Photo
Photo
Photo

303 / March 16, 2025

Sharad Sharma On India’s Digital Playbook, UPI’s Rise & AI’s Future

120 minutes

303 / March 16, 2025

Sharad Sharma On India’s Digital Playbook, UPI’s Rise & AI’s Future

120 minutes
Listen on

About the Episode

Can India keep its DPI edge?

In this episode of The Neon Show, we have Sharad Sharma, founder of iSPIRT, the technology think tank behind India Stack, Health Stack, and other digital public goods. We discuss:

How India Stack transformed the digital infrastructure of India, inspiring other countries to adopt similar models.

India Stack was built to prevent digital colonization, but without strategic autonomy, India risks becoming a trade colony of China and a tech colony of the US.

UPI revolutionized payments in India, but Brazil’s PIX outpaced it globally due to better adaptability.

DEPA enables secure data sharing, while MOSIP helps build independent digital identities, both ensuring open and inclusive digital infrastructure.

With the ability to unlock non-public data at an unmatched scale, India holds a key advantage in building the right AI ecosystem.

Healthcare AI presents a major opportunity where India can leverage its rich medical data.

India’s AI regulation follows a techno-legal approach and the AI chain framework, with an initial focus on smart safeguards for children’s AI use.

A strong AI ecosystem depends on advanced research institutions, industrial labs, and the right talent.

To sustain innovation, India must embrace a David vs. Goliath mindset and stay paranoid.

If you want to understand where India’s digital space is headed, how AI can shape its future, and why strategic autonomy is critical, this episode is a must-watch!

Watch all other episodes on The Neon Podcast – Neon

Or view it on our YouTube Channel at The Neon Show – YouTube

Sharad Sharma 0:00

We are not paranoid. Why are we not paranoid? Because we don’t have a cultural metaphor for David vs Goliath.

Our Mahabharata was in two equals. Even in Ramayana, by the time Ram reached Lanka, he had amassed an army bigger than Ravana’s army. So, in our narrative, this is not the narrative of David vs Goliath. Right? If you go to Vietnam, their narrative is David vs Goliath. They say, we first did it to the Chinese, then the French, then the Americans.

Now, bring whoever you want. So, their own thinking is, we will win the David vs Goliath battle. We don’t think like that.

Our Davids came and defeated Goliath. Right? How many people came from Afghanistan?

Handful of people. How many British were there? Even at the peak, there were only one or two lakh people.

We need to adopt a cultural mindset for David vs Goliath. And if someone leaves David vs Goliath, leaves open AI, he becomes humble.

Siddhartha Ahluwalia 1:04

Yes, he becomes humble. It takes two minutes to become humble.

Sharad Sharma 1:07

Right? So, this is the reality that we are looking at. So, you are absolutely right.

Siddhartha Ahluwalia 1:16

Hi, this is Siddhartha Ahluwalia, your host at Neon Show and managing partner of Neon Fund. I have a guest who I enjoyed the previous conversation a lot. Thank you so much, Sharad, for being here.

Just to reintroduce you, you are one of the founders of iSpirit, the organization which is responsible or in part like, you know, I would say, one of the architects of the digital public infrastructure in India, considering of UPI, DG Yatra and so many other things. And you have been an ex-entrepreneur, business leader, tech leader. Our last discussion we had, you know, about your journey, the journey of UPI, how from limping to today, you know, it’s one of the best DPI, digital public infrastructure in the world.

And today’s discussion, almost after one year, two months, the last podcast was in November 2023. We are sitting in February 2025. The world has completely changed, right?

A new US president, right? AI is a war cry for the nations. And excited to discuss, you know, what the next five to 10 years of India holds.

And, you know, just to recap some stuff, you know, we had a, you know, under innovation problem. And instead of accepting the trader and no R&D mindset, we created India Stack as innovation enablers and are exporting that to the world.

Sharad Sharma 2:49

Yeah. No, I first of all, thank you for having me. I enjoyed our last conversation.

As you know, I keep a low profile. I’m not generally on many podcasts, but you and Nansi were very insistent that I come back. And so here I am.

I usually don’t. I like working in the kitchen rather than talking about things. And, you know, my mindset is that create facts on the ground and, you know, the buzz will follow.

So, so I agree with you. Things have changed. But I think in a very predictable fashion, I think, you know, we could see many years ago that the source of country power, you know, which used to be often military power or economic power was getting supplemented by technology power.

And and today, you know, it’s safe to say that if you have economic power and military power, but no technology power, it doesn’t count for a lot. And if you have technology power, you’ll be able to accumulate the other two powers relatively easily. So it becomes very important for a nation to be to have strategic autonomy in many areas, to be dominant in many areas.

One way to think about it is in how many areas are you in the circle of five? Because almost anything that you take, there are only four or five players that really matter. Right.

And so is an Indian player amongst those four or five players. So what does it take for us to have an Indian player amongst the circle of five? Those are the kind of questions that animate us.

And and my feeling is that just the way our focus on India’s stack was to avoid digital colonization today. You know, the danger is that we would be a trade colony of China and a tech colony of the US. And and and we have to find our way out of this scenario.

And because this scenario is not a happy scenario for India, both economically, politically, geopolitically. So we have a lot of work ahead of us. Yes.

Siddhartha Ahluwalia 5:02

And let’s consider, right. Last time we discussed the India stack, its power. So in the last one and a half year, how has India’s stack gone global?

You know, some of the ideas like digital primitive for flow of people, money, information is now better recognized outside India in low income, middle income and high income countries. Right.

Sharad Sharma 5:22

So if you take us, you know, if I take you back to 2012, you know, at that time, the thinking was that we need to introduce three more primitives in this world of digital. The three primitives that existed, which is really flow of packets, which is TCPIP, flow of messages, which is SMTP. That’s how our email systems are based.

And flow of content, which is HTTP. Then that’s how web content is shared. You know, they were very useful, but they are not central to a digital transaction.

So, for example, let’s say, you know, I have to go to Chennai and I want to go by train. Then I would need to identify to whichever booking application I’m using, who am I? Right.

Then I have to give them some information about which day I wish to travel, from where to where. And then I have to do a payment. So I think our our realization 12 years ago that if we make these three things, flow of people, which is identity, flow of personal information, which is now, you know, in the financial world was the AA account aggregator, which is now getting extended to all sectors through the DPDP Act and and flow of payments, which is UPI.

If we are able to do this in a better way than what the West has been doing, then that would lead to democratization of digital transactions. That hypothesis has turned out to be true. You know, last time I I should have listened to my last episode before I came here.

But last time I’m hoping I talked about Rajni, a street vendor, which motivated us to do that.

Siddhartha Ahluwalia 7:17

I think we didn’t cover it.

Sharad Sharma 7:18

All right. So so the way we framed it for ourselves was that we want a street vendor. You know, we have a picture of her.

And if people want to look up that slide, it the best way to find it is to say Rajni, iSpirit, India Stack and a blue colored slide will come up. And if you really dig for it, you’ll find a version from 2013 on the net. It’s mostly unchanged.

And and the goal was that can we get her an intraday loan to happen? And to be able to do that, we needed to make sure she has a bank account. But 2012, she didn’t have a bank account.

Only 17 percent of the people had a bank account. So we needed to make sure she could sign a loan agreement. Therefore, e-sign.

She could store a loan agreement somewhere. Therefore, DigiLocker. She could she could receive the monies of the loan.

Let’s say 400, 500 rupees. That’s UPI. And then if there are two Rajnis, one pays on time, one doesn’t pay on time.

The one who pays on time gets a cheaper loan. And that would only happen if she threatens to take a business elsewhere. That means taking her financial data elsewhere and therefore the account aggregator and so on and so forth.

So so this conception of of how to make change happen on the ground made a very big difference. So those ideas, I would say that 2012 ideas have by and large played themselves out. That we could create a ecosystem where there’s private sector innovation happening to serve somebody like Rajni, which was not happening at that time.

The fact that we could create innovation enablers, which is India Stack. The fact that we could avoid monopolization and wrong kind of toll collecting because the primitives, in our view, should be public utilities, data blind public utilities rather than anything else. And and that’s also kind of worked itself out.

So therefore, there’s been an explosion of digital transactions, democratization at a scale that others don’t see. And other countries now, therefore, are very keen to replicate it. And and they are then able to understand that there was a Soch behind this.

And this Soch is very important. And and many countries are now embracing this Soch and taking it forward. So the manifestation of this is that something like MOSIP, which is India’s identity platform for the world, which is anchored in Triple IT Bangalore, is now in 26 countries.

There are 943 million people signed up for it. If you go to MOSIP.io, you can see probably 121 million people would have used it today. You know, so we have Indian technology moving out at scale.

And MOSIP has done quite well. We have serious traction on the data side. You remember our protocol for data is called DEPA.

And it actually has many that has three parts. One part is to do with personal data. That’s the account aggregator system in India, because that’s where the pilot was done.

And that’s doing well. So there’s another part to do with training data. And then there’s another part to do with non-personal data.

And so so the first one is doing well in India. People don’t appreciate the numbers, but if I share those numbers with you, you know, you realize how big it is. So we lost, we launched AA in the public domain on 2nd September 2021. Right. So when I came to you in 23, it was two years old. It’s now three years, four months old.

 

And, you know, in those three or four months, we have cumulatively done roughly about 140 million consent transactions. But in January alone, we did 10 million transactions. So our run rate is 120 million for this year alone.

Now, that’s not bad. But, you know, in a recent meeting that we had, you know, many people were convinced that we will end this year in December with a run rate of 50 million transactions. So I told them I’m coming to this podcast.

Can I mention it there? They said, yes, you can. So I’m, you know, I’m reasonably confident that number will be hit.

So this year is a breakout year for AA. And then, of course, the similar architecture is going to roll out in every other area because of the DPDP Act. And so that is big.

That is very appealing to Global North. And then we have other things that are appealing to the middle income countries like Vietnam, Thailand and others. So we have a Soch, a thinking that people are embracing.

We have we have different starting points for different types of countries. So in that sense, our DPI journey is going well. In fact, the word DPI itself is is relatively new.

