Nandan Nilekani | On Building Digital India, and the Next Population-Scale Opportunity

You should first bulletproof the idea in your head >> correct?

>> Before, you, bulletproof, it, in, the, real world. If you have focus, determination

world. If you have focus, determination long-term marathon running kind of thing, then you can bring change. We had

about 10,000 startups 10 years back.

Today we have 150,000 startups. So I

think it'll go to a million by 2035. All

the ingredients are in place because I think capital is there and more capital will come. Every time a internet company

will come. Every time a internet company goes public, it creates 50 new entrepreneurs. So you imagine that

entrepreneurs. So you imagine that happening at a scale. the Aadhaar

project. How did the scale of that not daunt you?

>> So,, first, of, all,, the, scale, was, crazy.

Giving 1.4 billion people an ID is non-trivial problem. I publicly said in

non-trivial problem. I publicly said in August of 2009 that we would issue 600 million IDs in 5 years, which nobody goes around in government saying I'll do something. So, I said I'll put my stake

something. So, I said I'll put my stake in the ground. When we design digital public infrastructure, different from a market. See, in the private sector, you

market. See, in the private sector, you focus on a market segment. When you're

designing public infrastructure, it's how do you get up and 4 billion people in the game? How do you level it out so that they're all leveled out and all have the same access? We had 1200 people

from 110 countries who want to implement digital public infrastructure. A lot of focus should be actually on applying AI to real world problems and that's where India has an edge because we have solved a lot of population scale real world

problems. So we do more of that with AI.

Abundance thing going to happen.

There'll be enough of everything. So you

can't derive purpose from that. But they

have you to find something else which gives you purpose. So I think go for purpose.

Nan welcome and thank you for doing this.

>> Sure., Um, you, know, Nand, what, I'll, start off by saying is that you know when as Ruchi and I talk about who uh deeply

inspires us and who we would like to model you know our long-term uh careers and lives after like your name is often one that comes up and it's because you

know you've retained that sense of curiosity uh energy enthusiasm up until your young old age you know so um it's remarkable so I guess my opening question was going to be what motivates

you?

>> No,, I, think, I'm, extremely, curious, and, uh I always want to learn new things, keep on top of things, be relevant uh and

leverage my, you know, I have a network of people I know now. I have a network of information I have and if I bring it all together, I can have great impact.

So, that's what keeps me going.

>> Amazing., So, I, mean, you, know, a, lot, of, I love that you start with curiosity because, one, of the, things, we, say, at, SPC is that when you're in that minus one phase kind of you know it's a what do

you tell someone the what is a common question you ask it's so easy to ask what's your startup idea what are you working on but we tend to reframe the questions internally as in the minus one

days you should ask what are you curious about like what is a thread of curiosity that you're pulling on so maybe to extend that then How do you go from that curiosity? How do you actually pick

curiosity? How do you actually pick things like you know you have worked on so many things?

>> Yeah., So, I, uh, I, mean, I, operate, in, four quadrants.

>> Okay.

>> I, operate, as, a, doer, in, business, and, also think about business strategy and write about. I'm also a doer in the public

about. I'm also a doer in the public space and a thinker in the public space.

So, so I operate in that's why I'm able to I have sort of much wider view of many things and uh so I look for things that have high impact

>> okay >> very, potentially, very, high, impact, which have which is uh playing the long game in the sense the impact may not be felt

immediately it may not be obvious to other people that this is impactful but I know that I have the conviction that I have mental road maps that say that if I persist in this for a few years, I know

that there will be a good outcome. So

that's you have to believe in that. It's

it's conviction from processing a lot of information and concluding that this is a desirable, plausible and achievable goal. So once I sort of it meets all

goal. So once I sort of it meets all these tests >> then, I, go, all, in.

>> Yeah,, I, think, that, the, ability, to, kind of be on the island of your own self-conviction for a while is very important, right? and to think long

important, right? and to think long term. Um, but maybe to get a little bit

term. Um, but maybe to get a little bit deeper into it, you know, what would you say? Are you a meticulous planner? Do

say? Are you a meticulous planner? Do

you kind of go and think a lot or are you more of you can just do things and just start doing something?

>> No,, no., I, think, I, spend, a, lot, of, time thinking about the plausibility of doing something in the sense before I plunge into it, I want to have a clear thought

process that there is I know there is a way to get to the end game. Once I have that map in my head then I so I spend a lot of time before I plunge into it I do that

>> I'll, give, an, example, uh, this, when, I, did the adha project >> that, was, going, to, be, my, next, question >> so, I'll, I'll, use, that, I'll, give, a practical example so what happened was

that uh I was uh actually asked whether I wanted to be the HR minister of India so I said why not let me go and fix education so then they said what else would you like to do I said okay let me

give you I knew that time They were they had this ID project which said give everyone an ID. But before I said yes to that I I had a bunch of friends

technology guys, everybody and said is this doable? Before I go and commit to

this doable? Before I go and commit to some doing something I want to make sure that there is a path to this and we spent quite a few days and weekend just thinking through how will this work? How

what are the risks? What are the pitfalls? you know what what the what

pitfalls? you know what what the what are the technology issues, political issues and so when I actually went to say I'll do this project I had a fairly good idea in my head that I know how I would get there. So I think generally I

before I accept something or plunge into something I I do that check.

>> You, know, I, swear, I, didn't, set, this, up.

It sounds like you had your own mini SBC coming up with that card because a lot of what we say is that before you start building just take some time surround yourself with smart people and just keep

on poking at the idea >> like, even, those, initial, conversations will you know iron sharpens iron right?

