Peter Slezkine: Arvind, welcome to the show.

Arvind Gupta: Thank you for having me, Peter. It’s nice to take this conversation from conferences to an actual one-on-one conversation, so I’m excited.

Peter Slezkine: Absolutely. And this time I’m in Washington and you are in, uh, India, so we’re—

Arvind Gupta: New Delhi, yes.

Peter Slezkine: New Delhi. Perfect. Yes, it’s good to have two cities, otherwise it’s an imperfect comparison. So next time we’ll record in person to really make this exciting. So you are the co-founder and head of the Digital India Foundation. Can you tell us a little bit about that organization and its place within India’s digital ecosystem?

Arvind Gupta: So, Peter, let me give you a background. I have been working on what is the Digital India project since its launch in 2015, pre-launch in 2014. And I spent a considerable number of years in that project. And this is a very unique project, and I will need some minutes to explain to you because it’s the participation of three or four key stakeholders that let this come alive. Of course, the private sector, the businesses, but also, of course, the government. And a lot of society at large, but also many projects which have been done in public-private partnership. So I was a founding member of an organization called iSPIRT. iSPIRT was responsible for working on one of the first big successes out of India called the India Stack. And I will explain that to you as we go. Digital India Foundation, post my full-time role in the government, we have been involved with all the technology policy and the issues which are current at that point in time. So technology policy, basically. But also through this role, I am able to sit and work with the public-private partnership that is building the new technology stacks in India. So a related conversation can be on what are the technology stacks, and something that we should note and do a conversation around.

So that’s what we do. We are a team of folks who work on technology policy, on digital sovereignty, on what is current and relevant to India, and sometimes a lot of policy disparities that occur between the digital world and the physical world. And lastly, we also do a lot of what is called digital diplomacy, slash digital outreach outside India, using India’s case study as a success story.

Peter Slezkine: You mentioned that this is an exercise in improving the public-private partnership. What precisely is your relationship to government? And here we should also mention that you were the national head of BJP’s Information and Technology Department from 2010 to 2017 and helped lead Narendra Modi’s digital social media and communication campaign in the 2014 elections, right?

Arvind Gupta: That is absolutely correct. So, you know, for about seven years, 2010 to ’14, Prime Minister Modi was aspiring to become the prime minister. I have been a politically right-wing leaning person, aligned to the vision of the Bharatiya Janata Party and of course to the vision of the prime minister. So I took up the challenge when the BJP was not in power and Prime Minister Modi was the Chief Minister of Gujarat to lead this new age thinking around how to use technology, social media, big data analytics at that point in time, to make it a competitive advantage for the BJP in the 2014 elections.

We can talk about it for hours, but precisely, we used technology to a level that we could create a very big difference in the 2014 election. And apart from many other things, technology and the use of digital and social media was credited with, you know, at least one third part of the victory of the 282 seats that the BJP won, which is an absolute majority, something that India had not seen for the past 30 years prior to that.

So the last majority government was 1984. In 2014, Prime Minister Modi got an absolute majority. The NDA got 310 seats plus, so that’s the alliance. But I’m a technologist at heart, so I’m a person of using technology, and that was a use of technology at population scale, using technology to make it a data-driven campaign, to use technology for making engagement with your supporters, with your workers, in a more professional and organized manner. So it’s a classic case of how technology was used to really gather a population scale momentum around the ideas and the ideology of the prime minister and the BJP.

Peter Slezkine: How did you use technology specifically, and what did you do that was innovative at the time? Was it mostly collecting data so the campaign had more information about their voting base? Was it mostly social media and getting the message out? What sort of tools did you use?

Arvind Gupta: There were so many of them, and it’s been years, Peter, but I’ll give you a broad example. And see, time passes, and you are an expert on Moore’s Law. 12 years is a long cycle, so people have probably been able to catch up, and in today’s world it may look like what you were doing then was basics and simple.

But from having organized surveys, that is research, to micro-research, microdata available at an aggregate level for each — India has about 1.1 million polling booths where elections really are fought, won and lost. So from knowing their micro dynamics and then creating content, knowing your set of audience there, what you’re talking, in what language you’re talking, what the issues are that are pertinent to them. To helping that data in candidate selection, to making the appropriate candidate for the particular constituency, to using it for communication, to manage the elections, and then using digital and social media for, of course, mass communication, but also party communication, to coordinate better with our volunteers and the workers.

