A couple of months after
Ai4, I was back at The Venetian in Las Vegas, this time for Money20/20 USA 2026. Same building, very different crowd — and for the first time in a while, it felt like every world I work in was in one building. Banks and card networks. Identity verification. Crypto infrastructure and stablecoins. And, everywhere, AI agents.
Those used to be separate conferences. This year they were the same conversation.
The show ran October 18–21, and what struck me wasn't any single demo. It was how practical everything had gotten. The talks I caught and the conversations I had were about plumbing: how to pay people, how to verify who's on the other end, how to let software act on someone's behalf, and how to make all of it work for people the existing system doesn't serve well. The
official program reads the same way.
First, a thank-you to the Money20/20 team for inviting me. You put on a great show — the expo, the sessions, and the people made the trip. I even got a haircut on the show floor. More on that later.
Banks and crypto are converging around infrastructure
Walking the expo, the old dividing lines were hard to find. My camera roll from the floor has American Express, Discover, BitGo, Ripple, OpenFX, Prove, Veriff and Persona, plus a long tail of smaller teams. Crypto wasn't off in its own corner. It was part of the payments, settlement and verification conversation, right next to the incumbents.
The deals of the last two years explain a lot of that. Stripe
closed its acquisition of Bridge in February 2025, and Ripple
announced a $1.25 billion acquisition of Hidden Road that April. Neither was news at this show, but they're part of why the convergence on the floor felt like real business instead of another round of pilots.
The use cases I kept coming back to were stablecoins, infrastructure, and real-world assets. On tokenization, my question is always the same: does putting the asset on-chain make it easier to verify and easier to use? A token doesn't answer who owns the thing, who gets repaid, or what happens when something breaks. The interesting work is in those answers.
Card networks, payments infrastructure and identity verification, all on one aisle.
AI agents that pay need more than intelligence
Agents paying for things on behalf of people and businesses came up in almost every conversation I had.
I
run a team of AI agents that write and review code every day, so I know how much changes when software stops recommending and starts doing. A recommendation is harmless. A purchase isn't. The moment an agent can spend money, you need answers to boring, critical questions: who authorized this, how much can it spend, who is it allowed to pay, and when does a human have to step in?
One agenda board I snapped summed up the direction:
“Micropayments Finally Have a Customer in Agentic Commerce,” with Citi, Artian AI, Coinbase and AWS on the panel. I only photographed the board, but the title gets at something real. An agent doing a job may need to pay for a bit of data, an API call, or one small task along the way, and most payment rails weren't built with that kind of buyer in mind.
Here's where I've landed: intelligence and authority are different things. A very capable agent shouldn't automatically get broad permission to move money. The question I want every system to be able to answer is this one:
Can this agent prove it's allowed to make this payment, to this recipient, right now?
The ArrA.I.val Stage. AI wasn't a side track at a payments show this year.
Stablecoin payments get interesting when someone gets paid
I also spent time talking about paying contractors and moving money across borders with stablecoins. This one is easy to understand: someone did the work, and they need to get paid in money they can actually use.
The
“Borderless Payday: Modernizing the Global Workforce” panel put BVNK, Upwork, Thunes and FXC Intelligence on one stage to talk about exactly that: payment infrastructure for a workforce that lives everywhere.
That's a much better place to start than asking anyone to get excited about a blockchain. Start with the person waiting on the payment. What currency do they need? What lands after fees and conversion? Can they spend it, or get it into their local bank, without a fight?
My bar is the whole experience, not the on-chain transfer. Fast settlement is great, but the payment isn't done until the person can use the money. Local payout, liquidity, onboarding and support are part of the product, not someone else's problem.
When this works, nobody notices it: the contractor just gets paid, and the infrastructure underneath disappears.
The practical question: how does someone working across borders get paid?
Portable credentials: the idea I can't put down
Some of my best side conversations were about identity, funding decisions, and risk in gaming.