So like if you after 3 months of vetting something you are still really into it then it has to be a stronger idea.

>> Yeah., I, think, the, idea, you, should, first bulletproof the idea in your head >> correct >> before, you, bulletproof, it, in, the, real world.

>> So, maybe, to, go, from, there., So, you, have you know you've kind of gone from you found an idea that you think kind of meets your criteria of being long-term impact large scale impact perhaps

something that is not obvious to everyone in the world today. you've come

up, after, that, something, has, met met, that criteria come up with a mental map kind of like okay these are the chess pieces I need to play what's the next step you often take after that is they finding a team well often these ideas are not

necessarily started by me okay because somebody comes to me with a great idea and then he he he or she obviously wants to do it and then after then I go through this mental process and then

when I'm convinced they I think this guy is on to something big then I will fully back uh that person. So these days I only do that. I don't do anything myself. I just figure out who the right

myself. I just figure out who the right guys to back, right men and women to back who have big ideas. Nowadays, you

know, I'm going to get you a cursor account. Everybody's writing code

account. Everybody's writing code nowadays, you know. [laughter]

>> Um, you, know,, one, of, the, the, quotes, that you had said in one of your previous interviews that really resonated with me was you said that all you had was one page. You kind of started with one page

page. You kind of started with one page >> in, the, Aadhaar, project., the, Aadhaar project and I would love to spend a little bit more time talk us through some of the early days now it's obviously it's such a huge success it's

being emulated by different countries but >> how, did, the, scale, of, that, not, daunt, you like you know it's kind of a >> yeah, so, first, of, all, the, scale, was, crazy you know giving 1.4 four billion people

in ID non-trivial problem but we had that as said we did all this debating confabulating arguing and we finally said okay there is a path to success

there's a road to success so that's when I took it up a few things one it had to be done at scale and speed because I was clear that we should do

this in a way that it reaches substantial size in one government term You know, so that's why I publicly said in July, August of 2009 that we would

issue 600 million IDs in 5 years, which nobody goes around in government saying I'll do something. So I said I'll put it up. I'll put my stake in the ground. Now

up. I'll put my stake in the ground. Now

to give 600 million IDs in 5 years, I had to build a system which issued 1 and a half million a day. and to do one and a half million IDs a day and assuming one enrollment station has 50

enrollments that's 35,000 enrollment stations had to be live simultaneously across every nook and corner of India so the scale was like daunting but we

thought through the whole thing we built an architecture for the scale we we built a way to remotely understand what was happening and this is long before you know you had connectivity and all

that so we had to build a offline method of ensuring that every data packet collected in every part of the country was, of, a, certain, minimum, quality, and and all those things. So that that was the front end part which is how do you build

a network of 35,000 enrollment stations everybody keys in the data in the same day we had 100,000 operators we had trained you know everything was at that scale so that was the front end part the

backend part was very different was how do we do biometric dduplication at scale so how do you take somebody's biometric data and compare it with all the people you have to see whether it's a duplicate

or not let's say we had 500 million people in our database and we did 1 million new uh enrollments there 500 trillion pattern matches we had to do every night. So the kind of computing

night. So the kind of computing infrastructure we had thousands of servers all you know parallel like doing parallel processing and all that. So the

technology was extremely uh complex and nobody had done this before. Nobody had

solved a problem of unique dduplication of a billion people. So we were on completely uncharted territories but whatever test we did we said okay this is probably doable. So we did that test

also. So I think we had to build this

also. So I think we had to build this whole infrastructure but that was the technology part of it. So that took 14 months to build the version one of this infrastructure.

>> Still, pretty, quick.

>> Yeah., Yeah., We, had, a, team, right, here, in Bangalore on outer ring road which was where I had I did one thing was I kept the political bureaucratic office in

Delhi. I kept the technology office in

Delhi. I kept the technology office in Bangalore. [laughter]

Bangalore. [laughter] I wanted these guys to do their work and not get bothered by stuff there. So we

while this was being built I went and met everybody because anyway I had nothing much to do. So I would go I went and met every state government met the chief minister met the bureaucrats I met

the politicians I met the judges I met lawyers I met activists I met journalists I met the world bank I anybody on the street I would catch and tell them about this Aadhaar stuff. So

because of that 14 month of effort, I was able to create a climate of opinion across all these people and I always went to everybody with what's in it for them >> you, know, to, support, this., So, I, created

good support across the country on this so that when we actually rolled it out I didn't face that much opposition. Yeah

I still think that it's the scale of it is actually just mindboggling. Like even

having worked at large obviously uh companies, you know, whether it be Facebook or Dropbox and a bunch of our portfolio companies, the scale of the Aadhaar card still kind of is mindboggling, right? Um I I think it's

mindboggling, right? Um I I think it's kind of hard to think of a project that has had that level of distributed scale and that needs to work more or less correctly because it's kind of hard to

make mistakes in that system. Every

every packet had to be uh absolutely solid, robust. It was encrypted

solid, robust. It was encrypted >> digitally, signed,, all, kinds, of, things, to ensure that it's not nobody can tamper.

It was tamperroof. Yeah

>> it, had, to, be, transmitted, from, across, the country to our data centers in Bangalore and Delhi. It was a very complex thing.

and Delhi. It was a very complex thing.

But we also had a digital map of this thing since we we could track every packet down to the last where it was what it was doing. So the granularity

and the telemetry was very high.

