Recorded at StableCon in Washington, D.C. The transcript below has been lightly edited for clarity and length.
Good to see you guys. What a setting. Best podcast studio ever, and our very first podcast at Coinbax. Taking it on the road, outdoors only. That could be the theme: middle of winter, out in the snow. Love it.
So how are you liking StableCon?
It’s interesting. To me it feels a little more mainstream this year. Last year I felt like I was the only person there representing traditional banking. This year I’ve seen at least a few more people who cross over into banking.
I think the location matters. I was at the first one in New York, and that definitely felt like the early movers. It was almost a who’s who of the startups in the space. Amsterdam started to get a little more institutional. Now, coming to D.C., it’s definitely heavier on regulatory and banking. It’s got that mix.
Are you doing any of the sessions? I saw Travis Hill.
Yes, it’s good that he’s here.
You chat with him a lot, don’t you?
Not a lot. I have a good line of communication with some of the folks there, but there’s not much new to talk about. It feels like we’re in a holding pattern. Until we get rules, there isn’t anything to really talk about, so we’re waiting.
I’ve had a few conversations about this over the last year, and there seems to have been some paralysis with the banks. There’s noise around tokenized deposits versus stablecoins: which one is better, should I do this, should I not. Right after the GENIUS Act there was all this energy of “I need to do something.” Then the tokenized deposit conversation moved to the front, it sounded complicated, and the banks decided to wait and see how it shakes out.
It is complicated. Very complicated. The FDIC has a tough job writing the rules around deposit liability, insurance and everything else. It’s not as simple as plug-and-play technology. It’s really a question of settlement liability, and that’s a complex set of rules to write.
One of the things he brought up this morning: with tokenized deposits especially, and the speed at which they could move, if a bank fails there has to be clarity on the point at which you lock everything down. That’s part of how they determine what liability they have.
That’s right. And if you ask different people what the system of record is, one may say it’s the core and another may say it’s the blockchain ledger. It can’t be both. Those are not simple questions to answer. If you have bank liability represented in token form, when does the real fiat move in the master account? How do you keep the messaging in sync? When do you extinguish that liability? All of that is really hard.
The other big one he talked about was the value of tokenization in collateral. That’s probably one of the more exciting parts. As a bank, you potentially have much more flexibility in what you can accept and who you can help. Unless a borrower is overcollateralized, liquid and diversified, it’s difficult in many cases. Think of a farmer: how do you put your land up as collateral in a way that’s comfortable for everybody?
Or think of the value stored in things like retirement accounts that isn’t used today. In tokenized form it would be much easier for people to pledge, especially as you get into fractional ownership.
Think about maximizing a balance sheet. Take a bank with great deposits but not much loan growth, and another bank with great credit and underwriting capabilities whose loan demand outpaces its deposits. Once you tokenize those loans, you can maximize both balance sheets and let each bank do what it does well. I’m excited to see some of those things.
You’re out on the streets learning a lot about tokenized deposit networks. They’re getting momentum, and there’s a who’s who of things being built right now. What’s driving the consortiums, the banks and the builders to put in the effort? There are derivative products you can build once you’re there, and there’s the speed of intrabank money movement. Is there one thing capturing most people’s attention?
It’s a four-letter word: fear. Fear of deposit migration to stablecoin reserves. That’s the driver today. We’re not at a place where banks are coming to the table with use cases, or if they are, not many.
So you’re seeing a ton of headlines. Cari announced a capital raise. The Bank Chain Alliance just made its announcement. Hazel DTX has announced technology partnerships. The credit union space has some consortiums. Lyric is out there. It’s not noise, it’s headlines. But I still think we have to keep our eye on stablecoins.
The tokenized deposit network situation presents some unique challenges that aren’t in play with stablecoins. The question I’ve asked everybody is: tell me how your settlement mechanism works. How do you settle in Fed master account money? There are different answers. How do you get fungibility? How does your network’s token transact with network B’s token? There are varying answers there too. None of those are obstacles in the stablecoin world today.
Staying on tokenized deposits for a moment before we spend more time on stablecoins: is the biggest threat that traditional sweeps and intermediaries get replaced if they don’t figure out how to tokenize or digitize themselves? Is the early use case just a better version of that?
I think it is a better version, in the sense that I can achieve programmability I didn’t have before and compose things if I’m all on the same ledger. Liquidity will be the centerpiece of it all, and the question behind that is: whose liquidity are we talking about? If I take T+2 out of a securities settlement, who gets the benefit? If I take a day out of settling a real estate transaction, whose benefit is that liquidity? We’ll see. It’s probably the depository bank and the customer.