You know, as a term, it was used for the first time on September 13th, 2022 in a press release that the government of India gave to announce what would be the focus areas for G20. And the credit for that, actually, you know, we had very few days to align on a single term. And and rightfully so, many people had different opinions on what that term should be.

You know, we had used DPG before, which we wanted to abandon because it was too broad a term. We wanted to pick something which was narrower. And and there were, you know, there were six opinions for five people in the room.

Right. And and Mr. Amitabh Kant, I really appreciate the role he played. And, you know, he took a stand.

He kind of banged heads together and said, no, we don’t have time to, you know, kind of endlessly discuss this. We’ve got to decide and move forward. And the goal was to create at that time what we call as a category brand.

And that has happened thanks to the G20 and all the efforts that went in. Today, you can go to any other part of the world. They know what DPI is.

They may have a misunderstanding of what it is. They think every open source project is a DPI. They think every government tech is DPI, which is not true.

Some government tech is DPI if it enables private entrepreneurship or private innovation to happen on top. Some open source is DPI and all DPI is not open source. So there’s understanding of that is not fully there, but at least the term is understood and people think they understand what that term means.

That’s the important part. So therefore, we have traction there from that perspective, definitely as a country.

Siddhartha Ahluwalia 14:54

What I believe is we need to market ourselves internally better that for our own citizens to feel the confidence that…

Sharad Sharma 15:04

Yeah, yeah, absolutely. And I think we have. Yeah, we have to galvanize ourselves to do that.

The reason is that. What is going to happen if you don’t do that is that actually the DPI, since it’s a category brand owned by everybody, think of it. EVs are a category brand, right?

So let’s say you were the first EV company, right? And everybody knows you as an EV company, but that isn’t a guarantee that you will be a leader in EVs. You know, it may have started EVs may have started in country A, but country B may be more important or company B may be more important to take it forward.

This is exactly what will happen with DPI. I think we started it and we tend to believe that since we started it and the conversation revolved around us, you know, it will be our calling card for the rest of the world. That’s not true at all.

And in fact, the fact that DPI is accepted as a term means that there are many parties who believe that they can drive the next wave of the conversations, the the implementations of what DPI will be. Actually, UPI is a good example of that. And, you know, today in the fast payment world, the the momentum is not with UPI, is with PIX.

It’s a Brazilian system which is inspired by UPI, as they say. And but they’ve done a very good job of the basics of taking it out. So we are on the back foot as a country.

You know, we managed to take UPI on a merchant level in very many places. So we can, you know, have you pay by UPI in Dubai or let’s say a few places in Paris. But those people who are collecting UPI payment from Indians, they’re just merchants who happen to be another country.

There is literally, literally no meaningful country which has adopted UPI as a domestic payment system within their country. And that’s not because UPI is a bad system. It’s because we became overconfident and we didn’t make the necessary changes that were needed for UPI to travel to other parts of the world, despite the fact that we made those changes with MOSIP.

You know, we had a we have now done an A-B test that weaponization of of DPI is a concern. So wherever we address it, like in MOSIP, it travels wherever we don’t address it, like in UPI, it doesn’t travel. And luckily, you know, in DEPA, we learned from that A-B test and we are trying to make sure that it travels. We don’t make the same mistake that we did with DPI.

Siddhartha Ahluwalia 17:56

And for our audience, can you explain the full form of MOSIP, DEPA and a few other terms?

Sharad Sharma 18:00

So MOSIP is modular. So MOSIP actually is an effort to take foundational ID. So the so the idea here is that there is almost like an ID chain, right?

And the ID chain is that if I go to IRCTC site or I go to Digi Yatra, I originally identify myself with an identification that already exists. So we call that the foundational ID. In India, that’s the best way to do that is through Aadhaar.

But subsequently, once you’ve identified me as Sharad, then the ID system that that Digi Yatra has or a bank has or IRCTC has can be their own. That’s called federated ID, right? So so I can have many, many IDs in the system, each one of it authenticated with a foundational ID.

So that’s like a chain. That we are talking about. So MOSIP is a foundational ID system for the rest of the world.

And why does it exist? Is because one of the promises that we make in MOSIP is that it is non weaponizable. So if you implement the MOSIP system in your country, there is no backdoor that India will have that will allow India to control that system as you go forward.

And and since identity is so mission critical, people demand that, right? And and since we address that very carefully in MOSIP, it travels. So when you do UPI, they want the same approach, right?

And so they say, OK, we’ll use your payment system, but we don’t want that you have some backdoor kind of strings that you can attach. In fact, the whole story of DPI is that it doesn’t have those strings, right? If you use a Visa MasterCard, the US government will potentially have strings that they can use to control one day, perhaps even shut the system down if needed.

In this case, the whole proposition is that if you use DPIs, you don’t give that control to anybody else. You own your own destiny. And that required us to open up the protocol in a in a very structured way.

We didn’t do that. You know, we are doing that now. But Brazilians with PIX did that I think in July or August of 2022, and so we got we got a little cocky and and overconfident about our own momentum in that area. So it’s been a humbling experience. But, you know, as long as we learn from it, that’s fine.

So coming back to your point, all of these are now called DPI sutras. So these are principles of DPI that are important for many people. They can go to a site called CitizenStack.world where they’ll find these sutras. So to to make sure that even as first movers, we remain as main movers. You know, we are doing four things to make it happen. First is to create an offering brand.

Right. So DPI is the category brand. But India’s offering is called CitizenStack.

So every country, we are encouraging every country to have a CitizenStack. India’s CitizenStack is called IndiaStack. But Ethiopia, you know, ask is, hey, Ethiopia, what would you want to call your CitizenStack?

You know, Philippines, what do you want to call your CitizenStack? So by creating an offering brand, we can have a more meaningful conversation. So that’s a branding kind of a kind of an answer to this.

And, you know, one of our branding whiskers is Mr. Amitabh Khan. So he’s had a role to play in making this happen. And the second part is that all DPI’s are not alike.

For example, some are good on the issue of weaponization. Some are not so good. Some allow for corporate monopolies to come up.

Some don’t allow corporate monopolies to come up. Some are easy to regulate through techno-legal regulation. Others are not easy to regulate through techno-legal regulation.

Some are privacy preserving and others are not. So we put all of these things, the five of them, as sutras so that the world can evaluate anybody’s DPI, including India’s, by testing them against these DPI sutras. So that allows us to guide the conversation based on our experience.

And we, of course, have more experience than others. So that is useful. Third is that not everybody will use our DPI’s.

They may say that, you know, we love what UPI has done, but we like to do Pix in the spirit of UPI, but it’s not UPI, right? And then the question becomes how do the two interoperate with each other? So how do we make that happen in different areas requires us to play a role in the standards space.

And so India is starting to create global standards organization that are based out of India. This is something that we’ve never done before. Obviously, maybe next time I come, we can have a deeper discussion on that, because obviously the best place for us to make this work in an authentic fashion is identity itself, because we have our own 1.4 billion people using Aadhaar. And then we have roughly a billion people who will be using MOSIP in the near future. So so therefore, if we account for 2.5 billion and MOSIP is still growing very rapidly, we’ll get to probably three, three and a half billion. But for the remaining three, three and a half billion who are going to build their own systems, there should be interoperability, right?

There should be a basis for us to say that there are some standards. And instead of adopting a standard from outside, India should take the lead in creating and socializing and certifying people across those standards. So that’s an effort that is underway.

And and we’ll know in a year or two how we are going to do that. We had a first international stakeholder conversation for that. TSO, they’re called TSO, Technology Standards Organization.

So India has created the first TSO, first global TSO from India. And so we’ll know in a year or two how that picks up.

We’re looking for a good CEO to head it. So let me plug that here. If somebody is really keen in leading a global standards organization out of India in the field of ID and then if this succeeds, we’ll obviously apply that to other fields.

You know, please reach out to somebody either in iSpirit or Professor Rajat Moona is a partner in crime there. He is the director of IIT Gandhinagar. So to either of us, if you reach out, we’ll love to consider that.

So so therefore, coming back, fix our branding, fix the principle sutra so that people can evaluate, take leadership in standards. And the fourth is that put some money on the table to control the funding ecosystem. We think there will be roughly about, you know, two to three billion dollars spent on the DPI ecosystem in the next three years.

And and and and we are very keen that that that we have some control on how that ecosystem emerges. Therefore, the government has been very thoughtful in allocating some money. But obviously, at best, we’ll have 100, 200 million because we’re not that rich a country out of the three billion spend that needs to happen.

But, you know, our private companies can come in and play a role. We are seeing very strong participation from some of the companies. TCS is one of them.

And and we can have startups play a role, you know, many startups are therefore getting actively involved in this. And we have to co-opt other funders to support our approach as we go forward. There are obviously some rich people who will support other players because everybody is trying to consolidate and make themselves the most relevant party here.

But I think we have some momentum right now to be able to shape how the funding ecosystem will develop. So these four elements will allow us to remain a main mover in DPI globalization. But it requires us to be paranoid.

We we we have a great difficulty in India to be paranoid. And, you know, we tend to think if things are going well, what can go wrong? Let’s not worry.

You know, we are a great optimist. And and to bring all these things to life, a new brand, new sutras, new standards, a new funding ecosystem. We have to motivate it, not because things are not going well, but they could potentially not go well.

And so therefore, paranoia is very important.

Siddhartha Ahluwalia 27:56

I think as a nation, not being paranoid has been a weakness for thousands of years.

Sharad Sharma 28:01

Yeah, yeah. We get a little bit of a chance. We go from being the underdog to the top dog and we think everything is sorted out.

And this is a problem even even in DPI. And so it’s very, very important that we maintain an underdog mindset. And and to your point, I think it’s a cultural problem.

I often say we must learn from the Vietnamese. You go to Vietnam, everybody has that underdog mindset. And if you scratch the surface, they’ll tell you, oh, we defeated the Chinese, we defeated the French, we defeated the Americans.

Bring anybody on. You know, that’s the part of a national ethos. We don’t have that.

And for because our history is different. Also, if you go further back, even in Mahabharata and Ramayana, you know, Mahabharata was not a David versus Goliath battle. And Ramayana could have been a David versus Goliath battle.