>> Why, you, know, it's, one, of, the, why, do, you think Silicon Valley is so bad at or perhaps even the American government I'd say why have they really struggled to

kind of build these kind of foundational layers that perhaps you know found I mean let's call it digital public infrastructure like the American government has remarkably little of

that. No, that's I wouldn't I I I I

that. No, that's I wouldn't I I I I don't agree with that completely because >> what, is, the, internet?, Internet, is digital public infrastructure. Internet

was designed >> 40, years, back., It, was, paid, for, by, the, US government. It was funded by DARPA and

government. It was funded by DARPA and others. And basically every piece of the

others. And basically every piece of the internet, the original internet, true you know, and the protocols of the original internet, you know, SMTP protocol HTTP HTML

all was public. So up till about 9495 the entire internet was public infrastructure. It's only after mosaic

infrastructure. It's only after mosaic was built and then and moved to the valley and started Netscape with Jim you know that guy >> Jim, Clark >> Jim, Clark, that, the, private, internet

began and then of course then you had Google and Facebook. So that private internet is only 30 years old.

>> So, it's, there, was, another, 10, years before that the public internet and all the early internet was all public. For

example, the worldwide web was designed at CERN in Switzerland >> and, was, the, protocol, was, put, by, Tim Bernersley as a public protocol.

>> It's, true.

>> You, know,, it, so, it, was, so, that's, one example of DPI. So, and built entirely by US government. That's true.

>> And, the, same, thing, is, true, of, GPS., GPS

was again paid by Department of Defense >> and, it, was, for, missile, triangulation, and all that. But fundamentally

all that. But fundamentally >> they, developed, the, system, to, answer, the question where am I?

>> Mhm., And, it, was, only, this, was, built, in the 80 in Star Wars days, you know, 80s and 90s. It's it was only put for

and 90s. It's it was only put for commercial usage in 2000 by President Clinton and two and then it became commercially available and 6 years later you had Google maps and 9 years later

you had Uber. That's right. So I think both G both GPS and internet are very good examples of digital public infrastructure built by state support

but which are opened up for private innovation. So we we took the same idea

innovation. So we we took the same idea forward. We said that there are pieces

forward. We said that there are pieces of the infrastructure which have to be done at population scale which if done everybody can benefit and take moves the

needle on capability. So it's not an old idea just an idea we took it forward.

>> Maybe, to, move, on, to, the, next, topic, about AI and India. One of the trends definitely is that every country seems to want to have some kind of like sovereignty over AI capability. Now

whether that means it's their own kind of like you know sovereignly trained model or certainly sovereignly trained infrastructure to deploy those models there's some aspect of essentially what

is my sovereign story here. So what do you think that should be for India like >> yeah, I, I, think, the, question, is, where does the sovereignity begin >> correct

>> right, if, you, are, truly, sovereign, in, your AI then everything in the stack has to be yours otherwise it's not sovereign okay okay you may build your model

>> you, may, train, it, on, your, own, data, but, if the underlying chip is somebody else's then it's no longer sovereign so it's I don't know, what, is, sovereign, here >> so, if, you, think, of, it, the, true, truly

sovereign stacks are just the in China and America that's it >> do, you, think, India, should, be, doing, more because you're right like I mean are you going to build GPUs also okay that's hard uh

>> so, where, do, you, stop, you, know, where's the fab you can go back and say the fab in fact nobody has the fab the fab is in Taiwan so >> so, so, I, think, the, uh, uh, I, I, my, own, view

is that the at least for the time being we should build models and there are many people building models serum and all of them >> your, you, have, these, Maya, guys, all building models so I think that's great

and I think it should be done >> but, a, lot, of, focus, should, be, actually, on applying AI to real world problems and that's where India has an edge because we have solved a lot of population scale

real world problems so we do more of that with AI and my view is that over time the lower levels of the stack will get commoditized >> I, think, we, already, seen >> we, already, seen, if, you, look, at, some, of the Chinese models they're amazing you

Every week there's a new one. There's

Deepseek and Kim or whatever it is.

There's of course Quen. So Minia Max, I mean like I and they are putting publishing everything. They're

publishing everything. They're publishing the model, they're publishing the weights, they're publishing the uh all the tricks of the trade, how they did you know mixture of experts. So I

think if this rate of knowledge permeates to the world, you'll commoditize the lower levels of the stack which means that those who build the upper levels of the stack are going to benefit. So we should build those and

to benefit. So we should build those and that's my view. It's a personal view.

>> No,, I, agree, with, this, and, I, think, that AI is a somewhat kind of if you kind of look at the generally how technologies kind of develop and who funds them and who gets first access to them. AI is

somewhat unique in the history of technologies in that it's actually completely you know it's generally available uh democratized and actually like you know commoditized almost from like day one right like it's not just that the military is getting to use it

or not only the uh kind of like the enterprises who can pay a lot of money like you know everybody on everybody in India can actually have access to uh chat GPG now if you have a smartphone and it's also really interesting I think

one of the common critiques I hear sometimes in India is that we don't have access to GPUs we don't have access to the capital resources But I mean a lot of these, Chinese, companies, are, training these models for like a lot lot cheaper.

You don't need billions of dollars to kind of build like these foundational foundation model pre-training runs. So I

do think I mean we'll get to the what you can build higher up in the stack in one second. But I think even in this

one second. But I think even in this realm of training models I think there's a tremendous amount of opportunity here.

>> Oh, yeah, definitely., I, I, think, uh, and, you know and also what's happening is with this mixture of experts kind of thinking you don't need to build one mega model

which is you know zillion trillion parameter kind of thing you can build many small models that are very focused and then bring them all together and orchestrate them to do the thing you want.

