From 15 October, larger merchants in India will pay for UPI for the first time since 2020. It is a small fee with a large question behind it: what are the rails worth, and who keeps the value that runs across them? Back from a week in Bangalore, VEF CIO Alexis Koumoudos on the repricing, why India is the natural home for agentic finance, and the new kind of company being built there.
What stood out most during your time in India?
The speed and rate of change, on many fronts, is exhilarating. Firstly, the best companies are innovating, shipping product and therefore able to grow faster than at any point I’ve witnessed before. It’s not uncommon to hear of companies 5x top line within 12 months, from an already substantial base. What’s particularly exciting is that they are doing it by utilising new tools and technology, and with a far lower incremental cost base than would have been necessary historically. Secondly, the regulator is following suit. I was there in the gap between Parliament clearing the way for a UPI charge in August and the rate being announced this month, and its planned implementation next month. I think the turnaround has been quicker than market participants expected.
How should investors read the new charge on large UPI merchant payments in India?
As a considered price on a piece of national infrastructure, and a fairly modest one. Merchants have paid nothing on small UPI tickets since 2018, when the state started covering the cost, and nothing at all since 2020, when zero MDR (the merchant discount rate, the fee a merchant pays to accept a payment) went from policy to statute. Parliament removed that statute in August and the rate has now been set: 0.40 per cent on merchant payments above ₹2,000, capped at ₹300 a transaction, from 15 October. Payments between individuals, payments to small merchants and everything below ₹2,000 stay free. By count, more than nine in ten merchant payments are untouched. By value, most of the flow now carries a price. That is the design. The kirana stays free and the e-commerce checkout pays.
Both the level and the timing landed higher and earlier than most of the market expected. I read that as a government confident that adoption is deep enough to survive a price, and a clear message that the rail has to pay for itself.
Why does that matter beyond the revenue it creates?
Because it forces a question the rest of the industry has managed to avoid: what is payments infrastructure actually worth? Zero was never the cost. UPI carries over 80 per cent of India’s retail digital payments, the industry puts the cost of running it at around USD 1 billion a year, and the state incentive meant to cover that was closer to USD 160 million. You cannot fund critical infrastructure on that gap forever. What makes India unusual is that it had to work the answer out from first principles, with no legacy fee to anchor to, and the answer it landed on creates a pool worth a couple of billion dollars a year, roughly double what the system costs to run. No other market has had to do that.
Where do you think the value in the payments chain settles?
Moving money is a utility now, and acceptance is a scale game. What merchants will pay for is what sits on top of the rails: reliability, risk, credit and software. That is where I expect value to settle, and it is where operational depth matters. India’s public digital infrastructure has made it easier for companies to reach customers, verify identities and access financial data, but as those tools become available to everyone on similar terms the advantage lies in how well they are used, particularly in payment success rates, routing, authentication and reconciliation. Our portfolio company Juspay is a good example, processing more than a trillion dollars in annualised payment volume for banks and merchants.
Is any of this relevant outside India?
I think it is the most interesting part. The rest of the world is drifting in the same direction through interchange caps and account-to-account rails, just more slowly and from a higher starting point. India got to zero first, so it has to answer first. The rate is the headline, but how it is split between the issuing bank, the app, the acquirer and the network underneath will be more telling, because that split is in effect a valuation of each layer of the stack. Whatever India settles on will be a benchmark other markets look to.
Agentic finance was the theme of this year’s Global Fintech Fest. Is India actually set up for it?
Better than anywhere else, and for reasons that have nothing to do with AI. An agent acting for you in finance needs four things: to know who you are, to have your permission, to see your data and to move your money. India built all four as public infrastructure over the past decade. Aadhaar for identity, UPI for payments, AutoPay mandates for standing consent to pay, and the Account Aggregator framework for sharing financial data with consent. Elsewhere those pieces are being assembled by card networks and big tech, each with its own protocol, on top of proprietary rails. In India they already exist, they talk to each other, and anyone can build on them.
The second reason is the regulator. The regulator has understood both the power and the risk of agents and is writing the rules before the market writes them for it, which is how UPI came about in the first place. Juspay is already building here, with agentic commerce work alongside Visa, Mastercard and HSBC, and I would not be surprised if the first agentic finance products at real scale come out of India, for the same reason the first instant payments at real scale did.
Where do you see the most interesting opportunities in India now?
The businesses that excited me most in Bangalore are AI native fintechs, built in the last one or two years by small teams who take for granted a set of tools that did not exist when we started investing in India. The public rails give them identity, payments and data from day one. Coding agents let a team of twenty do what used to take a hundred. Language models handle the customer conversations, the collections calls and the document checks that used to need a floor full of people. The result is a company that changes its product weekly, launches a new line in days and grows at a pace that would have been implausible three years ago. One company we spent time with has grown revenue roughly fivefold in twelve months, whilst barely touching their most recent round proceeds.
Capital has become more disciplined, which is healthy, and IPOs and secondaries have given the market a working exit route. Real companies, new tools and a way to get your money out. That is a rare combination in emerging markets.