Digital Money's Trio: Stablecoins, Agents, and Wallets
Exploring the next evolutionary step in finance.
The engine of technological change never stops. In the time I've been alive, I’ve seen clunky landline phones give way to the supercomputer in my pocket, cassette tapes give way to on-demand streaming, and shopping mall excursions disappear with near-instant doorstep deliveries. With time, life has seemingly moved to a state of being frictionless.
Yet some parts of our world remain stubbornly stuck in the past. No, I am not referring to the Beatles and their classic 'Hey Jude', but rather to money movement. Not so much paying a friend back via Venmo or UPI for a coffee, but cross-border money movement; arguably one of the most critical functions for our ever-globalized world.
With all the advances we've made, it only makes sense then that there has to be a better way.
Enter: stablecoins. Stablecoins are 1:1 fiat-backed assets that live on a blockchain; they operate 24/7/365, globally, and payments are instantaneous. It is not simply digital fiat, but a new financial primitive for digital economies (to all the fintech nerds reading this, am sure you’re looped in enough).
Let me set the stage: you are a freelancer living in Hanoi. One of your customers, a design agency in New York, pays your monthly fee of $1,000. Today, that $1,000 is sent via a 'network' called SWIFT. This may surprise many, but SWIFT doesn't actually move money; instead, it acts as a secure messaging system across banks.
For example, the Asia Commercial Bank in Vietnam will maintain an account with Citi funded with USD. When you get paid, a message goes through SWIFT instructing Citi to credit the equivalent funds to your local bank account. On average, you might pay $50 for this transaction—sometimes more if you’re in a less 'popular' corridor. With stablecoins, the US client will send $1,000 in USDC to the recipient's digital wallet. Accounting for any minimal gas fees (especially with Solana or L2's on Ethereum), you’ll almost certainly end up with closer to $1,000 than $950. Converting that into local fiat might cost another few dollars, but the settlement is practically instant—no more waiting days on SWIFT. Better yet, there are companies today that abstract away the stablecoin conversion aspect completely, making it feel like users are dealing with fiat throughout the transfer.
We are in the early innings in understanding the vast potential of this technology. Money movement is one, but as the benefits of stablecoins are clear and proliferated, we will see use cases across payroll, wealth preservation, and microtransactions, among a host of others. One area I’ve been fixated on is how stablecoins will integrate with the platform shift we’re witnessing in AI. I believe they’ll become the key instrument powering the next wave of autonomous activity—where agents handle more of our daily tasks, and wallets become vital gateways in this new digital money ecosystem.
Agents Don't Work for Free
Previously, I've shared my thoughts on agents: software that perceives its environment, processes information, and takes autonomous actions to achieve a defined goal. There is immense potential to apply agents to fintech, from single-use agents in accounting and debt collection, to KYC/KYB, where agents coordinate across multiple datasets. As we ride the "slope of enlightenment" in the technology adoption curve and agents become normalized, we will see autonomous, complex actions performed in the background; from planning a work trip to collecting information and filing taxes.
Back to my headline: I don't think agents will work for free, especially when tasks span company borders. For every agent to take a ‘fee’ for their work, microtransactions must prevail, and this is where my perennially optimistic view on stablecoins shines. Beyond being global and interoperable across chains, stablecoins settle at virtually no cost, enabling agents to operate at scale.
Let’s say a user wants to book a work trip to London: they add their dates, preferences, location in the city, and ask their agent to put together an itinerary. The agents scour for the best set of hotels, flights, and even offer to make reservations for lunch and dinner in the area. Once the user signs off on the recommendation, card details are shared across outlets, and the trip is effectively booked. In the background, each agent takes their 'fee' for playing their part.
I know, I'm getting ahead of myself. As illustrated below, today's payment experience is human-led, and although we've made headway in agent interaction, we are still in the world where most use cases are driven by co-pilot, where we need sign-off from consumers (a good example of this use case is here). Although this experience is leaps and bounds better than we have today, we are still at the tip of the iceberg and a few steps away from a breakthrough.
Agents Live in Wallets
The co-pilot model of requiring human intervention is comforting in that funds won't be spent where funds don't want to be spent. But the envisioned future of agentic payments will flip this on its head, where a set of pre-set instructions and thresholds can allow for funds to be spent as intended.
