the energy use, is not a global problem, but a local problem. At a national level, data centres consume a measly 4% of USA's electricity supply.
Data centres can raise local electricity bills, and it will not be a good thing if that energy comes from fossil fuels. You don't fix that by stopping data centres, which are soon going to be a national security asset, but by increasing clean energy supply.
And if we always want fix "execution" before adopting a new technology, you would never adopt anything. It's a degrowth mindset, which is idiotic, given we are a piss poor country with only $3k GDP per capita. We adopted Digital India before we fixed supply of piped water to villages. What followed was UPI, a global achievement which works on a scale most countries can't fathom.
When the ground is shifting beneath everyone's feet is exactly when you have to take a gamble. If the US populace is losing their minds over imagined concerns like water use of Data centers, it is precisely a moment where young country which is in desperate need of growth, capitalises on.
These arguments assume that we have time. We don't. We have 20 years before we lose our demographic advantage and have to contend with an aging population and low fertility, a problem which no country on the planet has figured out a solution to. We must get to the equivalent of 20-25 trillion USD GDP in today's terms before then.
I've read Masley. Your comment seems to itself be a form of premature imitation. Nothing Masley cites should dismiss concerns about energy and water usage by data centers at a local level. Comparing water usage by data centers to that of U.S. golf courses has limited applicability in a context where energy and water infrastructure is already strained beyond its limits, and cannot adequately serve the needs of the local population. If any of the announced $200B investment was aimed at upgrading this basic infrastructure to better serve Indians in addition to enabling the operation of new data centers, this might be a different conversation. But India's VIP culture and the execution problems Vedica highlights portend an inequitable distribution of any such investments.
I'm also still waiting on evidence for how constructing data centers will help India make use of and better serve the largely underemployed demographic dividend you speak of. Unlike smartphone factories, data centers don't need to employ large numbers of people to operate. By contrast, in the short term, AI capabilities seem well-equipped to undercut the kind of BPO work that has employed many young people in the country.
We don't need golf courses, because we have paddy farmers, whose produce the government is forced to buy at MSP. They cause more environmental destruction, and are among the most priveleged part of policy class, while providing measly economic returns. Neither you, nor anyone else would say 2 words against them, because that would be political suicide.
A lot of data centers were announced as partnerships with energy incumbents, because tech companies aren't stupid. We added around 52 GW of energy capacity. 48 GW of that was renewable energy.
Around 13% of the capacity in 2024 was installed in 2025 alone, over 95% of the newly added capacity was renewable. The capacity addition was so high, that DISCOMs have asked to slow down, so that batteries can be integrated, and the grid is kept stable. This is not enough, and more will be needed, but we are on the right track.
No one has the answer to your second question. All that is known is you create new companies on the basis of Al, creating new products and services.
Because you cannot put the genie back in the box.
You adapt, or you are automated away. If India doesn't do it, someone else will. It has already changed software engineering completely, and the coming 2/3 years are going to be even more disruptive.
That is to say nothing about military applications which have gotten so good that DoD and Anthropic are having a public spat. You cannot outshore critical assets, you will pay the price for it.
"I'm also still waiting on evidence for how constructing data centers will help India make use of and better serve the largely underemployed demographic dividend you speak of.." I'm not sure that's necessarily the angle I would take to look at it. DC's might not employ huge nubers but there is not doubt that they are a core infrastructure that we require. And burying our head in the sand with a view that BPOs will no longer employ people isn't exactly the best approach either right -- that would happen (if it were to happen regardless of whether we build DCs or not? (it's a seperate issue that we are also trying to move up from BPOs to GCCs)
The gap between ambition and execution shows up in small details too, like navigation and basic planning. That says a lot about where the friction really is. How do you think India can build strong research institutions that produce original AI work, rather than relying on startups and private players to carry that load?
Enjoyed reading this. I think you're right to point out the premature imitation problem but may be misapplying it to concerns about the environmental costs of AI, which seem far more salient to a nation that regularly faces shortages of water and energy than they are to Virginia. Instead, isn't there premature imitation in the rush to find uses for the shiny new tool before attempting to grapple with the mundane challenges of execution you identify? Won't those same problems of execution plague the adoption of AI as well and undermine any attempts to leapfrog?