>> Yeah., Um, what, would, you, I, mean, you mentioned about if AI is a democratizing and commoditized technology then we can essentially build a lot of things for it

up the stack. So as you and in particular I think India has a it's a very interesting confluence of um obviously a lot of technical essentially like talent but then also a lot of like

you know high-end kind of like you know white collar work that happens here. So

what role do you think we can play obviously to bring AI into those and I'm sure you see that at uh at Infosys but then what also like do you think could we then bring back to the world using

the unique capabilities and perhaps uh knowledge that we have? No, I think uh I think certainly the a lot of the

enterprise AI stuff will happen here because you know Indian companies are already familiar with the business needs of every every major corporation in the

world and all of them are embarking on this journey of AI transformation and it's not it's not magic you don't get up in the morning and some AI transformation happens you have to do a

lot of things a lot of uh stuff on data on responsible AI on you know making sure that the quality of output is good that there's no hallucinations you need

to create an architecture which allows you to use multiple models there lots of issues anyway so I think all that will happen here so I think we will learn how to apply AI at scale and speed to the

world's challenges and that will be a huge opportunity >> what, do, you, think, are, the, opportunities to and you you know you've obviously been involved with AI for Bharat and so on. But where do you think the

on. But where do you think the opportunities are to provide AI as a public good in India?

>> Yeah., So, I, think, there, are, three, four dimensions to this. One is uh providing data for everyone to consume to improve

their AI models. See today the way it's working is everyone is doing the same thing. They're all probably using the

thing. They're all probably using the same data source or they're all going to Reddit or something like that and building their models. So five guys are doing it. But there's a lot of data out

doing it. But there's a lot of data out there which is non-competitive and that data if it can be made available as a digital public good then everybody can use those models. So

that's what we do in AI for Bharat which is we are creating the data for 22 Indian languages as a public good. So we you know we have a whole infrastructure of collecting

samples from across the country. We have

we collected thousands of hours of samples and those samples are then uh put into a database and that database is available to everyone to use. So anyone

who wants to improve the quality of say Indian language capability can use that.

The government has set up this thing called AI kosh which is a repository of open data. So it's in that location. So

open data. So it's in that location. So

you can take that and so if five startups are trying to build language AI they can nuance it any way they want but they can use this as a core uh uh data.

So one way to think of public uh DPI in AI is to have digital public good. Now

this is not just this could be for weather or it could be whatever as long sharable data which everybody has equal access to and then there is the thing of

making it available as APIs. So you can create public APIs, AI APIs for doing different things or agriculture education and so on. So the the

different ways to think about uh uh DPI approach. So that's AI and digital uh

approach. So that's AI and digital uh you know using a public approach and then in digital public infrastructure you can use AI you know. So that's the other, way, around., So, for example,, if, you

look at Aadhaar authentication and Aadhaar does about 80 million Os a day uh you can do a biometric o or whatever all the biometric o is does livveness

detection, using, AI., So, for example, if somebody tries to use a gummy finger you can't it won't go through or if you try to use a photograph so there's a livveness detection that there's actually a proof of personhood kind of

thing. There's a person at the end of

thing. There's a person at the end of that the entire stuff is AI based. Yeah

I really would, you know, obviously as I've been here for the last week, a AI is on everyone's mind and this is kind of all people want to like talk about and I've been trying quite hard to

change the narrative that it's not that India is behind. I actually think we are incredibly well positioned in terms of the AI is a uniquely democratizing

technology, right? Um and I think that

technology, right? Um and I think that the both the level of technology talent we have here and kind of the workflows that we have learned how to do over the last 20 years I think it has some unique

advantages. So I I would really like to

advantages. So I I would really like to shift the narrative that we have behind in AI and actually to like give people the confidence that we have a ton of advantages right like we in fact should be thinking that we are much more well

positioned in some ways yes like just to use China as a counter example yes they're building a bunch of models but a lot of what AI is going to be used for if you look at for instance modern reinforcement learning a lot of it is about workflows what do people actually

want to do how do you create these smaller models these more specialized models and we're in a much better position to do that >> totally, totally, I, I, the, smaller, end, of the model spectrum will be done by everyone and then you'll figure out

you'll you'll then create orchestration layer that take some frontier models some smaller models some open source models it's all about how you then start orchestrating these things >> yeah, and, you, know, I, think that, the, one

area that and you know we were talking a little bit about this earlier but I do think that the level of uh demand for compute related to AI is going to be very large and I think if you're trying

to do low latency stuff. I think we do need to figure out how to colllocate GPUs within the country. But I'm already seeing a ton more data centers kind of come up and a lot of it is being nearshored in terms of GPU capacity.

>> No,, I, think, that'll, all, happen.

>> Capitalism, it'll, happen., Yeah,

>> I, think, yeah,, markets, will, take, care, of that, >> right?, Yeah., Yeah., So, maybe, to, now, shift

>> right?, Yeah., Yeah., So, maybe, to, now, shift gears to just talk about the India startup ecosystem. Um you know one of

startup ecosystem. Um you know one of the things that you had you know you had mentioned before that was actually I think really maps to how I think about the world is like why can't India have a

million startups?

>> They, will, we, will, have, a, million, by 2035.