The gaming conversation surprised me. I expected it to be about funding wallets with stablecoins. It turned out to be about risk assessment — pulling signals from different systems and deciding what they add up to. My reaction was immediate: some of that verified information should be portable.
This is close to home. I founded
CredZK around a narrow version of the same idea: prove the one fact a lender needs, cryptographically, and keep the rest private. The opportunity I see is the general version — a scoped, verifiable credential that answers one specific question, shared with the person's permission, instead of a copy of their history or a label that follows them forever.
The hard part is meaning, not cryptography. Whoever receives the credential needs to know who issued it, what it actually proves, when it was checked, whether it's still valid, and whether it fits the decision in front of them. A risk score isn't a fact, and a fraud check isn't a credit decision. To be clear, this is an opportunity I came away with, not something I saw shipped.
It also loops straight back to agent payments. If a person, or an agent acting for them, moves between services, the proof should be able to travel with them instead of every new interaction starting trust from zero.
Identity verification was everywhere on the floor. Making that verification portable is the part I keep thinking about.
The Startup Hub kept things honest
I loved seeing what the smaller teams were building next to the big institutions. The Startup Hub and the aisles around it were full of lending software, agent platforms, payments products and data tools.
A keynote tells you where an industry is heading. A booth conversation tells you what has to work on Monday morning. What data does this need? Where does it fit in the workflow someone already has? Who handles the exceptions? How does the customer find out what happened?
That back-and-forth between the big vision and the Monday-morning details is why I still go to these things.
The Startup Hub, where the practical questions get asked.
The talk I stumbled into
The session I keep thinking about is one I walked into almost by accident:
“Stablecoins and Women’s Financial Freedom.”The program listed Fereshteh Forough of Code to Inspire in conversation with Money20/20's Tracey Davies. The subject was Afghan women and girls: learning to code, finding remote work, and getting paid when ordinary banking and cross-border payments mostly don't reach them.
A
Code to Inspire pilot case study published by Binance Charity makes the problem concrete. Back in May 2022, that pilot sent 100 beneficiaries in Afghanistan 200 BUSD a month for three months through Trust Wallet — about $200 a month at the time. The report is just as honest about what went wrong. Of 65 beneficiaries surveyed, 40% said cashing out was their biggest challenge, and recovering wallet credentials was another one. Getting money into a wallet didn't make the rest of the problem go away.
What the session did was challenge how we pick problems. In the US it's easy to build on assumptions we never notice: a working bank account, a few cards, reliable internet, a phone that's always on. We pour enormous effort into making life slightly smoother for people who already have options.
What do we build for someone who has a phone, a patchy connection, and no bank that will reliably serve them?
That's where blockchain gets interesting to me again. Not as a way to get people onto crypto, but as a way for someone to get paid, work, and take part when the usual rails won't carry them.
Build for the people with the fewest options, not just the people whose lives already fit our product assumptions.
An unexpected session became the most memorable part of my conference.
Sessions worth catching up on
I couldn't be everywhere. Going back through the program, three sessions line up with everything above, and they're on my catch-up list:
All three circle the question I want to keep working on: how do we make money move more easily without treating authority, accountability, or access as an afterthought?
Good conversations, and a cleaner haircut
The best parts of a conference are rarely on the agenda. I met people I want to keep talking to and left with more ideas than I came with.
And yes, the haircut. One booth had a full barber setup right on the expo floor. I needed a cleanup, so I sat down. Best booth swag I've ever gotten.
Back at The Venetian for Money20/20 USA 2026.
Thanks again to Money20/20 for the invitation, and for getting so many corners of this industry into one room.
I left excited about AI agents that can pay, stablecoin payouts that actually land, portable credentials, and the institutions finally taking this infrastructure seriously. But the thought I'm holding onto came from that session I almost skipped:
The most important thing we build might not make a comfortable financial life a little easier. It might give someone an option they never had.
Disclosure: Money20/20 invited me to attend. This post combines my own observations and photos with linked research I used to check session and program details. Opinions and product ideas are my own. See the editorial policy for how this site distinguishes reporting from opinion. 