Moving from co-pilot to fully autonomous payments, agents will need to live in a 'home'. A 'home' that is programmable and acts as a command center for authentication. Ultimately, we need to trust in a cryptographic identity layer - a digital passport with immutable signatures - so every agentic transaction doesn’t require fresh KYC. This is what digital wallets offer today, and why that form factor will be the home to agents in the future, too.
Think of this: as a consumer, you add Ethereum, USDC, or an equivalent digital token into your wallet. Within the wallet, your 'virtual card' will contain a budget and specific sets of rules that the agents can comply with. And then, as a consumer, you simply prompt what you need - for ease, using the example earlier of booking a trip to London - and the agents from your wallet, speak to the agents of an airline, and with no opt-in, the receipt is sent to your mailbox. If this sounds too crazy, it isn't... here's a live example of this being done today. There are already companies like Skyfire and Payman building the infrastructure layer that will form the foundation for agents to transact with one another.
The 'why now' moment is primarily driven at the intersection of two core trends: 1) the proliferation of stablecoins; 2) the acceptance of agents.
My personal favorite definition of stablecoins comes from Patrick Collison: “room-temperature superconductors for financial services”. Just as a superconductor allows electricity to flow unimpeded, stablecoins promise to let value flow freely in a digital economy. And this value is being recognised - although just 15% of the broader crypto market is comprised of stablecoins, they represent close to 70% of all trading volumes.
As for AI agents, we’ll need buy-in across the board—from early-stage founders to the biggest tech platforms. We also need the brightest minds optimizing infrastructure, because the microtransactions and autonomous actions described earlier simply can’t run on our existing financial rails.
If agentic payments reach scale, money movement will change entirely. Agents won’t just ‘talk’ to each other—they’ll negotiate transactions based on user-defined thresholds. What if your AI-personal assistant can adjust subscription services based on usage patterns and allocate funds between your checking and savings accounts? Or if you're a crypto native, engage in automated yield farming or dynamic staking based on market conditions?
This is more consumer-facing, but in the world of B2B, the use cases get more complex, but the potential even more expansive; picture the coordination of logistics, where agents can trigger payments based on deliveries or renegotiate if delivery timelines shift.
Bumps in the Road
No matter how promising stablecoins and AI agents appear, broad adoption won’t be smooth.
First, to understand where we’re headed, it’s important to recognize why our current systems function as they do. To start, card networks like Visa and Mastercard don’t charge merchants an interchange fee for the movement of money. This fee funds chargeback resolutions, and any disputes or fraud consumers face. Think of it as a 'fund' for networks to dip into to support their user experience. Stablecoins don't have any equivalent of interchange, which means there is less of a chance of getting funds back in case of fraud. The saying goes, "faster money means faster fraud", and so this remains a sticking point for stablecoin adoption.
It's easy to imagine that stablecoins equate to decentralization, but that simply is not true. Today, Circle and Tether issue approximately 90% of all stablecoins in circulation. We will likely see a shift as adoption increases, but I believe the power law of the mighty few will still prevail. Even though stablecoins are backed by the US treasury, the buck still stops with the issuer. An assumption we need to believe is true then is that issuers of stablecoins will not mismanage their fiat collateral. Any mismanagement here could lead to a crack in the broader system.
And lastly, while the picture of agents interacting and transacting with each other appears seamless, this vision only holds if blockchain fees remain low. Those privy to using their DeFi wallets understand that 'gas fees' can get prohibitively expensive when things are busy. A risk that might hamper adoption is simply that the cost of each agent interaction is too high.
As we look at the world of stablecoins, agents, and the intersection of the two, we are very much in the phase of retakes and revisions. But with time, as more attention gathers to the innumerable use cases, and talent accrues to work on the right problems, I believe there is a chance that in a future not too far from now, money will skid around the world on blockchain enabled skates.
Thank you to Rahul Sanghi for reading and editing my drafts.
Well done. love it! 💸
Stablecoins are the future and the White House is already making moves to stay relevant in this future. By regulating stablecoins and making them dollar backed, they will proliferate the markets in every corner for the same reason stated in this blog. This will in turn preserve the US dollar dominance. What do you think this would mean for local currencies outside of the US? Does rupee and yen lose relevance?