We used digital to crowdsource a lot of ideas as well as money. And this was — the digital part of the campaign was self-financed in a way that it was funded by the crowd, and this happened for the first time in India. I still remember way back in 2010 when I was going to set up a payment gateway for the BJP, the payment gateway company came to me and said, “What is your refund policy?” And that was new for them, that a political party was asking a payment gateway. So for us, it was a method not only to collect money, but also to collect volunteers, and also see where our volunteer base is coming from, so that was analytics. So to answer your question, I can go on and on, but you use tracking analytics, a lot of big data analytics, in terms of what is happening both explicitly and implicitly, build on those trends, understand the pulse of the people. Very well. And then of course communicate back to them in what mattered.

And at the end of the day, in this, we had a product which is the ideology of the party, the promises of the party — we are in opposition — and the track record of the BJP and of the prime ministerial candidate who has now been a prime minister for 12 years, how he had performed in his home state. So it is a mix — it’s a classic case study. I mean, 10, 12 books have been written on it, and how it was made a ground-up campaign, from the base up, owned by the people, and became a people’s movement, and technology played just a huge role. And that was my contribution because I believe that to do such a big change, you needed to use technology. Otherwise you do mass campaigns, you do TV campaigns, outdoor campaigns, newspaper campaigns, and they are just one-sided. So this made it more engaging with your voter base and your audience. So I think to answer your question, that was the origins for me to see how technology can be used at a population scale.

And some of that thinking really, Peter, came into the Digital India stacking that we did, the India Stack, which is the first population stack that we made. And when I say we, this is where I explain to you that there are multiple people who contributed to it. There’s an official government program called Digital India program. There was a set of volunteers who are making this in public-private partnership. So I played multiple roles both inside the government and outside as a set of volunteers to define what would go into the stack, and how should it be used. What are the use cases? So once you make technology, how do you use it in the government and become the champion of using it? If you remember, in 2016, India did — on 8th of November, when the US elected a new president, on the same day, India did demonetization. That day our FinTech stack, which is UPI, became really popular post that.

Peter Slezkine: What does demonetization mean in this context?

Arvind Gupta: Well, what the prime minister and his leadership did was we had a lot of issues in terms of fake currencies and unaccounted for money in India. India was given a task for 58 days or so that we will basically change all our currency. So all high-value currency became not legal tender, and you had to change all currency through the banking system. So it basically flushed out all the unaccounted for black money, fake currency, and you had to put money through the banking channels back out. And that cleanup process of paper currency going back into accounts and getting new currency gave an impetus to the digital movement in terms of using digital currency.

I still remember in that November 2016, we probably did on the digital payment side maybe half a million transactions a month. That was the beginning of it. It probably became a million in December, and then in March we did a couple of million transactions a whole month. That UPI, which is part of our India Stack, and I will talk about it at length — last month did 22 billion transactions in a month. So what was a couple of million is now 22 billion transactions, which shows you the magnitude of how it has gone up. The idea that digital can of course move forward changed a lot of it also, but the fact that digital can be a very consistent, easy, non-friction form of payments was seeded at that point in time.

Peter Slezkine: Well, we’ll get back to the evolution of the digital ecosystem in India, but I want to quickly continue the subject of the use of digital technology in elections. So your effort in 2014 was clearly very successful. But at least from the Western perspective, it seems like there was a very interesting flip in the discourse around digital technology and democracy around that time. That before then, it was Obama’s campaign and the use of social media, the early moments of the Arab Spring, and the role that Facebook was supposed to have played, and social media, the internet, was going to bring democracy to the world and make electoral politics more transparent to all and more participatory. And then after Trump’s election, or perhaps starting with Brexit and up to the present day, it’s turned into a tool for disinformation, for election interference, and you know all the taglines that we hear here. So how does it look from India? And also, in the Western press at least in the last few years, there have been stories about how AI is being used in India to create multiple Modis who are present in every place in India at the same time, and this is seen as sort of a dystopian scenario. So describe the evolution of electoral politics in India.

Arvind Gupta: That’s a good question. In 2014, we also used hologram technology for the prime minister, 3D. India is a vast country, so to take the prime ministerial candidate’s message to every nook and corner, we did use 3D holograms for him to address speeches and rallies.