But by the time Ram reached Lanka, he had amassed an army bigger than Ravan. So in the end, it wasn’t a David versus Goliath battle. So we have to create stories to tell our own people about this David versus Goliath.

It doesn’t come naturally to us. And and therefore, you know, it’s a recurring problem that we have. And and we have to create these enclaves where people maintain their David versus Goliath kind of a mindset.

Siddhartha Ahluwalia 29:32

And historically, you know, Afghans conquered India, Mughals conquered India because there was not enough paranoia in those British conquered India.

Sharad Sharma 29:44

Absolutely.

Siddhartha Ahluwalia 29:44

Because there was no paranoia.

Sharad Sharma 29:45

Small groups of people could, you know, view the Goliaths then. And we were humbled by David’s, right? And so.

So in this today’s world, I don’t think in this multipolar world where even the U.S is not as dominant as it was, everybody will have to be the David. David versus Goliath mindset will determine who’s going to do, you know, in a pre-meeting we were saying, look at open AI, right, it thought it had arrived, and boom, you know, that’s not true anymore. So, so we can’t rest on our laurels, we have to keep working.

Siddhartha Ahluwalia 30:24

It’s a great example. Two years ago, Sam Altman came to India, Rajan Anandan asked him question, can a startup from India compete with open AI? He said, it’s not possible, don’t even attempt it.

Yes. And few days back, Sam Altman was here in India, and he’s saying we can collaborate.

Sharad Sharma 30:42

Yes, exactly.

You know, and you know what I spent, because of the citizens, citizens tech conference and the AI summit, I was in Geneva and Paris, you know, for these two weeks. And, you know, and the Europeans, many of them are lamenting, and they said, you know, why is it that we couldn’t have done DeepSeek, right, because they don’t have a GPU problem, they don’t, you know, their funding support is also there, especially in countries like France, is because they also succumbed to this kind of a thinking that we can’t be the Davids, the battle is already lost, right. So you have to admire the Chinese for not getting carried away.

And, and I think in India, while Rajan remained and push back, not only vocally, but also with his investments in trying to prove Sam Altman wrong. But there has been a very large camp, which said, let’s give up. Let’s only focus on diffusion.

And, you know, it’s been a very, very interesting thing. And, you know, and it took a DeepSeek to resolve this and say, no, we should not give up. Right.

It’s not a closed door. You know, it’s a worthy fight for us to undertake. And, but it’s very surprising how quickly are we willing to succumb to this narrative that, David is not for us, you know, let us stop trying.

Siddhartha Ahluwalia 32:20

And there’s a very interesting analogy that we launched Chandrayaan for less than $100 million whereas space missions of equal size are…

Sharad Sharma 32:28

Right. I would argue that we have succeeded only in those places where we become the David, in the David versus Goliath. You know, as many people, notably, you know, one of the persons who doesn’t get enough attention, but has had an incredible impact on ISRO is Kasturirangan, who lives in Bangalore.

And now, of course, I think he’s, you know, mid 80s or something like that. But the fact is that, you know, the stories that he would tell you is that the fact that cryogenic engines were denied from India was a galvanizing role for them to build their own cryogenic engines. You know, we’ve actually thrived only when it is that we embrace the challenge and we say it’s going to be horribly difficult, but we will keep at it, you know, and do something big.

And that spirit is very important. And we should not lose that spirit as we go forward.

Siddhartha Ahluwalia 33:40

So like India’s fintech ecosystem, if you see today, the boost it got, it was because of UPI, right?

Sharad Sharma 33:47

Yes. And you know, but that regulatory architecture. So let’s spend a little time on that.

See, what one of the gifts that Facebook gave us is that it said that you have a technology architecture of an ISP. Right? But you can either have a business model, you can have many business models on top of it.

And one business model is that who pays for the ISP, the customer pays. Another is the content provider pays. The third is that both pay.

And the US back in time had chosen a model where they said the customer pays. And they had a reason for it. They called it net neutrality.

And that had worked for a long time. And then Facebook came to India in 2014, 2015. And they say, you’re a poor country, your customers can’t pay.

And let us as content providers pay. And they literally asked for an exception there. And they called it free basics.

And they said the benefit of that is that the customer gets ISP for free. Right? Isn’t that a good thing?

And that led to a very significant debate inside the country as to what would be right or not. And a Save an Internet campaign came, you know, here in Bangalore, it was a very collaborative effort. In fact, 2 million certified signatures were submitted. Government of India took a stance in favor of preserving net neutrality.

That three page policy that came is taught in many law schools across the world as a very well written policy. Short but well written policy. And then we saw that President Trump in his first term appointed a person as FCC chair.

You know, fact that he was of Indian origin is not relevant, but he was. His name is Ajit Pai. And he genuinely felt that net neutrality had outlived its utility in the US.

So he made a case for getting rid of net neutrality. And of course, that meant that the content providers, which is Facebook, YouTube, Netflix and others will also have to pay the ISP, which they didn’t want to. So they went to court preemptively to justify preserving the net neutrality policy.

And the same Facebook, which had insisted on removing net neutrality here, was now defending net neutrality here using that three page policy that we had created to basically preserve our net neutrality. And so it told us that the same technology infrastructure can have many overlays of regulatory architecture. And there isn’t one regulatory architecture that is preordained.

The same company can think one should happen in India, another one should happen in their home country. And so it opened our eyes in a way that we had never thought before. And then it led to innovation in four areas.

It led to innovation, which many people, I don’t know whether they appreciate. And I don’t know if I covered this in the last podcast. Did I?

Siddhartha Ahluwalia 37:18

No.

Sharad Sharma 37:18

All right. So it led to innovation in four areas.

It led to innovation, of course, in this area of cable television, because TRAI had taken a role in deciding the net neutrality policy. And then they said, why shouldn’t it apply for cable television? Now, in India, cable television follows the must carry rule.

It’s good for the ecosystem, but not good for the consumer. You pay actually more for cable TV today than you were paying earlier, because that is how the system is set up. But it’s a good system.

And I think it is healthy for the cable television industry. Then at that time, we had Mr. UK Sinha as the chairperson for SEBI. He said, why should it not apply to mutual funds?

So India or SEBI embarked on a very serious implementation of the same concept of kind of open networks. Net neutrality is a subsidiary principle of open networks. It says that if you have an intermediary, it should be either a buyer’s agent or a seller’s agent.

If it’s an agent of both, it is neither’s agent. Therefore, very hard to regulate. Right?

So they said, in mutual fund, if you have an intermediary, it should either be a seller’s agent, in which case it’s called a distributor, or a buyer’s agent, in which case it’s called a RIA. You’re not allowed to be both’s agents. And, you know, market will decide which one will do well.

In fact, that system has done very well in India. And the fact that stock market has popped up is because lots of new money is coming in. And because that system provides clarity and good regulation of the mutual fund ecosystem in a way very few countries do it today.

And so that has worked very well. Mr. Jayant Sinha was the Minister of State for Civil Aviation. He was inspired by this.

And he said, you know, we should have the same thing for OTA. So I don’t know how many people in India know that when you buy a ticket from Make My Trip, Yatra, or Clear Trip, or any of these places, they are passenger’s agents. So you pay a convenience fee, the airline doesn’t pay them.

And that means that when a Akasa Air as a new airline comes, they are guaranteed visibility. So if I say I need to fly to Bombay, show me all flights, Make My Trip cannot say I’ll show you flights of my friends. They are obliged to show all flights because they’re my agent.

You know, they’re supposed to give me all options. Now, this is not the case everywhere in the world, right? And finally, it was applied very successfully to e-commerce.

And that is how GEM, which is now doing very well, is a manifestation of that. So these four gave us the sense that in phase two, we can extend that even further. We can extend that to lending, which is open credit enablement network.

We can take it to private e-commerce and therefore ONDC. We can apply it to health, which is called open health services network. So therefore, this regulatory idea of how does an intermediary behave is a very important idea.

The second part of the regulatory architecture that is equally important is techno-legal. I often ask in a talk, let’s say you have 100 people, how many of you have a license without appearing for a driving test? You know…

Siddhartha Ahluwalia 40:36

50 people go.

Sharad Sharma 40:37

Exactly. Usually 30-40% of the people will raise their hand. So in India, in places like India, rulemaking doesn’t guarantee following the behavior on the ground.

So we need a new approach to be able to do that. And most people don’t appreciate that our GST is a very good example of that approach. And therefore, GST revenues are very hard to fudge because we do the value chain kind of settlement not on a monthly basis, not even on an invoice basis, but on a line item of an invoice basis.

 

So which means that the tech system has to be fairly complex. And therefore, it took a little while to settle down. But subsequently, thanks to a wonderful Revenue Secretary, Mr. Ajay Bhushan Pandey, who was also the UIDA CEO at that time, he co-held those positions. And many people don’t appreciate because he’s a very quiet person. He never talks about it publicly. And I think people like him are, you know, who don’t beat their own drum are often not fully appreciated.

 

And his contribution to, of course, at that time, Aadhaar, but also to the GST system is absolutely seminal for the future of the country, because that allowed us to create a GST system, where revenue growth has been quite substantial. And I like to say that every physical infrastructure that you see is a result of actually GST revenues growing so rapidly. You know, not too long back, there used to be only 1 lakh crores a month.

Now, there are 2 lakhs crores a month. I mean, this is just a phenomenally good system. And…

Siddhartha Ahluwalia 42:26

All the highways that we see the result of the GST collections.

Sharad Sharma 42:31

Absolutely. And so therefore, you know, and there are simple ideas, this idea that we must do the settlement, not at the monthly level, which is what every other country does. You know, which means that if I grow cotton, I sell cotton to you, you turn this into, let’s say cloth, and then you turn this into yarn, yarn to cloth, cloth to kurta, there are 4 steps involved.

 

And in every other part of the world, you would just do the input credit based on the net business in a month between the first and the second party. So you would have 4 kind of tax filings that you would happen. But if you do it at the invoice level, then usually it will be 3 invoices a month, so it will be about 12.

But if you do at the line item level, then it’s about 36. So it did take time for it to settle down. And I think, as I said, Mr. Ajay Bhushan Pandey played a very seminal role in making that happen. And a group of us also made a small contribution thinking about the architecture, which then became part of the RFP that went. But I think, to me, the most important person here is Bharat Goenka. And, you know, he helped think this through.