Is the fear of disintermediation around stablecoins mostly about commercial money movement? The concern that money movement volume, and ultimately deposits, move to other providers? Unless you’re building products with tokenized deposits, like vaults and other novel products for banks, is the tokenized deposit purely about competing on speed? Stablecoins are about network more than anything: liquidity, network and reach. If you build a tokenized deposit network that’s a hybrid of permissioned and permissionless, your network effect is limited to those participants. I don’t know if that’s a competitive advantage if you’re competing on speed alone.
I think it’s beyond speed, because we can do the same things with FedNow and RTP if we choose to. Instant payments are here today. What we can’t do is develop real programmability, where an event triggers something to happen. The things you’re working on are good examples of that.
I keep coming back to Venmo and Cash App and all the things customers never asked for. I think you’ll see products come to market on the stablecoin rail that let customers plug in their bank account and do something they can’t do today. For example, I can’t do an account-to-account transfer between my own accounts at two banks. Somebody could solve that, and there’s utility value in just that.
I think Pete solved that one for us.
Those are the simple things. People aren’t really asking for them, but they kind of are. They’re obvious. But to your point, who gets disintermediated is part of the resistance. It’s also an opportunity for the banks that aren’t participating. Think about how many banks have no participation in payroll. Every two weeks, massive amounts of commercial dollars leave their bank and sit at another bank for a few days, and that bank and the payroll company benefit. In a stablecoin funding world, why can’t the company keep the dollars in its own bank account until the moment payroll needs to be paid?
I think that’s the biggest risk of tokenized deposit network fragmentation right now. If we end up with four, five or six networks that can only transact within themselves while the bridges get built later, then anyone with a brilliant idea is going to go build it with a stablecoin today. That’s where the distribution is. What they don’t have is bank connectivity.
But you can see that coming quickly, because you’re already building it.
I sound like a broken record, but I really believe that within the next two years every bank account will have a wallet address. Once you have that, you have global interoperability: bank to bank, bank to fintech, bank to a self-custodied user. You’ve created the perfect interoperable, cross-border payment network. We won’t need stablecoin sandwiches or on- and off-ramps. You might still have cross-chain movement, and you’ll still have FX with different stablecoins in different markets. But for the basic function of moving an asset from one account to another, the amount of friction removed will be incredible.
That’s also where we need to be prepared from a risk perspective, because now you’ve lost compliance at the edges. You’ve lost the trusted counterparties doing KYC and KYB on both ends. Now I can move money wherever I want.
But a wallet associated with a bank account has different implications than one that isn’t. It has both.
It has both. But you could still send to a wallet that doesn’t.
Just like people do bank account risk analysis, you can do the same kind of wallet analysis. People have been doing that for a while. Know your wallet. People will onboard. The question is whether the protocols or tokens being used will be limited to sending and receiving from wallets that meet certain compliance criteria. You and I talked about the FinCEN rule writing. We’ll see what they say. That’s where it’s going to come down, I think. Somebody will have to deal with that. I don’t know who yet.
I like it better, though. I am still, and I don’t have any other word for it, appalled that the keys to my kingdom, my bank account and routing number, fly around the world. People email instructions, and the numbers are printed on checks. It’s shocking that doesn’t cause more damage. With a wallet address, you can layer dynamic controls and programmability on top. You can build far more controls and protections than we have with a bank account today. I’ve wondered forever why I don’t have a separate sending account and receiving account. That alone would provide more protection than people have now.
I remember back in my Geezeo days, we used to ask why we have a checking and a savings account when you should have a spending account, a pay-your-bills account and a recurring expenses account. Everything was designed around a limited number of account numbers and a cost per account. Now you’ll be able to spin up wallets.
People used to think a database row had a cost to it.
What are you most excited about right now? There’s a lot of movement in the space and we’re starting to see it. Is there anything in particular?
When I think about what will drive stablecoin adoption, AI and agentic activity seem like the killer use case, because you get high utility value along with controls. For someone like me, I don’t care whether it’s a stablecoin. I care that it’s protected and the controls are in place. That’s where I think this is a better solution than most other stablecoins designed for agentic payments.
I’m excited to see where people take it. Not everyone will lose, but there’s so much focus now, and bankers are sitting up and trying to take action. Some of these headlines are telling. I’m just happy to see action. You were talking about fraud earlier. A lot of folks say it’s too risky, fraud is going to be awful, and that’s their reason for not paying attention. I’d argue fraud will actually be better. I’m seeing real products come to light. Maybe they’re not in market yet, but I like what I see.
If you step back with almost no knowledge of banking and look at how we do certain things, like how ACH works, you’d ask whether we could do it better. Yes, we could. We need creative thinkers in here, and it’s exciting to see what’s changing. I agree AI is part of the equation. It’s going to be a hard intersection, and it will put an end to lazy money sitting around in bank accounts.
We were talking to a banker recently who said they built the bank around market-based funding costs, so he’s fine no matter what happens. But banks that have been operating on lazy money, with an efficiency ratio built on it and loans essentially underpriced because of it, are going to be challenged. You have to embrace the inevitable.