I always have believed, and I’ll continue to believe in that, that the use of technology in a correct ethical manner is fair game, especially in a democracy. Now, your question pertains to misinformation, influence operations, or platform power. And we have fought against that since 2014, the abuse of platform and platform governance. Platforms should not have an opinion on this. We should have an opinion, the people. Platform neutrality is very important. You know, we had the big net neutrality movement in India, which was about the internet and the pipe. But the same thing also is now something that is coming about, and the world has realized that platforms were not neutral. They had their own biases and their own leanings, and they would promote certain content and subdue certain nature of the content. So this is real now. What we felt in 2014 — as I said, and since then, my multiple speeches and discourses have been around — that we believe that digital is only going to go up in democracy, and I kind of predicted that year on year. And with the advent of AI, and it comes with my technical background, we could foresee not only creation but also dissemination becoming super fast and just spreading like wildfire. Now, if it is being done with the right disclosures and the right intent, I think that’s fair. Everybody has a right to use technology to their advantage. But if it is being used without that and without the guardrails, then it becomes a challenge, and then the platforms can’t play a role of either a judge or an accelerator or a decelerator on both sides.

Peter Slezkine: So what are the guardrails now in India?

Arvind Gupta: So we have a lot of guardrails around fake news. We have a lot of guardrails around what can be done using technology during election period and non-election period. We have a lot of regulations around takedowns. So some things like national security, public interest, manipulating markets, and impacting democracies are against the law in India, and the platforms have to comply to that. They have resisted this many times in the past and India has been very strict with that. We have seen the interference of Chinese platforms in India’s security and public discourse and public behavior and trying to influence them. So in June 2020, we banned all Chinese platforms in India. Not one, but all. So there is a fair bit of guardrails that exist today.

And the points that you are making — let me address them because these are very relevant things for an open conversation. You talked about Prime Minister Modi. As I said, in 2014 we even did a hologram, so we are not shying away from that. We use technology. We have our own Indian AI stack called Bhashini. Prime Minister’s speeches are translated in that, and it is used for live translation also.

Peter Slezkine: You mean inside of India. So he speaks in Hindi, and then this AI is used to translate it immediately into all the Indian languages?

Arvind Gupta: We have tried it during live events also. Sometimes it has worked, sometimes it has not worked, but it is with a clear disclosure that this is a translated speech, this is coming with an AI tool. Now of course with that, the negative side comes in — it’s a double-sided sword, right? The second side of the sword is that we also see almost on a weekly basis some AI dubbed video of the prime minister or deepfake which is saying, “You participate in this scheme and get rich,” and all of that happens. So we have to be, as a country, on alert, and our fake news department at the Press Information Bureau is always in alert mode.

So what I’m trying to say, Peter, is twofold. Number one, the correct use, the ethical use with proper disclosures, is something that I personally believe is fair game. And more and more you put the right best practices into play, you are telling the audience, the general citizenry, that this is a genuine video — if it comes with a watermark then this is genuine, and if it doesn’t, it is not, or a photograph. So you’re setting the right thing, but if you shy away from doing it, the bad players are always going to override you. So just because there are bad players, I think we shouldn’t shy away. We should curtail the bad player. We should ensure that there are both regulatory guardrails as well as a takedown of the bad elements that are out there, and calling them out. But the use of technology is there. It’s going to happen. You can’t stay without it.

Peter Slezkine: So your argument in part is that this is an instrument that is available and powerful and will be used in any case, so the effort should be put into separating or clearly demarcating the difference between, let’s call them authentic illusions and illicit illusions. That there are going to be holograms, but you want them to reflect Modi’s actual speech. If it’s an AI-generated translation, it should be a genuine translation of the words he is uttering in Hindi and not, as you say, an image of the prime minister promising that people will get rich if they follow a con.

Arvind Gupta: That’s absolutely, that’s well summarized. Yes.

Peter Slezkine: So can you give us a brief history — I know it’s a very involved history — of India’s evolution as a digital power, from call centers to one of the most interesting ecosystems that exist in the world today?

Arvind Gupta: So I think, as an academic, this is a very important thing that you ask, and I will wear an outside hat to give you my perspective. India’s IT industry started — it probably existed in the ’80s and the ’90s, but really pre-Y2K is where it started gathering pace. When the Y2K problem was imminent, the Indian IT industry started making their presence felt in terms of low cost, high quality workforce available to you offshore to help you solve—

Peter Slezkine: Can you really quickly explain what Y2K is? Because I remember it, just barely, as a potential apocalypse that turned out to be a dud. But some of our listeners might have no idea what it is.