This is profoundly stupid. I'm sorry for being so blunt, but this fake narrative haks me at my wits end. The myth that Data centers consume water has been debunked 100s of times by now, and it will not die. You want to save water, then curtail agriculture. Two outlets of McDonald's in US use more water than a data centre. Even the energy use, is not a global problem, but a local problem. At a national level, data centres consume a measly 4% of USA's electricity supply.
Data centres can raise local electricity bills, and it will not be a good thing if that energy comes from fossil fuels. You don't fix that by stopping data centres, which are soon going to be a national security asset, but by increasing clean energy supply.
And if we always want fix "execution" before adopting a new technology, you would never adopt anything. It's a degrowth mindset, which is idiotic, given we are a piss poor country with only $3k GDP per capita. We adopted Digital India before we fixed supply of piped water to villages. What followed was UPI, a global achievement which works on a scale most countries can't fathom.
When the ground is shifting beneath everyone's feet is exactly when you have to take a gamble. If the US populace is losing their minds over imagined concerns like water use of Data centers, it is precisely a moment where young country which is in desperate need of growth, capitalises on.
These arguments assume that we have time. We don't. We have 20 years before we lose our demographic advantage and have to contend with an aging population and low fertility, a problem which no country on the planet has figured out a solution to. We must get to the equivalent of 20-25 trillion USD GDP in today's terms before then.
A short essay debunking environmental costs of data centres (Has links to longer posts with detailed arguments) - https://andymasley.substack.com/p/a-short-summary-of-my-argument-that
I've read Masley. Your comment seems to itself be a form of premature imitation. Nothing Masley cites should dismiss concerns about energy and water usage by data centers at a local level. Comparing water usage by data centers to that of U.S. golf courses has limited applicability in a context where energy and water infrastructure is already strained beyond its limits, and cannot adequately serve the needs of the local population. If any of the announced $200B investment was aimed at upgrading this basic infrastructure to better serve Indians in addition to enabling the operation of new data centers, this might be a different conversation. But India's VIP culture and the execution problems Vedica highlights portend an inequitable distribution of any such investments.
I'm also still waiting on evidence for how constructing data centers will help India make use of and better serve the largely underemployed demographic dividend you speak of. Unlike smartphone factories, data centers don't need to employ large numbers of people to operate. By contrast, in the short term, AI capabilities seem well-equipped to undercut the kind of BPO work that has employed many young people in the country.
Also data centres are built, operated and maintained by Humans. Discounting it completely is arguing in bad faith.
We don't need golf courses, because we have paddy farmers, whose produce the government is forced to buy at MSP. They cause more environmental destruction, and are among the most priveleged part of policy class, while providing measly economic returns. Neither you, nor anyone else would say 2 words against them, because that would be political suicide.
A lot of data centers were announced as partnerships with energy incumbents, because tech companies aren't stupid. We added around 52 GW of energy capacity. 48 GW of that was renewable energy.
Around 13% of the capacity in 2024 was installed in 2025 alone, over 95% of the newly added capacity was renewable. The capacity addition was so high, that DISCOMs have asked to slow down, so that batteries can be integrated, and the grid is kept stable. This is not enough, and more will be needed, but we are on the right track.
No one has the answer to your second question. All that is known is you create new companies on the basis of Al, creating new products and services.
Because you cannot put the genie back in the box.
You adapt, or you are automated away. If India doesn't do it, someone else will. It has already changed software engineering completely, and the coming 2/3 years are going to be even more disruptive.
That is to say nothing about military applications which have gotten so good that DoD and Anthropic are having a public spat. You cannot outshore critical assets, you will pay the price for it.
"I'm also still waiting on evidence for how constructing data centers will help India make use of and better serve the largely underemployed demographic dividend you speak of.." I'm not sure that's necessarily the angle I would take to look at it. DC's might not employ huge nubers but there is not doubt that they are a core infrastructure that we require. And burying our head in the sand with a view that BPOs will no longer employ people isn't exactly the best approach either right -- that would happen (if it were to happen regardless of whether we build DCs or not? (it's a seperate issue that we are also trying to move up from BPOs to GCCs)
The gap between ambition and execution shows up in small details too, like navigation and basic planning. That says a lot about where the friction really is. How do you think India can build strong research institutions that produce original AI work, rather than relying on startups and private players to carry that load?
Very insightful.