And his perspective was that if you don’t do it this way, you know, I know how people will beat the system. I’ll have to even put the techniques of beating the system into tally, ultimately my customers will demand it. So, you know, his thinking at the core, you know, in those early days, and then a set of new people coming in, who helped translate it all the way to a working system is a very important success story.

So we understand these tax liability chains now better than anything else. And that will be important for AI. We’ll come back to AI a little later, in AI regulation in a very big way.

 

So I think the point I’m making is that if you really want good ecosystem, you have to make decisions about what is the minimal stuff that should be in public infrastructure hands. Because that should be minimal. If you put too much of it, that’s like socialism.

 

You put nothing in it, that’s kind of unbridled capitalism. That’s also not good. So some part of it has to be in public infrastructure.

And it has to be regulated with a regulatory philosophy that is well stated. And we did a very poor job of that with Aadhaar, better job with UPI, better still with DEPA. And I think we’re getting better and better.

 

And I think the focus now as a country is that we have to really articulate it very, very well. Why? Because this globalization of DPI is forcing us to do that.

 

Because we don’t want people to make the mistakes we’ve made. And one reason for them to come to us is because we can share our learnings with them. So we have to become very good at doing this.

 

And I’m hoping that many of the old and the new institutions, I know, for example, IIM Bangalore has already set up a center to focus on this under Professor Srinivasan. But some of the newer universities, for example, Chanakya University, Kriya, and I’m hoping Ashoka, which is showing some signs of interest, but they’ll also come in and they’ll establish, you know, thoughtful centers to take this Indian idea and develop it further and become the epicenter for this thinking for the rest of the world. Today, that has not happened, but we need to make it happen in the next one or two years.

So our ability to crystallize our lessons on regulatory approaches and make them available to our own regulators, very often our own regulators don’t understand it. You know, if I take 100 regulators, I would say only about 20-30 of them have a good understanding of this. Another 20-30 have heard of it, but they don’t have a very good understanding and 20-30 have not even heard of it.

So we need to get this out there in a much bigger way so that our own implementation of this improves. And we can also shape other people’s thinking across the world as we go forward. So that’s an important concept.

Siddhartha Ahluwalia 47:08

A couple of important points here. Do our regulators, you know, have built a playbook on their learning from Aadhaar, UPI, that how does the regulator have to work?

Sharad Sharma 47:18

So in FinTech, if I take you back to Mr. Raghuram Rajan’s time, there was a regulatory approach that had been worked out, which was that we’ll have a three-lane system, right? There’ll be digital public infrastructure, which is things like India Stack was the name that we used then. Then there’ll be regulated entities and there’ll be unregulated entities.

 

And UPI, for example, you have had Google Pay, Phone Pay and others in the unregulated lane. And then you have banks in the regulated lane. And then, of course, there is the protocol, the switch, which is part of DPI.

Now, that system was supposed to be applied to everything else. In fact, there’s a beautiful BIS Bank of International Settlements paper, which kind of captures this thinking very cohesively. But it has not been followed.

And part of our pain in the FinTech system is that neither the FinTechs nor the banks, unregulated, regulated and the regulator have remained true to this three-lane model. While everybody says this is the right model. So either we should come out and say this is not the right model.

And you’re going to have an alternative. Or, you know, we must follow it diligently. And we haven’t done that.

And we have not done that partly because as things change on the regulatory side, the institutional memory is not kept up as much as it should. It has also happened because our FinTechs, you know, especially the Chinese FinTechs, you know, they exploited the system. And everybody paid a price for that.

And we didn’t call that out and bring enough self-regulation in place to address that. So ultimately, we knew we had a reference model. But we were not vested as stakeholders in following that model.

And since things have slowed down in FinTech, it is time to kind of take stock and fix things. In my opinion, go back to that old model, you know, in a very clean way. I think we should do that.

Because if you go back just 4-5 years, the talk across the world was that Indian approach is the approach for the world. Now we mucked it up ourselves. Nobody talks about that anymore.

And if you’re a FinTech in India, it’s become harder to raise money rather than easier to raise money. So this is a self-goal that we have done.

Siddhartha Ahluwalia 50:18

Every 3 months regulations for a FinTech change. So FinTechs talk about how do I operate in such a murky environment.

Sharad Sharma 50:25

Exactly.

So because we ought to be a country that we should be able to say that whatever is written in Gita, we will make a bhajan out of it. But we didn’t do that. And so we have, I call it a self-goal, frankly, in this area.

But I think there is recognition. In October, the RBI governor held a very good conversation with some of us. There was deep understanding of this issue and a desire to change.

I am super optimistic with our new governor for a number of reasons. He was instrumental in account aggregator being implemented by the public sector banks. As a DFS secretary, he played a very positive role.

He’s a person with an engineering background, IIT Kanpur. And I think a deep thinker and understands this. And I’m hoping that because things have to be turned around, you know, with this new energy, new people coming in, we’ll be able to do that.

Siddhartha Ahluwalia 51:42

Because the lead that, as you said, we had in the FinTech ecosystem is not there anymore.

Sharad Sharma 51:48

Yes. So I call it a self-goal. So my overwhelming feeling is that, you know, by not being paranoid, the momentum that we build from zero to one and sometimes between one to 10, we are not able to sustain it from 10 to 100.

FinTech is an example.

Siddhartha Ahluwalia 52:11

And AI has not even started yet.

Sharad Sharma 52:12

So that’s why we have to learn how to do these handoffs well. And these handoffs can start with, you know, zero to one can be a set of well-meaning people.

But, you know, one to 10 requires institutions. And we require strong institutions and strong Veda level thinking to go from 10 to 100. Because if you don’t know what we stand for, then to preserve it becomes very, very hard.

So DPI is now in the 10 to 100. So if you don’t crystallize the sutras and be able to state them very clearly, everybody’s interpretation will be very different. Even the definition of what is a DPI will be very different.

So if you want to go from 10 to 100, formalization, codification of the knowledge that underpins that thinking becomes very essential. And I’m hoping that this new energy in our universities will play a role there, especially the newer ones. Because this traditionally happens in the hands of universities, right?

They’re very good in understanding, codifying and taking it forward. And I think many of them are showing interest. So it’s very important they step in, in a big way.

So otherwise, we will start with a bank and then end with a whimper. That’s not going to be very helpful.

Siddhartha Ahluwalia 53:37

For example, folks had assumed a little while back, India became so successful in internalizing the applications of UPI, largest number of transactions globally, digital transactions globally, that we would take it international. For example, if I had to pay a Dubai driver, I’m still reliant on Visa and MasterCard for a credit card. Why was it not done through UPI?

Sharad Sharma 54:04

I would go one step further. See, look, I don’t know of any country that actually wants to rely on Visa MasterCard. I’ll tell you even back in time, Australia, which has repeatedly made statements and efforts to this direction, because the fact is that when your economy is reliant on Visa MasterCard, you’re paying a 1% tax to a third country for this.

And people understand this is not a good thing. And look, and therefore to find your own system, which is essentially a UPI transferred there is very attractive to people. But I think, despite that, we didn’t play the game right.

I don’t blame others, I blame ourselves. And so therefore, it’s okay, we made a mistake, we have to learn from that and move forward. So I think, but what underlies all this is we have to be way more paranoid, we get carried away.

You know, we do that even in cricket, we win a few games, and then we think we’ve arrived, nothing can go wrong. And of course, things will go wrong. And so we have to get out of that kind of a mindset, we have to remain paranoid.

Siddhartha Ahluwalia 55:21

I think the Chinese have done a very good job at it. The moment you underestimate China, like people say China is done, they’re not able to treat their own entrepreneurs well, see what happened to Jack Ma. The minute you do that, right…

Sharad Sharma 55:34

They bounce back and prove you wrong. So this is a culture that we have to bring. We get carried away with our own success too quickly. And, you know, if our goal is to become a product nation, which is, you know, what some of us, people like me are vested in and have devoted our time and energy to make that happen.

You know, we have a long way to go. We are, you know, as they say in a seven inning game, we are only in the first few innings. And so we must not, if there is one thing we must change is not become overconfident too early.

Siddhartha Ahluwalia 56:13

So, you know, want to discuss a few things, right? Why China is so better than US in doing the Darwinian ecosystem right?

Sharad Sharma 56:22

Even before that, let me pick something else before that. I think, you know, as we, you know, one of the things that I like to talk about is that is about ease of doing business because we have to fix it. It’s a basic, it’s table stakes.

You know, I know, I was in Vietnam, January ended up being a very busy travel month for me. But I was there for a few days. And we are actually going to do a citizens stack conference in Vietnam on May 9th and 10th.

And, and, you know, while I was there, there was all this buzz that was taking place about this Elon Musk department, DODGE, that is there. And, and I came back and next week, they announced a 20% cut in the number of people involved in regulation, so that they can be competitive, you know, with the rest of the world. Now, the point is that in ease of doing business, if anything is relative in nature, it is that right.

 

And in today’s world, if we don’t fix this, you know, we will lose talent and companies to outside, you know, you know, this even better than I do probably 50 of our 120 unicorns are technically not Indian, right. And now the reason is that there’s been an effort to fix them, fix the problems in something called stay in India checklist. For the past so many years, you know, the stay in India checklist came to being in October of 2015, few months before the startup policy was launched.

And the progress is pathetic. And if you talk to others, particularly Manish Sabharwal, who has also been focused on fixing all the regulatory problems that we have for employment law. You know, would also tell you that despite good intentions, there’s very little that has landed on the ground.

And so we realize that we need a new approach to do that. And, and that approach is finally beginning to catch attention. And I’m hoping that we can at least do a meaningful experiment and go from zero to one and one to 10 and 10 to 100 very quickly on that.

Because or something else, because doing nothing is not an answer. In this environment, there is almost a 550 rule, right? And this is not attributed to me.

But a very senior secretary in the government says that if we have to win in some of these new industries, let’s say AI, you know, if there are five big companies, it’ll be 50 deep tech startups that will make that happen. And in any area that we’re looking at, it’ll be that ratio. And therefore, the big companies can deal with all this regulatory compliance issues, because they’re big, they can hire people and, but the small companies cannot.