Arvind Gupta: I don’t know whether it turned out to be a dud. See, again, we always undermine maybe the people who solved all the problems before it turned out — because it turned out, and then it resulted, and then it turned out to be fine.

Peter Slezkine: Yeah, so maybe when I was waiting for the world to end in the year 2000 as a 13-year-old, I didn’t appreciate all the work that people did in India. So explain what Y2K is.

Arvind Gupta: So Y2K — let’s put it simply for the sake of this audience. The US uses MM-DD-YY format pre-Y2K. And that’s how our mainframes were programmed. We lived in an era in the ’60s and ’70s where even putting two extra digits was an extra cost to memory, and that cost a lot more. So the databases were designed to be MM-DD-YY in the date format. So if your date of birth is March 1970, it would store it as 03-10-70. And nobody had thought that these programs and software would survive till the year 2000, where 19 was the obvious prefix till 99, but then in 2000 it would become 00. So suddenly you’d be minus a certain number of ages and years. And so imagine the chaos that it would create in the year 00. So Y2K — the year 2000 problem — was making sure that all these programs are read, and wherever the date format is MM-DD-YY, it is changed to MM-DD-YYYY for the four-digit year.

And that’s a huge thing because date stamping is used in so much logic, so many calculations, from banking interest rates to flight bookings. Just imagine — you use it in a phone, in billing, everywhere. So to make the systems change required an effort. And the origins — I’m not saying the origins, the scale-up of the Indian IT industry really started with that, where you needed massive scale of programmers to look into code, to change it, and then to certify that this is going to both not break in year 2000 but also continue to do the same business logic that it was doing before that.

Peter Slezkine: So these are American companies who were worried about the dangers of Y2K in the two-digit year, and they started looking for large-scale labor in India?

Arvind Gupta: America computerized and digitized the earliest, and the bulk of the companies were there, but it was just not American companies. It was also European companies, Australian companies, everywhere around the world. Anywhere that used mainframe or midframe computing, which had that built into the software programming, they had to change. So the bulk of it was in the US, but it’s not just the US. Airline booking systems, for example, are global.

Peter Slezkine: But you’re saying that Indian IT labor was used largely to solve this global problem, and that was a big boost for the Indian IT sector because suddenly you have—

Arvind Gupta: More and more IT expertise, Peter, was used to solve for this problem, and it was the origins of the birth, or the scale-up, of many companies. If they were doing 100 million in revenue, certainly post-Y2K they started doing a billion dollars. Because now they had leads, they had clients who had seen them work, and the birth of the IT industry — really the scale-up of the IT industry — began there. So if you take the last 25 years, if the total size of the IT industry and the exports from India to the US, let’s just use that as a benchmark, were half a billion dollars prior to year 2000, today that industry is $300 billion. And it has of course matured as not just doing one bit of outsourcing, but total application management, infrastructure management, new program development, support of new rollouts, and AI now, of course. So everything that you can imagine today, the big five, six companies of India as a full stack offering, they do, and they compete with the rest of the world. So it’s a very sizable business. It’s one of the biggest exports from India. It is the biggest employer of India. This was the story of India’s growth. Today, directly within India, about five million people work in the IT sector, and that basically means, if you take the indirect employment, it’s 20 million along with that. So it’s a huge sector in India. And it’s part of the global supply chain now. It’s not just standalone — it’s important for India, but it’s important for the world. There are about 2,000 Global Capability Centers operating in India, which are apart from the IT companies. So everybody has understood that India is the talent pool for technology. We produce a million engineers every year, and that’s probably the highest in the world. So to be able to capitalize on that engineering prowess and the talent pool is what every company is doing. So outsourcing is still a cost-based model that the IT services companies do, but when they have development and their own internal needs, they are setting up GCCs, the Global Capability Centers. And the mix of the two is $300, $400 billion from India in exports.

Peter Slezkine: How do you differentiate between the IT work that is done for outsourcing, for export for the rest of the world, and that which is done for India domestically? You were talking about digital sovereignty before, so India has its own digital system, which is different, I suppose, from elsewhere in the world. But India also supplies the IT expertise for much of the global digital platforms.