Every day I get a marketing piece from somebody with a special rate. A 90-day special, then they drop it to the rack rate and assume you’ll forget and stay. Contrast that with a day when I have an agent working on my behalf that can shop, and banks are open to that agent coming in and asking what they’ll pay me on X, Y and Z. And that money can move, whether on tokenized deposits or stablecoins. That’s a very different mindset for the marketing side of the bank, but that’s where we’re going, and it won’t be long.
We were having this conversation last night: who provides the agent? Do you get it from your bank, from an independent party, or build your own? Does Amazon provide your shopping agent, or do you use some other agent to shop Amazon? Who are you going to align with? To me it’s probably not that different from all the payment methods in our wallets today that we chose based on loyalty or some other incentive.
I’ve got the billion-dollar idea for you. I’m running agents right now for work and personally. I’ve spun up capped credit cards for my agent to try shopping, and right now that experience is horrible. Almost none of the retailers let the agent through.
What kind of credit could your agent get on a credit card application?
Good question. I haven’t tried that yet.
This is probably a little irreverent, but at what point do ADA requirements for websites and content require you to support agents for someone who relies on an agent to do that work for them? I wouldn’t be surprised to see early-adopter retailers create MCPs and environments that let agents transact with their products. Amazon is the biggest offender at blocking agents. You can’t just have your agent go shop on Amazon.
Who’s going to build the headless, agent-first Amazon, where a retailer with a store on Amazon can sell through a new distributed system that lets agents find the best deals? That’s the next billion-dollar idea, I’m telling you. I’m sure somebody’s working on it.
In the near term, could you build it as an affiliate? That’s your gateway.
For sure. I saw someone build a pretty neat Chrome plugin that eliminates fake products. There are a lot of shops on Amazon that rip off a brand or are no-name shops selling junk, basically repackaging products from Temu. The plugin lets you search Amazon but filters those out, using AI to identify the pretend shops. You’ll need something like that for AI agents too, and I think Coinbax smart contracts would be a good fit for know your agent.
What am I excited about? It may sound silly, but I’m excited about wallets being everywhere and becoming the new foundational method of money movement. Look at how much work payment companies did over the last 10 years trying to create interoperability between Venmo, Cash App and Zelle. These are siloed systems, so you had to build workarounds that hold funds in one place and send them to another. As a big proponent of blockchain, permissionless architecture and standards, I think we’re creating an environment that wipes all of that out. You’ll be able to move money, and it doesn’t need to be USDC or OUSD or any particular one. They’ll all work on the same rail. To me that’s exciting.
Once you have that, you layer in agentic payment infrastructure, and it changes everything. In 10 years we won’t recognize what we’re looking at today.
In the short term, think about the account-to-account example from earlier, just for your own purposes. There are so many ways you could use agents to optimize your own finances, and that’s a great place for people to get comfortable that this works.
That call that just came in was my bank. I’m trying to send a wire with some excess money I needed to move for earnings, and I’m two days into getting it confirmed, because I’m in D.C. and it’s coming from an unusual spot.
I have one that’s a week delayed because every time they call me to confirm, I’m on an airplane.
Recently I had to get cash. It’s good to have some on hand; you never know. But banks always get a little cagey when you call and say you need cash. It got me thinking: if you wanted to tokenize your cash, could you have an at-home ATM that knew how much money was in it, so it could become a yield-bearing asset while it sits in your home? Has anybody done that?
The closest I’ve seen is in Europe, where they built an ATM network among retailers. I grew up in a retail family, and my dad brought money home from the store every night as the cash drawer for the next day. These retailers are open later than the bank, so if you need money on your way home, you can see who has cash available, go there and make a withdrawal, and the retailer gets compensated by the bank later. That’s probably the closest network equivalent.
There are probably people with a lot of money in safes at home. Tokenizing their safe money is no different from tokenizing gold that’s still in the ground. You just have to be able to prove it, like the armored car service that pulls a convenience store’s safe and certifies how much is in it.
Or cash will just go away at some point.
We’ve said that. We said checks would go away and cash would go away. They’re shrinking, but I think it takes generational change. We just got rid of the penny. I’m a reasonably early adopter, but I still don’t instinctively pull out my phone to pay at the point of sale. I pull out my card, and my kids pull out their phones. Sometimes changes like that take a new generation that starts on something and stays with it. When mobile payments first came out, there were more steps to pay that way. You really had to want to do it, and tapping a card isn’t hard.
It’s frustrating when you use Face ID or Touch ID on your phone, tap to pay, and still have to enter your PIN. Why is that? A lower-cost transaction?
Probably. Less fraud means lower cost, but the friction is high.