>> Yeah., So, and, I, think, that's, hugely inspiring because you know there's one strain of thought in Silicon Valley. I

think there's, you know a bunch of smart people I think feel that the number of true entrepreneurs in the world is limited, right? And the goal of venture capital is to try to find the 20 or 30, you know, Elon Musks and kind of

put all money into them. I tend to disagree with that. I think there's a lot more talent that can be unearthed and essentially like nurtured to build startups and like so I think that

there's a lot more like there India should have you know an order of magnitude more startups than they have today and I hope South Park Commons can play a part in building that and I don't think of it as a zero sum game I think

more startups means there's more kind of like areas to fund more opportunities to tackle so this is a long-winded way of saying Nan which is what do you think is the biggest factor holding us back in India right now

>> actually, I, I, think, I'm, fairly, confident that see we had about 10,000 startups 10 years back today we have 150,000 startups so I think it'll go to a

million by 2035 I think the all the ingredients are in place because I think capital is there and more capital will come I think capital

will come by itself I think the uh virtuous cycle of entrepreneurship is happening Because every time a internet company goes

public, it creates 50 new entrepreneurs.

All those guys who made money there then leave and start their own company. So

you imagine that happening like at a scale and I I've studied the data on number of entrepreneurs that came out of some of the early companies like Flipkart and all. It's large.

>> It's, incredible.

>> It's, incredible., So, that, that's, a virtual cycle. Then many of the early

virtual cycle. Then many of the early entrepreneurs who made money are now becoming seed and angel investors. So

there's a virtual cycle of capital. is

the virtual entrepreneurship and I think there are enough things to be solved. So

I think that will take care of itself. I

think the place where we need to focus on which is really more of a government thing is improving ease of business. How

do you reduce friction to business in India because that's has a huge multiplier effect. Imagine if there's

multiplier effect. Imagine if there's millions startups all of them have less friction to business then dramatic thing right. So that's a high leverage uh

right. So that's a high leverage uh activity.

>> Yeah., And, you, know, it's, interesting, I when I talk to when I talk to entrepreneurs here they all of them point out there's been a lot of progress in terms of ease of starting a business but there is a huge gap like you know I

tell people in America like if I wanted to start a company I can have that idea by 1 p.m. by 3 p.m. I will be incorporated have a bank account and I can take money from a VC and spend my first AWS bill. You know that still takes a while in India.

>> Some, of, these, things, are, better., It's

more the day-to-day stuff of running a business where you're dealing with tax authorities. You're dealing with uh you

authorities. You're dealing with uh you know all kinds of government departments. That's where the rubber

departments. That's where the rubber hits the road.

>> I, mean, I, had, to, sign, 70, pages, of documents today just for you know some uh opening an account here. Um let me push on that for a second. So there's

obviously the governments sorry we are seeing a lot of kind of positive inputs into the system. uh virtual cycle of capital, virtual cycle of essentially people who know what good looks like in

terms of companies. This is fantastic.

This is what Silicon Valley runs on, you know, but do you think and then you also like you know the government is making a concerted effort. Fantastic. Are there

concerted effort. Fantastic. Are there

cultural elements that you think we need to change here?

>> No,, I, don't, think, so., I, think, uh, I, think uh the entrepreneur has come front and center in society which is a big deal.

>> That's, Yeah.

>> Yeah., Because, you, know, when, we, started everybody said are you crazy? Go and

take a job somewhere. you know what I trying so so I think from those days and I was just looking at the you know the the the topper at uh IIT

10 years back said I want to be a scientist today's says I want to be Elon Musk so clearly there's a sea change in uh mindsets so I think culturally I

don't see an issue at all I think uh entrepreneurs are now celebrated they're front and center uh and also very importantly I think it's It's okay to fail.

>> Yeah.

>> Now, in, many, so, many, cultures, if, you, fail you know you're like it's a problem. So

then you don't become an entrepreneur you're worried of failing. I see so many guys whose companies have not gone through but they come back in a new avatar very quickly. So that's also a

sign that failure is accepted which I think is a huge part of it.

>> Yeah., The, most, positive, thing, that, we have seen at SPC which was that and this was a kind of a hypothesis on on our part when we started a year ago was that you know to join SPC you kind of have to

uh in some ways admit that you don't know what you want to work on or you are thinking about it and you have to almost drop the facade of uh of knowing everything right so that was my that was

one of my open questions will that model kind of like of like come to SPC and for 6 months kind of wander a little bit try to like you know come up with like ideas stress test them and you might have to in some ways you're implicitly saying I

don't know all the answers right but that requires a certain degree of courage and I would say self-confidence and I was curious about whether it would work in Bangalore and it's been amazing people are willing yeah it's amazing

>> yeah, no, I, every, day, I, meet, I, meet, just for curiosity again you know what's going on right so I meet entrepreneurs every day and uh it's extraordinary the the scale of the ambition the it's quite

remarkable I mean and these are some guy just out of school not even in college you know that kind of thing so so I think, it's, it's it's, a, great, it's, it's, a golden era for entrepreneurship >> yeah, I, mean, we, just, met, Shri, you, know, 23

years old and building airplanes it's pretty incredible uh >> totally >> yeah, do, you, think, that, you, know, what, are there traits of entrepreneurship that

you think are uniquely required to succeed in India I'll give you an example I think anywhere an entrepreneur needs to be resilient they need to be able to roll with the punches is they need to be very ambitious. They need to

be internally self-convicted. They need

to have a long-term view. Do you think some of these traits are more like required in our ecosystem here?

>> Yeah,, these, are, generic, traits., But

also, I think why entrepreneurs are successful is because they can deal with chaos.

>> That's, true.

>> You, know,, it's, just, getting, to, office, is chaos. So, [laughter]

chaos. So, [laughter] so when you can deal with chaos, that actually makes you even more stronger.

>> 100%., Uh, and, I, think, the, same, reason, why Indians are successful as CEOs in America because they can deal with the chaos of today's world. Right. This is

very true. Are there particular areas or verticals within the Indian ecosystem that you think represent perhaps bigger areas of opportunity that not enough people are tackling right now.