Arvind Gupta: So there are two different paths. The first one is a business model, it’s an industry. Digital sovereignty is both policy and vision-driven, right? Now that we have so much expertise — and just to give you some numbers: in 2014, inward-looking, we were probably consuming about three percent of this whole IT industry in India, and most of it was export. And everybody you asked was working for a client overseas. Today that world has changed. For example, it’s about 15% domestically and 85% globally of a larger pie. So what has happened over the last 12 years, since Digital India has become a prominent program within the government of India, is that we’ve started building — and this “we” means businesses, banks, which were already using technology but started using it at a much larger scale. And of course the benefit of leapfrogging has also started happening, which is that they missed a lot of buses, so now they sometimes went straight to a cloud-native architecture, to AI-native architectures.

Digital sovereignty, Peter, if I define it, is driven by, as I said, policy as well as some kind of a vision. But for critical sectors, we need to be dependent on ourselves. Which means that if a conflict happens in the Middle East, if the data center pipes are cut off between the US and India, my financial sector should keep working nonetheless. So for example, the financial sector has been identified as a critical sector. From data centers to the applications, the cloud running on top of it, the talent pool that manages it, to the outcomes — and the data storage is also required to be kept in India. It can’t be outside. It can be mirrored and used for analytics outside, but there has to be availability here. So that also ensures, both from a technology perspective and from a regulatory perspective, that if we need to do fraud investigations or money laundering investigations, or make new software for that, we are not dependent on somebody else to give us the permissions to use the data for the analytics. And that is where digital sovereignty started as a thought process.

Now it’s expanded. And as I said, there are a lot of rules. Many things happened — policy-wise, we banned all the Chinese platforms. So what you need today from the stack of technology infrastructure is to have some control over the supply chain of the chips, to make sure that they are clean, there is no malware. Because chips are now going into every car, every fridge, every TV, every computer, so no compromise on that. So what can be indigenized, replaced, and made in India, or if not, certified to make sure that there is nothing in any product in India that has a backdoor into it. Down to the telecom connectivity — same thing. We have actually built a complete 5G telecom network with no technology from China. And we are probably the only network in the world like this, and the cheapest still. So we’ve broken the myth that you need Chinese components and Chinese telecom networking equipment to make a cheap network. India has the cheapest internet connectivity in the world today, and without any Chinese components in 5G. So it starts from the chip infrastructure, the telecom infrastructure, which is the connectivity infrastructure, to the data centers, to the data storage facilities, to then building up to the applications, and having full control over there.

Now, control does not mean having it with only you developing it. It can be developed by Indian software companies, and we have global software companies developing in India. But the final control of the whole stack needs to lie either with that business or with the government. And it is a relationship where data is available easily. It becomes a big challenge in the world of social media and the world of digital media, where data is all global. And when data is flowing back and forth, there is content flowing back and forth — if something happens, what is not good for me could be okay and acceptable to a country across to our northeast or to our northwest. So we have to have some rules and guardrails that if we say this is not allowed here, it is not allowed here.

So I think we can keep going on and around talking about what sovereignty means here, but there are a lot of components to it. What India has been able to do, either through partnerships and local development through the India Stack and the digital public infrastructure of India, is really start controlling certain elements which are essential. The data funnel in India, for example — India always follows as the data capital of the world with the diversity of data that we produce, by language and with a billion people on the internet. It’s something that we have really made sure, with policy and with local infrastructure, that we have control of. Now, AI is the new battleground. Data is going to flow across borders. But what we are doing through both domestic development and international relationships, where geopolitics comes into play, is to make sure that we are in good control. I won’t say absolute control, but in good control of our digital sovereignty. We have removed certain players from the Chinese side, so we know that the bad elements are gone. The others work with us both from a law enforcement perspective and a co-development perspective. And such is the game.

Now, on a scale of zero to 100, I think that technology sovereignty is not achieved 100% by anybody. At best, you can hope to be going towards a 90 passing A-plus grade mark. We are probably at 40, 50. The US is probably 60 on the whole technology sovereignty. I mean, even the US cannot claim 100% technology sovereignty. The talent pool lies sometimes in India with the GCCs. The hardware is getting made in Taiwan. Within Taiwan, the lenses and all are coming from Korea and from Japan, and the machines are coming from the Netherlands. So all countries are aspiring to go as close to the holy grail of 90-plus. The US is probably closer. India is not there, but we want to keep moving up the ladder.