And, and they either they die, or they have reduced outcomes, or they abandon India and do it elsewhere. And now in this new environment with us literally calling for capital and talent, you know, we will lose our talent. To the US, if not other places because other places are also offering, they’re wanting Indian talent to come there.

 

Siddhartha Ahluwalia 1:00:03

For example, France.

Sharad Sharma 1:00:04

For example, France, right? And Singapore and of course, Middle East, right? They’ll give you a…

Siddhartha Ahluwalia 1:00:11

Golden Visa.

Sharad Sharma 1:00:12

Golden Visa without any problem. So the point is that we are in a competitive space, we have to have a national effort to fix this and fix this very quickly. And if you don’t, then no matter what plans get made, you know, they’ll not translate on the ground.

So the four big learnings to summarize before I answer your question from our globalization of DPI, because that gave us an outside in perspective is that, okay, we do well with the technology architecture, you know, protocols and platforms and how to think about them. I would say we get an A there. We do pretty poorly B- for the regulatory architecture on top of that.

 

We do our ecosystem building organizations, which are supposed to give us scale, you know, especially, you know, 10 to 100. We have a mixed bag of organizations, we have to do much better.

 

Siddhartha Ahluwalia 1:11:14

We have a B-there.

Sharad Sharma 1:11:16

We get a B- there and EODB, we get a D-.

So the table stakes is that we got work ahead of us, right? We have got some traction, but we also, you know, need to do a lot of work. Now, when we compare this battle between US and China, which you pointed out, I think US is doing very poorly.

Now, of course, in EODB, they’ll go from whatever their current score is, they’ll jump two levels, because they’re going to make a serious effort. And, you know, and some of it will certainly succeed, right? And so that’s a good thing.

But that may come at the cost of very poor regulatory architecture. Right? And, and that regulatory architecture, because if the job is to, instead of smart regulation, do no regulation, that is not a better outcome that we are talking about.

And in the rush to show progress, they add that is what they may end up doing. And that has implications. In fact, you know, we should realize that has had implications even in the past.

So about 23 years ago, had they not challenged the monopoly of Internet Explorer, I can tell you, Google would not be an important company, it is today, you know, you won’t have had Facebook as an independent company that you have today, you know, even the current big tech ecosystem that they see is all a result of being able to challenge the monopoly of Internet Explorer.

Siddhartha Ahluwalia 1:02:47

And can you share more on that?

Sharad Sharma 1:02:48

If Internet Explorer at that time had 90% kind of market share, right? Now imagine for a moment, then you had to show some ad, they would show up on a browser. So Google would have been merely a tech supplier of the advertising infrastructure, because you couldn’t actually, there would have been no Chrome, there would have been no other browsers, it would have been controlled by Microsoft itself.

Siddhartha Ahluwalia 1:03:14

And what did US government did to?

Sharad Sharma 1:03:15

So they filed an antitrust case against Microsoft. And that forced Microsoft to give up its control over the Internet Explorer. And that is how the browser market is no longer controlled by Internet Explorer anymore.

And at that time, Facebook, you know, when Facebook became a unicorn, a billion dollars in valuation, Microsoft had owned 15% of Facebook, they would have done to Facebook what Facebook did with Instagram, saying, sell me or else you have no future. They would have pushed a sale, and there would have been no independent Facebook that you see today. And so the reality of this is that there are very powerful forces that cast a shadow on innovation.

So I’ll give you examples. The very powerful Wall Street, poorly regulated, forces good companies to embrace financialization and lose their way. And, you know, and I think Sridhar Vembu talks about this beautifully how Boeing lost its way, not because somebody else did it to them, because they became it was run by bean counters.

 

The same thing happened to IBM. IBM is a shadow of its former self, because they did not invest enough in innovation and R&D. And why didn’t they do that?

Because financialization took hold. So who drives that? It is the Wall Street that drives that.

 

So that casts a shadow on the US innovation ecosystem. The second shadow comes from the military industrial complex, you know, where they’re focused on big projects, not the small projects, the F-35s, right? And anything small doesn’t get their attention at all.

And I therefore fallen behind, you know, as we’ve seen from the Ukraine war. The third, of course, is the big tech, you know, big tech had a well-known, although they won’t admit it, but the fact is that they had an organized effort to restrict AI to their own set of people. They said, let’s erect entry barriers so that more people don’t come.

 

Right? This is, this is, they will never accept it. But this was a reality, right?

This, I would say is still a reality. And so therefore, the US is not letting the Darwinian ecosystem operate in many new areas like it should, because the incumbents there have become very powerful. China to its credit, even as a communist country, is letting the Darwinian ecosystem operate, right?

The BYD is a result of a very Darwinian ecosystem, you know, that has thrown it up. And that’s, by the way, true for almost any area that you can see, that was true for Huawei and ZTE, that is true for AI, that’s true for many areas that they’re talking about. Yes, of course, once the national winners, once the winners are there, they make them national winners, help them travel the world.

But to get to that stage, you know, you have to win on your own merits. And we got to make that happen as we go forward. So I think our lesson, the innovation model that US brought to life, which is often called the Vannevar Bush innovation model that you have some public research or public goods or public infrastructure and private innovation on top of that is a great model.

But today, I would say the US is implementing it very poorly. And others, especially China is implementing it much better. So I would believe that we have to make sure we don’t, when we take lessons from the US, we don’t take those bad lessons, we take the good lessons.

And similarly, we take only the good lessons from China and not their bad lessons, because both sides have some lessons to take and craft our own thinking and approach to take this forward, which ultimately is not that complex. It says you need research, especially pre commercial research, because no company in the world actually does pre commercial research on their own dime, public money is involved. And therefore, we have to enable that to happen.

That doesn’t happen today. Because CSIR, I’m sorry to be so blunt is defunct. Our academic research is mostly hobby research.

And so we have a problem, we don’t have a research infrastructure. We need our market players to focus on on commercializing pre commercial research much more aggressively than what they have done in the past. And we need policymakers to build the right policies to facilitate especially leveraging our domestic market in the right way.

Right. And, and, and that is particularly hard, because the government has two sets of people that come to them, the assemblers, and the original makers, that both are ours, we should favor who amongst the two? So for them, it seems to be, you know, choice between right and right rather than right and wrong. And we got to learn how to be more, more discerning about those choices, and make the right ones, we haven’t made them very well in the past.

And so, so we can only do that once we know what our goal is, our goal is that in this area, we got to have strategic autonomy. And, and how do we have strategic autonomy, there has to be a game plan, that game plan has to be understood by the policymaker by the market player by the researcher, all three have to understand the same game plan. Now, your question is, do we have a game plan like that for AI?

Right? And is it understood by all the three stakeholders? And the short answer is not yet, we don’t have a game plan, we know some elements of the game plan.

 

But we don’t, we know those elements are necessary, but not sufficient. And I can talk a little bit about that, if you like.

Siddhartha Ahluwalia 1:09:22

Let’s let’s continue our conversation on right, the various legs that our AI progress has to stand on one is regulation, what are the legs and how do we go direction?

Sharad Sharma 1:09:34

I think one learning that we have is a very simple thought process, that ecosystems produce winners. Yes. Right.

And we have to build the right ecosystem. And that will produce winners over time, right? I love what Subrata, for example, used to say for even Flipkart in those, you remember, Axel had many, many investments in, in eCommerce companies.

And he would say that, you know, till very late, he didn’t know Flipkart would be that winner. Similarly, I could say that for UPI, I remember last time, I talked about five companies that only two and a half agreed, but there were many others. But if you asked me that these will be the winners, it was not obvious.

In retrospect, you tend to say that these are the winners. This you as a VC, know this, right? Every investment that you make, you think will be a winner.

Siddhartha Ahluwalia 1:10:29

It’s the same condition.

Sharad Sharma 1:10:30

But only a subset of them will be a winner. In retrospect, you know, you can justify it, but in prospect, you don’t know actually which one it’s going to happen. So we got to build ecosystems, right?

I think one of the learnings from the last 10-15 years is that a good ecosystem is when there is coherence between the underlying tech architecture, market architecture, and regulatory architecture, right? Now, very often, we don’t control the market architecture. That’s a derivative.

What we control sometimes is a tech architecture, only if we can have a DPI there. If you don’t have a DPI, you don’t even control the tech architecture. And you of course, control the regulatory architecture.

So the question that is animating many of us is that, you know, if India has to have a game plan, if we are going to be a late entrant, we are a weak player in AI, and we have to have a play in AI, what is it that will get us there? Now, there’s some things that are clear, we must have good regulatory architecture, because we have bad regulatory architecture, we’ll screw this up. Luckily, there is good consensus on it.

Siddhartha Ahluwalia 1:11:39

Okay.

Sharad Sharma 1:11:39

And I’ll explain what that consensus is in a minute. And your listeners can even give feedback and say, you know, does this make sense or not? The second is, we understand that the data architecture or data economy will have a very big bearing on the AI economy in the future.

And we have an advantage because we have continental size data, you know, as a country, we because of our work in India stack, you know, we are data rich, while we may not be economically rich. So that is a asset that we have, which we can build on as we go forward. We know from a tech architecture perspective, that, that there will be, in addition to training data, there will be pre training models that will be there.

And then there’ll be post training that will happen. And we know that the market will split, and there won’t be one model that will drive all for example, you know, we know that the video editing AI for video creative is going to be very difficult for different from AI for customer engagement, right? You know, it will not be the same at the at the product level, it will not be same at the data training level, it’ll, if at all be same at the research level, where common research thinking is going, but it will be split at the research level, certainly data, model and applications, right.

And same as the case with health, there are going to be pockets of health, that will be a completely different thing altogether. Now, knowing all this, you can begin to decide where to play and where not to play, right? So for example, does it make sense for us to play in health?

The answer to that is definitely yes. For a number of reasons, because, you know, let’s say we want to find a way of using AI to do cancer detection earlier. Then we have the training data, because India maintains cancer registries, some cancer registries, especially in the southern states are very good.

