Speaking about the evolution of tokenisation and the Finternet at the Global Fintech Fest 2026, Nilekani said creating tokens alone would not be enough. The ecosystem would need interoperability, liquidity and the ability to operate at very high volumes if tokenisation is to extend across different asset classes and use cases.
“If you're going to make this population-level tokenisation, then you need very high volumes,” Nilekani said, adding that the architecture being developed for the Finternet is designed to address this.
He said tokenisation is already gathering momentum, pointing to initiatives involving certificates of deposit and corporate bonds. Globally too, large banks and asset managers are experimenting with tokenised assets. SEBI recently launched
Nilekani clarified that tokenisation of assets is different from the “tokens” used in AI, where a token is a unit of work or consumption. In financial tokenisation, an asset and all its attributes are bundled into a single, portable digital package that can be transferred to another party.
Also Read
“If you're not careful, we'll have a lot of token issuers, but no activity. You don't have the interoperability, you don't have the liquidity, and all that,” he said.
Nilekani cited three use cases being developed using the same underlying infrastructure. First was tokenisation of cattle to enable access to credit — in this, a cow and the information about the cow can be tokenised and the owner can then get credit based on that. Second, portable tokenised warehouse receipts that can be offered to multiple lenders, and third, a marketplace for securitisation of assets. The objective, he said, is to create an architecture that can work across multiple asset types and use cases.
Nilekani argued that such infrastructure should use public chains rather than private chains if the objective is to create a large market with adequate liquidity. The architecture also needs to be interoperable, public-chain agnostic, capable of handling high transaction volumes, quantum-safe and designed to address cybersecurity threats.
He also sees AI agents playing an important role in creating demand for tokenised assets. Agents operating round the clock could, for instance, package an invoice as a token, offer it to different lenders and obtain financing.
“Tokenisation by itself will only handle issuance. But tokenisations on public chains supported by agentic transactions will also create demand,” Nilekani said.
He said agents and tokens would have to come together to create the volumes needed to justify such platforms.
AI can create level playing field between large and small businesses
The larger opportunity from combining tokens and AI agents, according to Nilekani, is to level the playing field between large companies and small businesses.
In the past, a company needed functions such as a treasury desk, research team, credit rating and sales team to participate effectively in financial markets. A small business can now potentially use a 24x7 AI agent, access analysis and have a token that can establish its credentials in a marketplace to raise a loan or undertake a transaction.
“So the main thing, why we believe that tokens and agents are important is actually about inclusion,” he said.
Nilekani also linked the growth of small businesses to the debate over AI's impact on employment. Large companies, he said, can use AI more effectively to reduce jobs because their activities, tasks and jobs tend to be more structured and therefore easier to automate.
“If you want to create an economy that's lively and is generating jobs, it's not necessarily done by having a few big companies. It's going to come from having millions of small companies,” he said.
Jobs at small businesses tend to involve employees doing different and more dynamic tasks, making them harder to automate, he argued.
“Those millions of jobs in millions of small companies are actually safer than a few million jobs in large companies,” Nilekani said, adding that creating an “AI-proof economy” would require job creation to be driven by millions of small businesses.
He expects the number of startups in India to continue rising sharply. Nilekani said India had about 10,000 startups in 2015 and 150,000 in 2025, and projected that the number would reach one million by 2035.
Both startups and traditional small businesses would be “the employers of the future”, he said, making it important to use technology to make it easier for them to operate and compete with larger companies.