Peter Slezkine: Well, as you seek to improve your grade, what holes would you like to close first, or what are the problems that you think should be the greatest priority to solve?

Arvind Gupta: I think one of the biggest priorities to solve — some of the bigger problems, like the data centers, data control, are getting solved. It’s the smaller ones now which are embedded into our daily lives — the chips. Chips that go into autos, into TVs, into phones, into CCTV cameras, lenses. This is small, and it’s an easier substitution than trying to build the next H100 or similar high-end chips in India. So which will happen eventually, but you start with making the fabricated memory, the low-end memories, all the daily use chips that we currently import — that is the first thing to do.

Peter Slezkine: And that seems to be the problem for everybody. The Americans are trying to solve the chip problem, the Chinese, the Indians as well. That, I guess, is everybody’s priority at the moment.

Arvind Gupta: That is a big priority, and for us it’s both an economic decision as well as a sovereignty decision. Because the economic decision is — we used to import almost all mobile phones into India. Today, in the Android market, India is about 97, 98% Android-based user base. We domestically manufacture 95% of that in short six, seven years. Same thing in iPhone. Three years ago, four years ago, we didn’t have a single iPhone made in India. Last year we made $52 billion worth of iPhones in India. So it is something that we aspire to do. We have the talent pool, we have the resources, we have the intent. But you are right, chips are a big thing, but it’s also both a job creation tool and an economic decision, because you don’t want to be importing so much, and eventually a sovereignty thing. And technology feeds into that score that I just told you about.

Peter Slezkine: So digital sovereignty is in part the domestic production of hardware infrastructure. It’s reliable partnerships or a diversification of partnerships with foreign companies and states. It’s legal agreements, licensing, and a longer list. You mentioned control of data as being very important for digital sovereignty, but also it’s one of the greatest assets that India possesses. It has a huge amount of diverse data. So can you explain once you control that asset, how you create value out of it for domestic purposes and as something that holds value elsewhere that you can share with others?

Arvind Gupta: So give you one example. We need better algorithms to give better credit to India. Credit is one of the most underserved things everywhere in the world — matching the right credit with the right rates to the right person. If you have more control on your data and more data points with alternate data, not just financial data, you can actually rate people much better. That’s a very first major use case for data and alternate data.

Peter Slezkine: And this is data that would be useful in the US, let’s say, for solving the same problem? Or is it so context-specific that—

Arvind Gupta: Everyone wants to give more credit, to India and to the US. Companies want to rate and give more credit. So everybody wants to do that. But going into that, there are so many other data points which could be sensitive, which we don’t want to share. Location, right? And that is where India has put together an architecture called DEPA — Data Empowerment and Protection Architecture. And this is the guardrails I’m talking about. It’s a techno-legal framework, part of the India Stack, which has identity, KYC, UPI the payments, and then the consent architecture built into it. This is the whole India Stack. And in this, DEPA is the central part — which is how do you empower the users to use data for their own empowerment, but also protect it so it’s not misused and shared? So if I share my personal data for bank A to give me better credit, bank A cannot go and tie up with Google and share the data to Google or to any other partner in the US.

So that is the framework that we have. But Peter, you have to understand — there was a time, as I said, and this is very interesting that you asked the question about Y2K, where storing two less digits was a priority for computers, and that saved memory, and that was a critical business decision. Today, I want every single thing. Whether I’ve raised my hand or lowered my hand, where I’m standing at 5:01 and where I’m going at 5:02. The cost of compute and the cost of storage has come down so drastically that we produce more content, more data every year than the last ten years combined. So nobody’s deleting stuff any longer. So you want a lot more data. And I think that is why every data that people keep collecting, big tech will find ways to monetize on it. And we are trying to just make sure that at least when it comes to personal data that people are sharing, it should have the advantage to the consumer rather than technology companies.

Peter Slezkine: You mentioned that kicking out Chinese platforms was one of the first key steps on the way to Indian digital sovereignty. To what extent was it a matter simply of exclusion or to some degree imitation? China was already on the road to creating its own digital sovereignty. Was China a model as well as a competitor?