And if we can find a way to unlock that data, and give it to people who will be the modelers for that data, we would have more data than almost any other country. Furthermore, this is one place where there is difficult regulation called HIPAA in the US that will prevent US startups being able to do well. So is this a battle that we should try and win and be in that circle of three or circle of five?

The answer is unquestionably yes, right? So we have to be able to pick our battles. Similarly, I would say, in many corporate use cases, you know, will we be able to win?

I don’t think so. For example, Salesforce has got such a lock in on customer facing kind of corporate use cases. And since our way of thinking about those use cases is more or less the replica of the American use case, we should, we will struggle, right?

So we will not have the same potential dominance in that area as would be the case in health, right? So first is we have to do a triaging of where is it that we can win? And where is it that we may not win and not that we don’t try in those areas, but the ones that are lower hanging fruit, we should go for them in a very big way.

Now. So, so clearly, the area that is indeterminate, and, and I’ll put this on the table, some, some of us believe that is a worthy area to go after. And many people believe it is not.

And maybe next time I come, we’ll, we’ll have a conversation about that. Is that look, it is a fact that most individuals will rely on AI to book, let’s say, the restaurant in the evening to open a bank account, you know, it’ll be the equivalent of, let’s say, what, what Google Assistant is, should we let Google Assistant and WhatsApp dominate that? Or is it possible for us to take a different approach, so that Indians are served with AI assistance of India?

Yeah. You know, even though they may be running on Android and, and, and iPhone. And, and there’s some of us, you know, Umakant Soni is one of them and a few others believe that we have a game plan, we can have a game plan to make it happen.

I’m not prepared to air it today, because it is still evolving. There are daily meetings that are taking place to give it shape so that we can move very quickly, because we have only 12 to 18 months to get that right. And, but, but that’s an example of looking at it and saying, you know, on surface, it may look that we can’t win.

But if we follow this game plan, then we have a chance of winning. So let’s put everything that we have to make it happen. Now, what’s everything that we have?

Everything we have, of course, needs compute. Luckily, for us, we are getting compute, the government is very supportive of making compute available, we already have a line of sight to 18,000 GPUs, if we need, I would say even 10 times more, I think the government is willing to step in to make that happen. We have a we have number of good people, right, we have a wonderful minister, who understands this, we have a mighty secretary who’s exceptionally supportive.

You know, we have an AI mission led by Abhishek Singh, you know, himself an engineer, very, very knowledgeable, you know, one of our best officers. So I think, you know, we can, we won’t, we won’t have an advantage with GPUs, but they are strongly committed to taking the disadvantage away to give us more or less a level playing field, you know, even if it takes public money to be brought in to make it happen. And I think that there’s a desire to implement that well.

 

But that will give us table sticks, right. And but where will we get the edge, we will essentially get the edge with talent. And we will get the edge with with data.

So with talent, we have gaps, we have very good Indians, you know, NRIs in every place, we need to bring them back. Now to bring them back, we need to bring them back in four places. So for AI talent to work, you need almost a talent tree.

The talent tree is that some people need to come back to be founders of startups. Some people need to come back to be not founders. They say I would rather be the guy who’s the engineering guy who delivers the damn thing.

I don’t want to be the founder. I don’t want to be the founder CEO, I want to be on the engineering side, then we need people who come back for industrial labs, type R&D, and then academic type R&D. So if you’re really looking for 10 people, they’re roughly broken up in these four categories.

Right now, we don’t have an industrial lab here. One of the reason France, there’s optimism about France in the future is because Yann LeCun, you know, he ended up establishing a Bell Labs equivalent for AI, which is called FAIR, Facebook AI Research, which has done very, very well to the point when he was here, you know, he spent, as you know, a whole day with iSpirit in Delhi. And he explained that, you know, his legacy will be less about the, it’ll be, of course, about the technical contributions he’s made for which he has a Turing Award.

 

But he believes his bigger legacy would be the institutional structure that he created, and the fact that FAIR is working very well. And because he’s from France, he ended up creating in France, that’s a tremendous asset that they have, right, which is a gift that they got from Yann LeCun, right, and Mark Zuckerberg, right, we need to create that ourselves and our big companies like Reliance, you know, our leaders there are very interested in doing that. You know, in Tata’s, they’re very interested in doing that, and others and I’m hoping that they’ll make the right big moves to create these kind of industrial labs for that.

Our academia also needs an upgrade, right? While we have very well-meaning people in academia, you have Ganesh, in IIT Mumbai, you have, you know, you have Professor Ravindran in IIT Madras, and so many others. They are there, but we need to up our game in academic research as we go forward.

And IIIT Hyderabad is trying to do that. So we have some crystal formation that has happened, but nothing that has become significant. So talent means that we should be able to draw people back.

Today, a startup is able to attract a good guy to come back, either as a co-founder, or as a CXO, you know, CSO or CTO or Chief Engineering Officer. But we don’t have places where we can get the other two people back. So we have to create these places that are run so well, that they actually are willing to come back into those setups.

So we call them corporate FROs, Corporate Focused Research Organizations, and academic FROs. And we need to be able to create these new type of FROs, because they don’t exist today. Today, somebody is not going to come back from MIT to be in even in Institute of Science.

Why? Because the life is very different, right? And so we need to create a pocket in ISC, which is part of ISC, but is independent enough to create its own culture to be able to take that forward.

That has happened. Institute of Brain Science is an example of that. So it is possible to do that.

But we have to do it at scale as we go forward. So I think we need to roll out the red carpet for high quality AI researchers of all these four types, because they’re all in some way or fashion, involved in shaping the ecosystem and get them back. We must have a red carpet for them as we go forward.

So that’s the talent part. And I think if we play our cards right, we can do that. So that’s one part.

I want to spend a little more time about data. Remember, the first version of AI is on so called public data. And that made everybody had the same access to data.

So if you’re sitting in Dubai, you had access to the same data than you had in the US. And therefore, you know, you could build Falcon as a good foundation model, you know, like what Mistral can do or what, you know, open AI was able to do and others are able to do. Because the data was more or less the same, then it’s a question of applying the right techniques and throwing the right compute.

So it was all about who has more compute, who has more gumption to throw compute. But now things are changing in that changing because either the research is telling us that you don’t need as much of compute or you can get more bang for the compute that you have. And we need that kind of thinking.

Or the new models that are coming in are based on data that is not in the open domain and may never be in the open domain. So for example, if I want to do something for cancer, you know, that cancer data is in the hospitals. And because of privacy reason will never be in the open domain.

That doesn’t mean that me as a startup can’t go to hospital one hospital to hospital three, and sign a contract with all of the three to be able to get privileged access to that data. Great, but I’ll have to be a reasonably well funded startup to be able to do that, right? Let’s say three startups are able to do that.

And nine such contracts have been signed. That’s great for the lawyers, because they made lots of money in helping make that happen. And, and let’s say that system works.

 

But what then happens is that you actually put a middleman in place. And that middleman comes and says, look, I will sign the contracts on your behalf, so that you have to only sign with me. Right?

Yeah, that’s more efficient. Because if the three startups, they all sign a contract with the middleman, and the middleman signs three contracts, only six contracts are to be signed. And if you want to scale this, of course, this scale is much better than the bilateral model that we talked about earlier.

And this is starting to happen. Databricks, Snowflakes, many other companies are starting to do that everywhere in the world. But what will happen is that if these companies are profit maximizing entities, they will suck the value away from the edges.

Because if you on the other hand, make them a public utility, then the value is either with the person who had the data in the hospital, or with the person who made meaning of that data, which is the startup, you know, and the middleman is just a middleman. Right? Now remember all this conversation that we had about regulatory approaches, this kind of an approach becomes essential.

So the Indian thinking is to put DEPA in the middle, as not only as a utility, but also one that is data blind. Because by doing both, you are able to create a trusted low cost system that unlocks data that would never be in the public domain, but makes it available to startups as we go forward. So that’s the story that is there.

Now to make it happen in a way that it will be a public utility and it will be data blind is not so easy. Right? And so to be able to do that, that middle entity needs to be a sophisticated user of a new branch of mathematics called differential privacy, you have to be a large scale user of a new area of confidential compute, which is hardware based confidential compute, you need to be able to introduce payment mechanisms and modalities and electronic contracts that can be created on the fly as you go forward. So it requires amongst many things, just these three things coming together.

Now luckily for us, in the last four or five years, this has all come in place. And this is now in the zero to one stage, where there are closed pilots taking place. And not just in India, but now in Japan and France as well.

And so therefore, you know, anybody can look at this, if you really want to learn what it is, go to depot.world, you will see inference and training, go to training. On training, you will see getting started. In getting started, it will take you to the GitHub code that will allow you to use that.

Now, I must warn you, it’s not so easy, because you need to know differential privacy, you need to do confidential compute, but it’s available for you to use it. And people across the world are able to use it. Because they’re sample data sets, you can actually get a hang of this as you go forward.

But I think over time, it will become easier to use the claims that are being made about it being confidentiality preserving it being privacy preserving will be more proven. And it will be easier to scale this as we go forward. If you are able to do this, then India has ability to unlock non-public data sets at a scale that perhaps no other country in the world can.

Right. And, and that may then be an impetus for, for whoever other model builders in India. Now, of course, this means that we should give some preferential treatment to Indian model builders over other model builders.

And one of the significant criticisms that many of us face is that we did not do that for UPI. Right. And at that time, we were age blind and colorblind.

And which meant that if you’re a small startup wanting to be part of a close pilot or a big company, we didn’t care. And whether you are Indian or non Indian, we didn’t care. Right.

So age blind is startup versus big company. And colorblind is, you know, in brown or white, right. And, and, and I think I would confess that this is not a settled issue, even with the nice spirit, there is a case to preserve that old thinking.

But also, I would say it is trending towards saying that we should be age blind, but maybe not colorblind. Yeah, when it comes to this new infrastructure that we’re trying to create. But the mechanics of how we’ll make it happen, we have to build a consensus with the policymakers with everybody else.

So, but I think we’re moving in the right direction. By end of this year, we’ll have a fully finished zero to one transition. And that will be very similar if you go back in time to where we were with Oaken.

