Good afternoon, and it's great to be back at the Global Fintech Festival, which every year seems to grow from strength to strength. Today, I'll speak about Finternet. Actually, this is my third talk in this series where I have spoken about Finternet, and I thought I'll use this time to give you an update of what's happening in the world of tokenization and in the world of the Finternet.
So now it's very clear that for tokenization, the time has arrived. As you have seen the very exciting announcements that have come from both RBI and SEBI, RBI is talking about tokenizing certificates of deposit, and SEBI is talking about corporate bonds. They're launching it; three or four companies are already launching these bonds, so clearly there's a lot of action in the era of tokenization.
Now, one thing we should remember is that when we use the word tokenization here, it's not the same tokens that the AI guys talk about. AI guys talk about tokens as a unit of work, and token consumption in the AI world is measuring how you consume AI. When we talk about tokenization here, what it really means is that you take an asset and describe all its attributes and put it in one package along with the asset so that that token, which is both the asset and its descriptors, is all bundled together and is portable to give to somebody else.
So this tokenization is very different. To make it clear what tokenization is, for example, a government file is a token because a government file on a particular topic has all the documents over the years, or what has transpired on that topic, the comments of various bureaucrats—it's all in one place. When that file goes from A to B, the entire information travels in the file. So, in fact, the file is like a token.
Or your currency note: if you have a 500-rupee note, it has everything about the note in the note—whether the value, the fact that it is certified by the RBI, its number, and all that. And so when you give that 500-rupee note to somebody else, the other person is confident that they have received the full value of that. So that's a token. Tokens are known to us. What is different is that in the world of computers before tokens, we unbundled this, put them in databases, and did all kinds of things. In some sense, we're going back to the future by defining tokens as self-describing assets with all their attributes.
So that's really this kind of tokenization, and this is something which is happening. There are other things happening on frontier technology: the RBI has come out with an AI rulebook—a very impressive rulebook—and SEBI has worked on a quantum-safe action plan about how do we make sure that over the next two to three years, as cryptography gets broken by quantum, all our financial systems are safe. So that's very important. You can see that a lot of things are happening, and it's related because tokens need cryptography, and therefore they're all related, and this is something which has arrived.
Now, this is a global phenomenon. People are all moving in the same direction. There's about $600 billion of tokens today; most of them are like stablecoins, but the expectation is that there'll be a lot more. One estimate is $19 trillion of assets on tokens in the next seven to eight years. However, this is being done by big banks, this is being done by large mutual funds—so JPMorgan, BlackRock, Franklin Templeton, and so on are all doing some kind of tokens which they're putting out. And in India, of course, I mentioned to you that multiple issuances of tokens are happening by various companies in the coming months. So really, it's all moving in the same direction.
Now, the people working on this idea of Finternet came up with the idea in 2023, and we published a paper in 2024. I co-wrote that paper with Agustín Carstens, who at that time was the head of BIS, because Carstens was trying to think about what the future of the regulated tokenized world would look like. From a central bank regulator's perspective, they were worried about the rise of cryptocurrency and stablecoins, and their view was that if these things are not regulated well, you could have all kinds of issues. Therefore, they wanted to use the power of tokenization, but in a regulated world with all the checks and balances that we have. Hence, he co-authored this paper which came out in 2024.
I talked about that paper right here at this event last year, and we talked about who all are participating. Today, we'll actually talk about three live cases where people are implementing the Finternet ideas to do tokenization across three different disciplines, and explain to you what we are doing there.
The three pilots we are talking about: one is in dairy, and how a cow and the information about the cow can be tokenized, and how you can get credit and so on for that. Second is how you can get a portable warehouse receipt, so you have some grain—rice, wheat, maize—and you keep that in the warehouse, you get a receipt, that can be tokenized, and you can get a loan against your inventory. And the third is a marketplace for securitization of assets.
I'll give you three examples, but what is important to note is that in all three cases, the underlying technology is the same. That's the whole idea of creating an architecture which is usable across multiple asset types in multiple situations. So the way the asset details are stored, the way the tokens are stored, the way the business rules are implemented, how escrow is done—everything is the same standard.
Another important thing about the architecture that all these three use is this is done on a public chain. This is not about creating a private chain. If you really want to create a global market for tokens with enough demand and liquidity, you need to put them on a public chain. All this is really using public chains. It's also public chain agnostic—that is, it doesn't matter which public chain you use; we can use any public chain. We're not really worried about whether it's in Solana or Ethereum or whatever; it's a layer above that.
It's also designed to do very high volumes because, ultimately, if you're going to make this population-level tokenization, then you need very high volumes. Typically, the volumes on these blockchains have not proved to be that great, so the architecture allows for very high volume. It's obviously designed to be quantum-safe so that cryptographic challenges are handled, and it is designed to handle cybersecurity—now we have new types of attacks, including agentic attacks, so they need to look at all that.
The theory is that if you want to accelerate tokenization across all asset classes for all use cases, if you have underlying technology and rules which can bring in industrial-strength, population-class capability, you can use that rapidly. So we have done that. These three examples which I'll show you—all of them have the same underlying infrastructure, and they all use purpose-based programmable transfers. That's where, when you release the money, you can also say this money can only be used for these things; it's like a voucher which is purpose-specified. That's a horizontal capability that's used across these applications.
This is the first one, and it's about dairy and a cow, so I'll just ask for the film to be shown.
India built digital identity at a population scale. Then instant digital payments became part of everyday life. Their power came from what people and businesses could build on them. The next opportunity is to make productive assets and economic rights identifiable, trusted, and usable.
In a village in Tamil Nadu, Uthangarai, meet Dhanalakshmi. She keeps cattle and leads a dairy cooperative where 40 families bring milk each morning. She wants more cattle, more milk, and a proper shelter, but cannot get the finance she needs. The value is in front of her. Yet, a distant lender cannot easily verify the animal, confirm it belongs to her, or assess its condition and earning potential.
A Dvara E-Dairy field officer arrives with a phone. The muzzle pattern of each animal can help distinguish one from another, much like a biometric marker. One of Dhanalakshmi's cattle is tokenized, creating a unique digital representation linked to trusted information about the animal: its identity, ownership, age, health, and milk productivity. A digital proof helps make later changes detectable and traceable. A verifiable cattle credential is issued for the record and made available to her.
She decides what information to share, when, and with whom. With her consent, selected information moves across an open network to the lenders she chooses. Participating lenders can discover her request and respond without her visiting them one by one. Two lenders express interest. She compares their offers and chooses one. After Dhanalakshmi gives further consent, the selected lender verifies the information, carries out its usual credit assessment, and approves the loan.
She reviews and accepts the terms. Once she accepts, the lender makes 50,000 rupees available to Dhanalakshmi as a purpose-based digital loan balance visible in her app and usable for cattle feed and shelter materials. At an eligible supplier, she scans a QR code with her phone. The transaction is checked against the agreed conditions, and the approved amount is paid directly to the supplier. Soon the cattle have a roof. The value was already there; the trusted record made it easier to recognize and use.
And cattle is only one example: a machine powering a small business, rooftop solar, a registered fishing boat, an accepted invoice. When trusted sources digitally represent assets and rights and their records move with consent across open networks, participating lenders, insurers, fintechs, payment networks, and marketplaces can discover new opportunities for finance, insurance, payments, servicing, and verified exchange. The value already exists; the opportunity is to make it visible, trusted, and usable.