>> Well,, I, mean, I, can, tell, you, where, I'm working. So that means I've therefore

working. So that means I've therefore come to the conclusion that these are areas that are worth working on. Right.

So obviously uh in AI the work is on how to use AI at country scale to really change uh people's lives. So we are working on language to make sure that

see if every Indian can access a service in his own language you know like your Maya guys are talking about then suddenly you have expanded dramatically inclusion right everybody can

participate talk to the talk in Tamils Hindi my whatever and then that just changes the whole equation of just like touchphones change because people could

you know put on YouTube videos without uh knowing how to use a keyboard so same thing is happening with voice now so I think language. Second is education and

think language. Second is education and some of the work we're doing is how do we improve basic literacy and numeracy with AI tools because if more pe more kids can read, write, do arithmetic then

big benefit and the third is agriculture because India is such a large farmer base. So this is what I'll call as my

base. So this is what I'll call as my applying AI for social impact. So that's

one bucket where I operate. The second

is uh I'm involved with designing something called the fintet. Yeah

>> which, is, the, financial, internet., The

idea being that we have seen the power of tokenization and the public blockchains have shown us that you can tokenize, you can create a verifiable asset. It's immutable. You can sell it.

asset. It's immutable. You can sell it.

You know, it really, you know, removes the friction of trading an asset. But

this is all being done in the crypto kind of world which is outside the mainstream world. And now, of course

mainstream world. And now, of course some of it is seeping in. in a way that people are saying I'll buy crypto as an asset class and all but if you really want to change the system where every

type of asset whether it's gold, real estate art deposits bonds everything can be tokenized. Is there a way to think of an architecture for doing this

that is agnostic to the asset? So any

asset can fit into this and that any asset can trade and trade at very very high volume. In the real world we need

high volume. In the real world we need high volumes. So the internet

high volumes. So the internet architecture provides for that and that we have set up a global uh institution called networks for humanity with the internet in that which is to make create

this and there's tremendous global in fact I'm going to Singapore day after to speak at the Singapore fintech festival on on the internet.

So I think that is going to have a big impact on on on financial things and that's leveraging blockchain but in a regulated world.

It's not creating a parallel world of cryptocurrency. It's regulated in in the

cryptocurrency. It's regulated in in the reg. That's one. The second one. The

reg. That's one. The second one. The

third thing is we're doing a lot of work on open networks. How do you how do you create more inclusive solutions which allow people to participate without an

aggregator sitting in between. So, which

is our open network for digital commerce does about 15 million transactions. uh

you know so you can buy and sell directly between consumer and seller. So

example namayatri which allows you to book an auto without paying you know 30% to some guy in the middle. So those are all things uh you know different but in that the one of the biggest

opportunities is what we call as the digital energy grid. I was going to ask you about that.

>> Yeah., So,, so, I, would, say, if, I, look, at what I'm working on, it's internet, open networks including digital energy, AI for public good and then taking DPI

global >> you, know, because >> we, learned, a, lot, of, lessons, in, DPI., So

we're saying how do we take it to other country. I'm just back from Cape Town

country. I'm just back from Cape Town where we had a global DPI summit. I came

back on Friday and we had 1200 people from 110 countries who want to implement digital public infrastructure. So that's

what so these are three four areas where I do some stuff. So some follow-up questions on that. So on the agriculture part in particular you're right like India obviously huge it's a huge segment

of our kind of like our labor force but is there a way to like how can we help entrepreneurs also on the more capital side like capital capitalism side of that?

>> Yeah., No,, see, I, think, what, do, farmers need, right? They need access to

need, right? They need access to information.

>> Yeah, >> they, need, to, know, markets,, prices, weather, soil conditions. They need

access to inputs, better seeds, better fertilizer and so on. They need access to capital, be able to borrow money to invest and they need access to markets.

Whatever the thing is, how do we make these accesses easier? So through ONDC I should, be able, to, access, the, market.

>> Mhm.

>> Through, the, account, aggregate, system, I get access to credit through this thing called Vista which by the way is a government project. You get access to

government project. You get access to information.

>> So, if, you, can, at, scale, reduce, friction to information, market access, capital access and so on then something will happen. Markets will also play a role. I

happen. Markets will also play a role. I

mean remember everything that we do in digital public infrastructure is enabling market innovation. Today

>> the, market, value, of, companies, built, on top of India's DPI is $150 billion. It's

not small. I mean you know you saw you know people like Pine Labs going public.

You saw grow going public. They all use phone pay.

>> Phone, pay, is, going, public, shortly., They

all use the underlying EKYC or this and that that we built.

>> Yeah., It's, interesting, because, I, often feel as though that you know at least in some of the characterization of India 1 India 2, India 3. The question is like in that India three bucket like you know it's uh I mean I feel as an entrepreneur

just focus on the first two and then the India three is kind of like listen the propensity to pay kind of like the disposable income all of that is too low then the question is well is there a way

to in some ways bootstrap the market mechanism there by using the government either for aggregate demand or something like that but maybe that's something that can happen in the future yeah >> but, also, I, mean, when, when, we, design

digital public infrastructure different from a market see in the private sector You focus on a market segment. Y

>> and, then, you, try, to, get, product, market fit for that market segment. And you

obviously want a paying market segment will buy your stuff.

>> When, you're, designing, public infrastructure, it's how do you get everybody in the game? How do you get a billion 1.4 billion people in the game?

How do you make how do you level it out so that they're all leveled out and all have the same access? And how do you then create an insurance method for them to improve their lives? That that's a different model of thinking. Then

>> it, is.