Oaken 4.0 is the first time Oaken’s been made available to everybody. Now 4.0 means that the pilots resulted in many improvements that came along to result in 4.0. 4.0 is the first time it’s in the open domain, till then it was all closed pilots. Similarly, in the account aggregator it was in September of 2021 that it became an open system, till then you could only do it with closed pilots. So in DEPA we are in closed pilots, but I encourage anybody who feels that they can make the getting started piece run, that is a precondition, and they’re interested in being part of the closed pilot, come and join this.

And we are very keen to do this with other strategic countries. Our hope is that those strategic countries will be Japan, France, but we are also very keen on Vietnam. And so we call it a three and a half kind of thing, India, France, and Japan and Vietnam.

Vietnam is the half there. And that kind of consortium will kick off. We have a lot of good support from some of the big tech companies.

These are not the traditional big techs, and these are companies which feel very responsible on how they treat their customers’ data. And so they want to optimize for trust, and some of them are going to announce in the coming months that they will follow our system, DEPA, on a global basis, even in jurisdictions that don’t require them to follow, right. And so that’ll be a big win for us, and we are quite excited about that.

So I think coming back to AI, to summarize, we’ve got to triage the problem, pick the places where we can win. Second is figure out what would be our regulatory architecture, which I’ll talk about, figure out what will be our data architecture, make sure we are not disadvantaged for GPU, and finally bring a new class of talent back into India. Because frankly, we need super smart people coming back to drive this India AI regulation.

We have super smart people, but we need more, right, because we need various kinds as we go forward. So this is not to ding the people that we have, but we need more as we go forward. So this brings us back to regulation.

Now regulation, the Indian thinking which has now been put out, there was a regulatory committee that was formed and has been there after much consultation within the government. Government has not formally agreed to that approach, but they have now put it out in the public domain. The Indian approach is the following.

So Indian approach relies, A, on techno-legal regulation, B, on this concept of an AI chain. And the concept of an AI chain is that in this AI chain, data starts as personal data, which increasingly will be in our DBDB system. Then it becomes aggregate data.

Then it goes into a model, model becomes an app, app creates an artifact that gets distributed. And so this chain is there. Now in this chain, different countries have taken different approaches as to where should they regulate.

So for example, in Europe, they said, we must regulate the app. App for them is like a car. So they say, we’ll not put a car on the road until it is safe for its occupants.

And we can’t trust the carmaker to do the testing to determine whether it’s safe for the occupants. So it should be third-party testing. So they’re saying the app should be third-party testing.

The US, which has tried to control AI and its dissipation, you know, Sam Altman’s statement was just a manifestation of that. And the GPU diffusion order is a manifestation of that, said, look, ultimately, all these apps will have only a few models. You know, there’ll be a handful of foundation models.

So let’s only regulate the small number of foundation models. Because which app do they go into, what difference does it make? So let’s just regulate the engine, there’ll be hundreds of cars that will use the same three engines.

We don’t have to worry about them, because we are regulating at a deeper level at the engines. And if they’re smaller engines, let’s not even worry about them at all, right? An analogy would be, let’s only worry about combustion engines that go into a car.

But if you have small combustion engines that go into a leaf blower or in a small lawnmower, let’s not worry about them too much, right? So that is their strategy. Our view in India is that we don’t know how the ecosystem will evolve.

There’s a lot of unpredictability. Therefore, we must think of it as a complex adaptive system and retain the right to intervene wherever it makes sense. For some years, it may make sense to do it at the app level.

Some years, it may make sense to do it at the data level. We don’t know. In some cases, especially for children, it may make sense to do it at the distribution level, right?

So what we are saying is that really our effort, rather than bogging down on where to intervene, is to create this AI chain and make it visible, at least to the regulators, so that they can do pointed and minimal regulation at any one point and get the job done. Now, we have an analogy with that, actually, with the stock market, right? So there is a way to do that.

You want to control insider trading, you do it in a different way, and so on and so forth. So we know that we can learn from those complex adaptive system design regulation and keep ourselves open-minded as long as we bring the AI chain to life as we go down. And if we can do that by having electronic contracts come up in a big way, that would be very helpful.

Electronics contracts are a gray area, but some of it is going to get sorted out. You know, the GIFT City people, they formed a committee for asset tokenization, which I chair, where one of the focus areas is to get the regulatory approach to contracts sorted out. So hopefully, things will come in place as we go forward.

So if we can take smart regulation, which is very focused on the harms that get created. And initially, we are saying, don’t worry too much about adults. Focus on children, right?

Because I don’t know, your viewers may be aware of it, but India had 23 million babies or so last year, right? And China had 9 million. If you take the next four, five countries, they total up 23 billion.

So we are, in some sense, responsible for humanity. You’re the largest number of people under 15-year-old. And while AI will lift their performance, it will also can expose them to things and ruin their life.

So we have to take some agency away from the children and give it to the parents. Because if you were in the West, we would give it to the teachers. If you were in China, we’d give it to the state.

But in India, you have to give it to the parents. So we have a particular thinking about that, and age tokens, and many other things are already being thought of or in implementation right now. So that tells us that that kind of children regulation is happening at the tail end of the AI chain, right?

And we know if we create a DPI-based data economy, we’ll be able to preserve privacy at the early part of the chain. So we may not have to do very much in the middle at all, right? That is a possibility.

I won’t stand here and claim that is the right way to do it, because we don’t know. But that is quite possible. That by looking at the bookends on the distribution side and at the data side, we may allow for a fairly liberal system in the middle.

And I think there’s reasonable consensus that is building around that. And Professor Ravindran was the chair of this committee that spent a lot of time, lots of healthy discussion, and some convergence that has emerged there. So I think coming back to AI, we have to now turn this thinking about the tech architecture, about pre-training, post-training, about data architecture, which is also part of the DPI, about the market architecture, which areas to focus on double down, which areas to take a shot at, not knowing whether we’ll win.

And some areas in the triaging may not make sense at all. And regulatory architecture, how do we bring coherence in all of this to make it go forward? And I think we are still four or five months away before all this will coalesce into a common form.

But I think a lot of discussion is happening. Lots of people from outside India, particularly, I would say in Europe, it’s France, in Asia, it’s Vietnam, and Japan are paying a lot of attention to our thinking. And in this case, we may have others who may go along with us for the ride, with me playing a significant leadership role there.

Siddhartha Ahluwalia 1:39:40

I believe what you said, if you’re able to identify top 100 Indian origin brains in AI in the US, for example…

Sharad Sharma 1:39:49

And bring 30 of them back. And I think we have to do that, because that red carpet treatment has to be given to these people. The problem that we have is, we don’t have landing spots for them.

And we have to therefore create the industrial labs. And if I was to signal to people, I’m reasonably confident that will happen. And I would say, remember, based on a larger discussion that’s been happening about revamping our R&D infrastructure, last January, I remember, January 4th, we had a meeting where we had some of us, that included Kris Gopalakrishnan, who is in some ways, the deepest thinker about the R&D infrastructure in India.

Because he put all his philanthropic money to work there, right? And he is also the chair of the governing board of ISC, also the chair of the governing board of IIIT, where MOSIP is based, and you know, it’s a big success story. And he is deeply reflective, it’ll be hard to convince him to come for a podcast, but I think you should try.

And you know, he is also the guy who likes to work in the kitchen, rather than talk about, you know, what he’s doing. And so people like him, you know, we went and said, look, so much experiment that has happened with philanthropic capital has to move into public capital, and perhaps we need to allocate 10,000 crores, then we were told that’s too little, it should be, so we said, okay, 25,000 crores is what you should do an experiment with. But over 10 years, you should budget for 100,000 crores.

And this was on 4th January, 1st February, they allocated 100,000 crores. So there is an understanding that our future is going to be based on what we do on R&D infrastructure. See, what we were able to do as digital infrastructure, and that has become an asset.

Even for AI or anything else that we do, even in areas outside of all this, it has given us the self-confidence of being able to do it outside the ISRO and the BARC model. You know, we now need to spend the next 10-15 years strengthening our R&D infrastructure. And I think there is a serious effort that is taking place.

It’s still very early days, but we are trending in the right direction. So I think, you know, so the way I look at it, is that if we do five things well, right, one of them is based on you, but the other four are based on others, then we could have a deep tech system that will really be in the golden era. So one of those things is, of course, the risk capital itself, right?

That requires you to go out and secure money from the LPs and then deploy it successfully. You know, and I really appreciate what you’re doing. You’re already in AI, you know, the second most ranked, you know, VC firm, right?

Is that right?

Siddhartha Ahluwalia 1:43:12

Yeah, by Economic Times, we were considered the 2nd.

Sharad Sharma 1:43:14

So clearly, you know, Godspeed to you, and may you do this at larger scale, and we need many of you, frankly, right? So that is one.

Second is that a culture has taken root, and which is that our entrepreneurs share knowledge with each other.

Siddhartha Ahluwalia 1:43:32

Yeah.

Sharad Sharma 1:43:33

And I would say that peer learning culture is growing stronger and not weaker. And in the original place, which is where it came, and we talked about it last time in my podcast, you know, where people like Girish and Pallav Nathani and Shekhar Kirani and others had, Avinash, of course, had a role to play, they continue to make it work in their core area, but it has spread to many new areas. So I think that peer learning is very important.

Third, that we have had, we have to fix EODP. Because if we don’t fix EODP, right, you know, all these efforts are wasted, right? So we’ve talked about this.

Fourth, our DPI approach matters. And we got to continue to build on it, especially for things that are to do with digital. So quantum will benefit from this, AI will benefit from this.

And of course, we have unfinished agenda in the core areas like health, digital health, digital finance, many other areas, digital education. So there’s unfinished area. And the fifth is that we got to get the R&D infrastructure in place.

And I think for the first time, there’s a recognition of the problem, right? And there is money available for this problem. We got to make sure we get it right the first time because time is our enemy.

This is next few years, our make or break for India. Because as the supply chains are up in the air, we can find a place in this world, you know, by dominating some parts of the supply chain. If we don’t do this in the next four, five years, you know, it will become exponentially harder for us to do that.

And then we would remain become just a kind of a services nation. And, you know, there is, it’ll be like Thailand, it will not be like Korea. And so we have to be, you know, we have to give it everything.