Arvind Gupta: Any comparison with China, Peter, is actually always wrong. China lives in its own walled garden. China doesn’t have the same democratic norms that we have. China has a very different way the capital and the state capital work. There’s hardly any way to distinguish it. We respect China for its talent pool and its innovative thinking, but that also leads to problems for other nations.

So to me, I don’t think we are trying to imitate the Chinese model at all. Let me give you my definition of the three models that exist in the world of tech. And on a lighter note, I’ll give you a fourth model. There is a model which is the US model, which is called data for capitalization. That’s clearly the model of big tech. Use data, capitalize it, make more money. And that’s why you have trillion-dollar companies, $4 trillion companies in big tech. The Chinese model is data for surveillance. Capitalism plus surveillance. You can just go about and do anything with the data that you have in the Chinese ecosystem. I don’t think India is trying to do that at all because we have our model, which is called data for development. We need to use the model of technology which is for human development, societal development, population scale, bottom-up. These are the things that India has presented to the world. How do you make technology more inclusive? Make the bottom of the pyramid participate in the technology revolution, use it as a tool for development and empowerment.

And I can go on and on and give you examples. In which country would you have used technology to create 550 million bank accounts with zero balance permissible to the poorest of the poor? It’s possible only with Indian technology and Indian staff. That’s the bottom-up approach. The KYC layer, the Aadhaar layer, which is our identity layer, is something that is enabling the banks to serve the low-income communities. So those are the three models. The fourth one, as I promised I’ll joke about, is the European model, which is data for regulation. Regulate and make money out of all this big tech and nothing else about it. So the three models that exist — we couldn’t be more different than the Chinese model. The Chinese and the American models are both actually top-down models. We are a bottom-up model of technology. That goes into the thinking of everything that we do in India.

We are the model which says let a hundred unicorns bloom on the Indian sovereign stack, the FinTech stack or the India Stack. It’s open for other companies to participate. We are not only thinking of creating five or ten big tech companies. The model is maybe there should be 500 big startups, but not just one. So it’s a very different approach to technology, and that is something I think we are very proud of in India.

Peter Slezkine: Where does Russia fit in? I know that on both the Russian and Indian side there was great hope for cooperation in the digital sphere, for some sort of mutually beneficial exchange. So what is India’s relationship to Russia in the IT sector, and what do you make of the recent Russian crackdown on various internet platforms?

Arvind Gupta: Well, the Russian crackdown on the internet platforms — they should be speaking for themselves and understanding it. I think they also became too dependent, before the conflict that they started — the Ukraine conflict — on US technology. I have always been very, very cautious. Big tech is nobody’s friend, and that’s where your question of sovereignty comes into play, right? Is big tech the extension of the state? Chinese are clearly an extension of the state. In the US there is a heavy mixture of now Washington and California deciding how they want to operate. India has navigated that well so far, and we have ensured that Indian priorities are taken into account. Russia became too dependent and paid the price for that. And so if they are coming down on platforms which they thought were conducive to them but turned out not to be, that’s a decision they should have made in the past and more aggressively. But that’s for them to decide. I’m analyzing it as an outsider.

On the India-Russia side, I think India has been — and this is a double whammy for India because a lot of Indian companies also used to serve the Russian clients. Because of the sanctions, they had to stop doing that. So that cooperation really came down, not because of India and Russia’s bilateral issues, but primarily due to global considerations. There is cooperation happening on technology, on AI, on the quantum mission, on the scientific side of things. And that’s where it is limited to. It could probably grow and extend a lot in health tech and in many areas, especially around physical AI, in smart manufacturing. But there are global headwinds to it, which are understandable given that India lives in a world where capital is coming from all across the world and they are sometimes subject to — if a company takes funding or a client from Russia and the funding has come from the US, it’s not allowed. Those are the things at play right now. But government to government, certain projects, and scientific community to scientific community, they are progressing well. Hopefully, and again, that’s a hope that all of us have, that the conflict comes to an end and there are no sanctions, so that the world of technology can progress from both sides.

Peter Slezkine: Well, to end, let’s address the topic with which we probably should have started, which is India’s extraordinary importance to American big tech. And perhaps we could look at this at three separate levels and three different spheres. One is the idea of India as the new China. You mentioned the number of Apple iPhones produced in India now. There has been an idea now for several years that what was produced in China could now be produced in India. So how promising is that path? The second is the prominence of South Asians, Indians, at the very top of American big tech as CEOs of all kinds of major companies. And finally, at the sort of middle or bottom levels, the H-1B question, the ubiquity of Indian engineers in the United States actually powering these companies, and the way that this helps drive them, but also creates tensions politically in the United States, and has raised this H-1B visa issue to the highest level of politics.