>> Yeah., So, when, we, think, about, it,, there's a lady in a village. She doesn't have ID. She gets an ID. With ID, she gets a

ID. She gets an ID. With ID, she gets a bank account. With ID, she gets a mobile

bank account. With ID, she gets a mobile connection. Then she starts a small

connection. Then she starts a small operation using borrowed money. She

starts selling stuff. When she sells stuff, she gets it a digital payments using UPIs. That gives a digital trail

using UPIs. That gives a digital trail of her payments which she then can monetize and get a loan. So you know, we created a ramp for a billion people to improve their lives. So when you think DPI, you think about how do you create

these kind of things. Then on top of that markets will operate. So different

guys will come in and build applications use cases for the segments.

>> Yeah., On, the, on, the, internet, side, Nan, do you think that you know um and because you're right like you know kind of crypto in some ways uh poison the well where like essentially all the rampant

speculation defy this and that kind of in some ways offiscated the fact that this is actually just financial infrastructure. This is global

infrastructure. This is global decentralized permissionless essentially like scaled out infrastructure and that's fantastic because we haven't seen innovation in that part of the stack for

a long time but I know that India hasn't had a ton of innovation in that particular kind of like you know even the even the stable coins or kind of the emerging parts of like I would say uh

crypto rails that are happening in America. So do you think that there is

America. So do you think that there is appetite for regulatory change there?

Well, I think there there is some because it'll also be forced because if if if, you, look, at, what's, happening, in the US, there's a huge adoption of uh these kind of things. The Genius Act which has essentially regularized stable

coins.

>> Correct.

>> So, if, a, dollarized, stable, coin, is, around and creating a fast payment rail, then you know countries have to respond.

>> They, need, to, have, their, own, version, of that.

>> So, so, even, if, you, don't, like, it,, you, may be forced into it. But in any case when we think about internet it's it's we're solving it globally. You know we have labs around the world. We have people

doing early projects in Singapore, US Europe. So it's it's a global thing.

Europe. So it's it's a global thing.

>> Okay., And, maybe, the, last, follow-up question on your on your three areas.

Can you tell us a little bit about the digital energy grid?

>> Yeah., So, I, think, the, digital, energy, grid came from the realization that the nature and this is really about electric energy. Now electric electric energy is

energy. Now electric electric energy is the fastest growing part of the energy ecosystem because of AI, data centers EVs, heat pumps, everything needs electricity. Electricity is a growth

electricity. Electricity is a growth industry.

>> Now, electricity, has, not, actually, gone through any big change for 100 years.

Last time electricity had issues of uh technology was 100 years back when Edison and Tesla fought with each other.

>> CCDC., Yeah., Exactly.

>> Yeah., So, that's, about, 120, years, back., So

but now there's a dramatic change happening because of the nature one is of course renewables coming but more importantly distributed energy. So every

home will have a battery, every home will have uh uh battery and an EV car.

Every home will have a rooftop solar.

Every home will have a smart meter. So

every home now is actually a producer consumer, buyer and seller of energy.

And price of energy in the old days was one bill at some rate. Tomorrow energy

price will change every millisecond because when the sun is beating down power will be available cheap. When

there's shortage it'll go up. So dynamic

pricing will happen every day and everybody is a buyer and seller. So it's

going to become a very different world of you know fast fluctuating transactions and we don't have the digital infrastructure for that. So

digitally designs the a world where everybody produces buys, sells all in real time.

>> Yeah., Know, I, I, couldn't, agree, with, you more. I think the combination of uh an

more. I think the combination of uh an order of magnitude ubiquitous on demand electricity coupled with modern kind of like actuators like motors coupled with modern power electronics will actually

make a lot more devices in home. So

whether that be humanoid robots quadriped robots like I think that the possibilities are really endless. And we

seeing you know we there also we have a network of startups.

>> We, we, have, we, have, this, thing, called, the unified energy alliance.

>> So, all, all, the, there, are, many, people working in the space. Many startups

working in the space. So they all come together and already there's lots of transactions happening all on this infrastructure and now India is coming out with something called India energy stack. Interesting. and I'm an adviser

stack. Interesting. and I'm an adviser on that on how do we create across the country this modern infrastructure and of course there's also global uh conversations >> because, a, lot, of, these, standards, and

interfaces and protocols for connecting the grid are just very haphazard right now okay um okay so we have about 10 minutes left I'm going to open it up to

two questions from the audience maybe three okay um I saw you raise your hand first >> uh, I'm, 23, I'm, massively, inspired, ired, by your work. Uh my question to you was a

your work. Uh my question to you was a lot of startup is about cutting down a lot of signals and focusing on a niche.

So what ideas have you previously rejected working on and why? And like

how do you reject ideas? Oh

>> you, mean, as, as, an, investor, or, as >> No,, no., As, something, you, would, focus, to work on?

>> Oh,, okay., Okay., So, I'm, slightly different because I I'm more doing this public stuff, right? And I have a grid

2x2 matrix like consultant guys [laughter] low impact high impact low friction high friction. So I work in

that quadrant which is high impact low friction. You get me high impact but

friction. You get me high impact but there are many things that are high impact with high friction. What happens

when you have high impact, high friction is 80% of your time goes with fighting with somebody who doesn't want it to happen. So I don't want to waste my time

happen. So I don't want to waste my time on that. So I focus on high impact, low

on that. So I focus on high impact, low friction. And by definition, this

friction. And by definition, this quadrant is something that most people can't see because if people can see them, they'll say, "Oh, this is going to affect me, so I'm going to stop this thing." So you have to be outside that

thing." So you have to be outside that realm only. So So that's what I do. So I

realm only. So So that's what I do. So I

I select for ideas in in that quadrant and I then I look for uh opportunities where if you build some basic public rails market innovation can flourish

and uh then I you know I work with the best people who want to work on these problems. >> Thank, you, for, the, insights, Nandar., My

name is Manu Manu Meron and I'm building an AI product for leadership development. Here's my question to you.

development. Here's my question to you.

