Are micropayments about to have their moment?

Micropayments have long been a trope of “future of money” discussions. The term tends to refer to payments of less than a dollar. Yet micropayments never took off, chiefly because transaction costs make micropayments prohibitively expensive.

There are, in theory, workarounds. Customers can fund a digital wallet on a closed-loop network so they can avoid transaction costs when they make payments on the platform. Or platforms can aggregate many micropayments into fewer, larger, economically viable transactions.² But these solutions are hard to scale, and adoption of micropayments has remained limited.

Recently, however, that has started to change. Account-to-account (A2A) payment systems are making it feasible to send small amounts of money in an affordable way, since the fees are so low. For instance, the Central Bank of Brazil processes six billion transactions per month on its Pix platform, many of them for a single cent. Interestingly, this is often for an unintended purpose.³ Since messages can be attached to Pix transactions, people use it as an informal communication channel. Everything from marriage proposals to insults have been sent via one-cent Pix.

Microtransactions could also unlock various new opportunities for businesses, usually involving “on demand” models for access to resources, content or services. Media outlets could offer readers an alternative to buying an expensive subscription: instead, just make a micropayment for each article you want to read. Websites could offer to remove intrusive advertisements in return for a micropayment. Or consider charities—the ability to make regular small donations, perhaps as part of rounding up the value of a purchase, could create valuable new revenue streams.

Just when we thought the problem had been solved, however, a new use case has emerged that may require a new solution altogether: AI agents. Broadly speaking, an agent is an AI model that can use digital tools to achieve a given goal. Imagine you want to arrange a business trip. You could explain your needs to an AI agent, and it would carry out its own research to create an itinerary and perhaps even make the bookings. Or maybe you could ask an agent to complete a complicated research task that would otherwise have taken you a month. These are all use cases that leaders in the AI space say are coming.

In achieving your goals, agents may need one-time access to gated APIs or premium data sources. They may also need access to specialized machine learning models, perhaps for enhanced abilities such as price negotiation. It seems reasonable to think the agents will have to pay to access these things. But this presents a new challenge. An agent may need to ping any number of APIs and models to get something done. Some imagine that we may have many thousands of agents working in concert to crack harder problems, meaning those pings will quickly rack up. It’s unrealistic to take out subscriptions for all these services because that would quickly become too expensive, especially as a resource may only be needed once. Instead, analysts imagine a world where agents pay per use. And that the economics will dictate these payments take the form of really small micropayments—perhaps a fraction of a cent. Call them nanopayments. The problem is that even the low fees associated with A2A rails are likely to be too great to make nanopayments economically viable. And, of course, A2A is not available everywhere.

How could we solve this issue?⁴ The venture capital community posits that blockchain—particularly stablecoins—could prove effective. Stablecoin transactions can be completed at exceptionally low cost. They are also a good means to pay for services located potentially anywhere in the world, since they circumvent conventional payment rails and therefore don’t encounter currency conversion or international payment fees.

The idea of bots continuously exchanging value for services that we cannot fully imagine right now is heading towards the mainstream. Fintechs are already securing venture funding to build payment networks specifically for bots to make autonomous, stablecoin transactions.⁵ If this future transpires, it will create a range of new concerns for payments professionals. Take treasurers, for whom it will be important to optimize liquidity in light of high-volume, low-value payments. They will also need to consider how best to integrate stablecoins into their treasury set-ups. As with all things AI, there is one piece of advice you can rely on, however events unfold: track the space closely and be prepared to act fast.