Siddhartha Ahluwalia 1:45:30

And you have given last 20 years of your life to making India a product nation.

Sharad Sharma 1:45:35

Yeah, I mean, I’m not the only one. See, so I think what is happening here and so therefore, one of the learnings is that there are, so far, this revolution has been driven by people who are willing to be sutradhars for this journey, right? Now, many of us were lucky that we could play the sutradhar role without demanding any money back.

And I think people like Chris, for example, you know, I would say in the geospatial area, it’s Lalitesh, you know, and many others. In space, it’s Pawan Goenka. Many of the sutradhars, they also are not even drawing attention back to themselves, right?

They are, I would say, no greed, no glory, right? And that’s a good way to do this because you are only an instrument to doing something. You are not the story.

The story is what happens to the ecosystem. You are merely a sutradhar to make it happen. But we need many more sutradhars and our thinking is we need 15, 20, 30 more sutradhars.

 

So we’ve taken a big leap of faith. He said, instead of restricting the sutradhars only to people, first of all, you can’t be a sutradhar in your 20s. It’s very difficult, right?

Because you have to deal with complex things. You have to draw many stakeholders together, right? You’ve got to take a whole of nation approach.

So you’ve got to be in your mid 40s to be a sutradhar or older. So right now, we are restricted to people who by that time have been financially successful to say, okay, I don’t need more money. Yeah, I’ll do this on my own.

And that is restricting us to a very narrow set of people. So we said, why not think like Singapore and say, if this is an important thing to do, find money to pay them. Yes.

Not pay them an upside, but pay them a fair market salary to do that, right? And luckily, as this idea took root, we had phenomenal support for this. You know, I remember a conversation where this was, you know, we’ll need 50 crores to do this.

And, you know, one very meaningful person said, look, I personally promise that I’ll arrange 1000 crores for this, right? So because people understand this has a very, very high impact. So I would mention to people that they should go and check out something called vigyanx.one. Okay. Where we have just put a call out for sutradhar. Okay. We’re trying it out only in three areas right now.

And those three areas are power electronics, AI compute, a narrow version of that, and to do with fabulous design. And we’re going to experiment, try to get some sutradhars who are in this category, who are not, you know, not post-economy, which means they’re living on their own, but they are getting market salary. Of course, it is, the bar is very, very high.

It’s very, very selective, but it’s got a great set of search people involved. You can see and on that site, and we’ll start putting out more videos as we go along. But, you know, surprisingly, we got 16, we got many responses of which 16 look very promising.

So I don’t think we are there yet, but love to hear from more. And we want to even look for people who are willing to come back and do this, right. And so the goal is to, we need a sutradhar who makes zero to one happen as a very big role to and then remains engaged from one to 10 in a slightly less active way, but is still there for the 10 to 100, because we need that continuity for 10 years.

So we need people who will make a 10-year commitment. And, and, and so, so really, you know, to take this journey forward, here is the ask, if you are in your mid 30s, come back to our next generation R&D centers that are going to be created, the FROs that we talked about. We need people in mid 30s.

These are people who would be near what in these centers, the role, the lead is called the PI, principal investigator. So we are looking for people who would be somewhere near that. And so we want principal investigators to come like that.

We know this can work because NCBS here in Bangalore did that. And some of the people who made that move are part of design of this thinking. If you’re in your mid 40s, consider becoming a sutradhar.

If you’re in your mid 50s, consider becoming the leader of these ecosystem organizations, because you have to create many more. And we need people who are who have some gray hair, they know how to build institutions, and good institutions. So this is a call out in this mission for three types of people.

And of which one is a formal process, the vigyanx process, the other two are still informal. And, but in the past, we’ve never done this. We have always looked for it, you know, through organic model.

But I think time is of the essence, right? And I think we have a good chance of winning. And therefore, we must…

We put all resources… By the horns and no longer, you know, do it the old way, you know, do it the new way. So that’s really, in some sense, the opportunity before us.

As I said, our digital stuff is doing well, on the back of it. You know, we’ve made good progress on telecom, I talked about it last time on neurogenerative diseases, you know, where you know how drugs are made is changing completely what used to take years is now going to take months. We are beginning to do well in some pockets there, not in all pockets.

But one of those pockets is neurogenerative diseases. And, you know, we’ve built up from research onwards. And that looks very promising.

We have a geospatial area that is looking possible, and very attractive. And we have a very good game plan for quantum. It’s not yet reached zero to one.

But the game plan is quite credible. And so these five areas, but there are many, many more, you know, if you allow me, I can tell you how many more. But I can give you examples.

And I want to infect people’s thinking here. If you are okay, can I take a few minutes to do that? You know, India’s bet big on solar plants.

Now, we’ve such large solar plants in Gujarat and Rajasthan that we have a secondary problem. The secondary problem is we have to evacuate power. To evacuate power, we need HVDC systems.

Siddhartha Ahluwalia 1:52:48

What do you mean by evacuate power?

Sharad Sharma 1:52:50

Evacuate power means we have to move power, electricity from Gujarat and Rajasthan to the rest of the country. Right? Because they have surplus power.

And soon we’ll have very large. So to move that power, you need high voltage DC systems. We want to build this HVDC system without Chinese power electronics.

Is that a reasonable approach? Yeah. Very difficult to do.

So we, as a country know that we have to basically encourage our own ecosystem to create the companies that will build the high voltage power electronics that is needed because we need it in India. If we need it, the world needs it, right? And so what does it mean?

It means that you don’t focus on the current generation, you focus on the next generation. That means R&D. That means startups that commercialize that R&D.

It means that, you know, the country buys only local rather than foreign, which is what we did in telecom as we go forward. So we know we can pull this off. We have a playbook for it.

And now we need to make it manifest and bring it to life. You know, India is rolling out, you know, one of Prime Minister’s pet project is this rooftop solar. And you know, the way this is designed, you know, the goal is to get to 10,000 installations a day.

Siddhartha Ahluwalia 1:54:14

Okay.

Sharad Sharma 1:54:15

You know, and they have his handpicked people to do that. And I know within two, three years, we’ll get there. But to get there, you know, you need power electronics in the house.

Today, the power electronics in the house is good for five KVA houses. We want to do it for one KVA houses, which means nothing exists. Again, there is a willingness to take a long term 10-year view to create a power electronics, low voltage power electronics that will serve India.

And if it serves India, it will serve the world. Right? Now, these are tangible.

These are all high tech things. They’ll need research. They’ll need deep tech startups.

They’ll need market making to happen. They’ll need the right policy. They’ll need a 10-year view.

But I think we have the playbook to do that. And so, therefore, we are even looking for a sutradhar to make it happen, right? Or sutradhars to make it happen.

So, the question is that these kind of opportunities are there now more than a dozen. So, and why the confidence is growing is because many of these building blocks, the five things that are talked about, which is VC funding, you know, the fact that peer learning is happening, competitors talk to each other and learn from each other. The fact that we have to do EODB hopefully will happen, at least for these people.

They should be given special privileges to deal with that. And, you know, with respect to wherever we need DPI, and then wherever we need research, and especially pre-commercial research. So, I think that playbook is coming together.

We need the best people to go and attack these most important problems. So, don’t go and build another marketplace. Come and solve for this.

You know, one person you should bring on your podcast to talk about this is Mukesh Bansal. He has his own podcast. But Mukesh is an example of saying, look, you know, he’s built two very valuable companies.

Myntra, of course, he had a role in Flipkart. Furthermore, CultFit is a success too. But he is now committed to this deep tech startup ecosystem that we are talking about, where we want to get strategic autonomy.

In fact, our VigyanX effort, you know, he is one of the key volunteers driving this right now. So, look, we need more people to say instead of doing what are so-called cool sectors from a US perspective, instead of going after sunrise sectors of any and every kind, we must go for those sectors where India needs strategic autonomy. Because India’s future depends on us succeeding in strategic autonomy by creating the product companies in that area.

You should create product companies, believe me, in every area. But I think right now, for the next four, five years, call of the hour is to do it in strategic autonomy sectors. Is that fair?

Siddhartha Ahluwalia 1:57:20

Is that fair? And you think you have capital, enough capital to support it?

Sharad Sharma 1:57:23

I think the good thing about capital is if we get the story right, capital is most fungible part of this. You know, when we got the story right for FinTech, there was no shortage of capital. Capital will come.

We got to get all our other things right. And as we begin to do that, you know, as our sutradhars become more effective. And who do they have to do?

They have to work closely with the policymakers. They have to work closely with the researchers. They have to closely with the market players.

And they have to bring all of them on the same game plan. And that is not so easy, right? And first of all, you need a good game plan.

And to get everybody aligned is not easy. And then keep them aligned is not easy. Then you got to produce results relatively quickly so that you can build on that.

It’s an interesting role. And I think, however, many people are beginning to get traction on that. And as we do that, and we do it much more systematically, we will see more success.

Siddhartha Ahluwalia 1:58:26

I think if somebody has the vision of saying that these are the four, five critical areas where India needs to be independent, and we’ll create 100 new public companies in this in next five years. Now we are looking for talent and capital. And because there is a large outcome possible, come join us in this.

Sharad Sharma 1:58:46

Absolutely. 100%. And so this is really, and there are many, many such things that are underway.

As I said, geospatial is already underway. The data economy is already underway. But there are many more like HVDC, which need or power electronics that need to get underway.

And I think, you know, hopefully, we can seize this opportunity as we go forward.

Siddhartha Ahluwalia 1:59:11

Thank you so much Sharad Sir. I think this conversation, I loved our previous conversation, but I respect this conversation so much, much more, because it’s a call for action. Right now.

My job as a VC and through this medium as the podcast will be to get as much as talent possible aligned to this. So that we are a formidable force, you know.

Sharad Sharma 1:59:40

So let’s make India proud. Yeah, we can only do it together. Thank you so much.

Siddhartha Ahluwalia 1:59:44

Thank you. Thank you so much.

Sharad Sharma 1:59:45

Appreciate that.

Thank you.

 

Vector Graphic Vector Graphic

Know when new episodes are released. Subscribe to our newsletter!

Please enter a valid email id