The world of work is getting fully upended because of AI. Being AI aware is almost becoming as important as being digitally aware. So for a highly

digitally aware. So for a highly motivated professional today, what does it mean to differentiate in this era?

Well, I think my my feeling is again this is again my view it's not is that the middle ground is very vulnerable in

the so on one end you must have the mental models to think from first principles because first principle thinking is

becoming more and more important whether you're an entrepreneur whether you're working and that AI is now going to do that because it requires first principle creative thinking that's one part and on

the other end human skills, leadership empathy collaboration motivation getting a bunch of diverse people to work on the same goal. So you have to if you develop those skills. So if you're

able to deal with first principles on one side and able to deal with this human skills on the other side, you're safe.

>> I'll, add, one, thing, to, this, which, I, I mean completely agreed with Nandan.

One of the cool things about AI is that no matter who you are in the world, it's only been around for three years. Like nobody has that much of a

years. Like nobody has that much of a head start. So I often think that like

head start. So I often think that like if you really want to be AI native I mean I honestly feel that like spend two hours a day just learning how this thing

works like actually go and you know like some of the people who in some ways understand the most they have spent you know tens hundreds of hours like poking around with the chatbot and it's also kind of unique you don't need to be a

programmer to kind of understand how it works you can literally sit there prompt it play with it so I think a lot of it is actually coming back to having a beginner's mindset and to think about

this AI not as something that is that much more powerful but it is a tool with the world's most natural interface to kind of learn about how to do things right like figure out like when does it

make sense to use it in reasoning mode like figure out like when does chat GPT should it go think for you for 40 minutes versus like you know think for you for like 30 seconds figure out which areas it actually does well in which

areas it needs your supervision and I think that it really is a green field right now nobody knows that much about it in my opinion. Um, so I think it's just a beautiful time to have a a beginner's mindset. I spend a lot of

beginner's mindset. I spend a lot of time just like playing around with these models trying to learn about them. Uh

and I think it's it's actually really like fun because it's it's a unique moment in time. We have one more question for you.

>> So, you, talked, about, AI, for, public, good in different sectors. uh I worked in this area of building better technology for people with disabilities and if you see in a lot of developed country like

US UK it's very well developed like the infrastructure the environment for them not to a large scale but to [snorts] some extent in India we don't even have

those basics for a lot of people with disability so what is your opinion on how AI can help a lot of people with disabilities to better you know interact

in different environments Well, I think uh uh it it depends also on what it is, right? I mean because there can be uh different kinds of

disability and the response to different the be different. I mean if you know providing wheelchair access is not may not be able to solve with AI. You have

to provide wheelchair access in in your office or whatever. But I think for example uh I know many people are working on visually challenged disability and they're using AI to improve the ability of uh the visually

challenged person to be able to read stuff and all. So I think there's no one answer to this but I think we'll have to look at the whole matrix of disabilities and figure out which of them AI can play a role.

>> I, have, another, question., Why, do, you think there are less women entrepreneurs in India and also in the world? Like

what are the gaps there and how can we solve it?

>> I, think, I, think, men, better.

[laughter] [laughter] >> Okay., Well, said., Okay., Uh

>> Okay., Well, said., Okay., Uh

and then two two questions to finish up here. You strike me as a person who is

here. You strike me as a person who is so focused on kind of curiosity and continuing to build that you very much look towards the future, right? But if

you had to kind of like think about what would you like, you know, what do you hope from your work, what are the ways that you know what would you hope like people are inspired by

>> you, know, the, fact, that, you, know, you, can bring change.

>> Yeah., and, that, that, you, know, it's, not you don't have to get up and say nothing is possible you know you if you h if you have it if you have focus determination long-term you know marathon running kind

of thing then you can bring change so I think that that that should be the lesson for everyone >> you, can, just, do, things, if, you, think, yeah I I deeply respect that uh and last

question I often ask people you know I have three young children and uh it's it's quite it hard to extrapolate. I

mean, you know, you asked a good question. Some things will remain

question. Some things will remain fundamentally human in terms of how we work with people, in terms of what does leadership look like. Um, but it's in some ways it's also quite hard to know

like what the world will look like 15 years from now as kind of this new very powerful technology radically changes the landscape. I'm not saying work will

the landscape. I'm not saying work will go away, but will look very very different, I think. So, what advice would you have for me as a parent and perhaps a bunch of other parents in the

room about what what do we what should we equip our children with in terms of maybe skills, maybe knowledge, maybe empathy? I'm curious.

empathy? I'm curious.

>> I, think, I, also, don't, know, the, answer, but I think >> having, a, sense, of, purpose.

>> Yeah., Because, because, purpose, will, trump everything else if and the purpose need not be you know some dramatic purpose change the world. It could be a simple

purpose but and there's enough evidence showing that purpose leads to better health longevity and better mental uh you know mental

health you know so I think because many people will no longer because the abundance thing going to happen there'll be enough of everything. So you know you can't derive purpose from that but they

have to find something else which which gives you purpose. So I think go for purpose.

>> I, love, that., Um,, thank, you, Nand., It's

>> Thank, you, very, much.

>> A, pleasure., Thank, you, so, much.