Arvind Gupta: Okay. I know we’ve been speaking for the last 50 minutes, so let me give you a briefer answer on this. It would be wrong to say India is the new China. When we work with countries, we work in a trusted manner. We are not putting backdoors, we are not putting malware in it. I would like to position it and say India is now a partner in resilience, in digital resilience with the West. Whether that means hardware supply chain or software supply chain, it’s up to us to decide. But both the hardware and the software supply chain — and these are broad terms — India is a partner for resilience, a partner in resilience and a partner with resilience, with friendshoring as a concept. We are not just a low-cost partner anymore. That’s where we have moved up the value chain. We are a friendshoring partner. We are a partner that you can work with at all times. And I think geopolitics has been a great tailwind in this, to help us come to this kind of perception and cooperation.

On the Indian CEOs in the US — a lot of them are my friends, my college mates — I think it’s natural. Indians have generally been good as a technology talent pool. I would say 80% of the CEOs of Indian origin are in the tech industry, 20% in banking, and some in many other industries, but primarily concentrated towards technology. Indians are also good managers, good communicators, and I think that has naturally led them to grow organically into those roles. You’ve seen what Satya has done, what Sundar has done, what Nikesh Arora has done as CEOs. Shantanu, who’s retiring from Adobe, has done it. So all of them have really, really served their companies, their boards, their shareholders and stakeholders very, very well. And they are themselves products of the whole migration, H-1B, students studying in America and then getting the first visa as H-1B, and then becoming US citizens, and then really going to the top. So they see the benefit of that. But that’s not a benefit that they see accruing to themselves. It’s the ecosystem that benefits.

Now the US has, of course, its own considerations when it comes to domestic jobs. And so somewhere there they are also advocating for a balanced approach because Indian talent is now being sought after everywhere in the world. And as I said, it’s no longer just a low-cost replacement talent. It’s the high-end talent also, the brains that really do it, but also at a certain volume. So I think the argument for H-1B is always that you get the right people. They are seeking the top-end jobs. They’re creating more value than what they are in the US economic ecosystem, the digital ecosystem, the technology ecosystem. So that argument is holding good. It’s a US policy to decide whether the trade-offs between local jobs and H-1B jobs and the larger value — my example is very simple: in a tech job in India, it creates four downstream, upstream and downstream, jobs. So if that’s the case, if the simple argument is that if you get high-end talent and that creates more jobs for you eventually as an economy, I think it makes a good argument.

Peter Slezkine: How did the H-1B issue reverberate in India when the policy was changed, when some people who meant to work in the US had to go back to India? There are Indian businesses that are involved with this pipeline. Did it have a big effect on the economy? Was it something that was covered much in the media?

Arvind Gupta: Well, it did. But Peter, you have to realize — policy shocks, one has to prepare for. In any business, I sit as a board member in many businesses which are exposed to so many other things. I mean, you have to now prepare for a world where geopolitics and policy has to be in the boardroom discussion. So if businesses found this shocking, I think they were unprepared, and they need better preparation at their boardroom level for this because you need to have this as a plan B. Of course, individuals get dislodged and their own aspirations don’t get fulfilled, which is regrettable. But at the end of the day, sometimes we have gained as a country. So if you think about it, we now have a trend of people returning from the US and forming unicorns and multi-billion dollar companies in India. So it’s a two-way story. And yes, there is merit in the H-1B argument for the US. But that’s for the companies and the US to decide. Individually, of course, it gives them a shock. But the Indian policymakers also have to see the benefit of what is more important sometimes.

Peter Slezkine: Well, very good. As you mentioned, geopolitics can be headwinds or tailwinds, so it’s good for boardrooms to prepare for all kinds of climates. Thank you so much for spending much more time with me than promised. Thanks again for appearing on the show.

Arvind Gupta: Thank you, Peter. We had a great conversation, and I thought, as you said when starting the conversation, we’ll do it about 40 minutes. My clock is showing 57, 58 minutes. So I think time just passed by. A good conversation leads to that. So thank you for doing that, and look forward to more such conversations in person.