A while back I wrote about the quiet spread. We were looking at data on dollars moving from banks to crypto exchanges, and the most surprising finding, coming back to generational change, was that the biggest tickets and the most dollars moved out of money market accounts. These are not kids. These are people willing to go to the bank and wire money to an exchange to buy a speculative asset. That was shocking.
If you were going to build a customer-facing application today, you’d have to build it on a stablecoin. We don’t have anything else that works. If yield continues to be paid, people get comfortable, and managing wallets isn’t eliminated by the CLARITY Act or some other mechanism, don’t you think adoption comes simply from people getting comfortable? I use it, I trust it, and suddenly it’s not a thing anymore. It’s fully backed. That’s what the reserves are for.
My daughter is 23. She’s been dollar-cost averaging into ETH, Bitcoin and Solana on Coinbase for the last four years, putting about 10% of her paycheck in every month and occasionally selling. A little under a year ago she said, “I saw Coinbase is offering 4.2% on my account. Do they have a debit card I could use with it?” I said yes. She said, “Then why am I only sending 10% of my money there instead of all of it? I’m not keeping it somewhere that pays me zero when I can get 4.2%.” In her mind she isn’t thinking stablecoin, money market or savings. She sees yield on money that’s sitting there, whether she spends it tomorrow or later. That’s a perfect example of someone starting with a speculative asset and turning into a deposit customer.
I think behavior has changed. I’m old enough to remember the good, better, best account strategy, where you got life insurance and other perks with your account. Now it’s: I don’t care what the account is called. I don’t care that I got a $5,000 life insurance policy. What I care about is yield and how I move my money. Give me a card and pay me some yield. I don’t need a toaster.
And back to agents shopping for rates, the relationship ends up not being a bank account. It’s a fintech constantly reallocating your assets to maximize yield within your risk tolerance, which might include the Fed rate, money markets and maybe some lending protocols. Next thing you know, you’re earning 10% at low risk.
Take that to a corporate treasurer with $10 million. That’s a better example. Their choice today is a ZBA sweep: pay the line down, and whatever is left goes into a money market account or a money market fund through the bank. Then there’s settlement timing. If I want yield anywhere else, I have a T+1 or T+2 settlement window, and I may need the money back.
Roll forward, and that money becomes a tokenized deposit or stablecoin. I can take it out to a ledger where I can get T-bills, a money market fund, a mutual fund, a slice of BlackRock, whatever I want. In real time. If I need something to happen over a weekend, it will. It’s all programmed. I’m dragging and dropping how I want it, and if I have an investment policy, it’s queued up.
Is anybody close to that right now?
I wouldn’t say close, but a lot of companies are working on it. Treasury management software companies are moving in that direction. They have to. And banks are paying attention, because if the company’s relationship is with the software company, the software drives where the money goes, and the bank becomes a commodity. The bank then has to get into that business and say, let us provide that for you. But it’s not fully built out.
The questions will be things like how you create incentives. We’ve seen increased redemption requests on private credit funds, and hedge funds typically require quarterly redemption requests. In a tokenized world, will everything have equivalent liquidity? At a price, I guess. Some things are hard to price, like alternative private investments in early-stage companies, which have high price volatility. If you really need to get out, tokenized or not, you have to wait or take the clearing price.
The automation goes all the way back to your accounting system. Today you go to different entities and ledgers to reconcile and get your marks done. All of that can happen automatically. One ledger, multiple assets, atomic settlement.
To your point, if I have to get out, I take what I get. But I can get out, and I get a good mark.
Something could tell you the relative discounts, and then you choose.
And a lot of the function living in bank treasury management platforms today is shifting to corporate operating systems. I connect my bank account with an API, and my operating system does the work. That’s one of the risk points for banks. If we don’t build some of those things, those companies will.
The bank definitely can’t build it alone, but it can get built. The question is whether the bank can sell it.
I’ll give you two answers. First, look at the networks and ask every one of them who’s going to build product. If the answer isn’t “we’re creating an open ecosystem and we want fintechs building on our tokenized deposit network,” I’d put a big question mark there.
It’s the ecosystem. We were talking about this last night. FedNow’s biggest issue is not building the ecosystem. If you don’t open it up, builders will go build on stablecoin rails. That’s just what will happen. There are great ideas floating around out there.
So can banks sell it? I’m a former banker, and I went through a 40-year career where nothing really changed. ACH was ACH, wire was wire, cards were cards. Regs changed, I got new rules, I adapted. This is like going back to school to learn a foreign language as a senior executive. If you’re at the tail end of your career, that’s daunting. How many bankers are conditioned to learn new products every day and figure out what this stuff does and means? I worry about the motivation to get people out of their seats to do that. Many are barely able to sell deposits, period. A lot of lenders are being retitled chief banker or deposit officer. And as I was saying to you, I don’t think a single bank has a deposit committee. They have a credit committee and a loan committee.
Maybe a deposit pricing committee.
Maybe they’ve decided they don’t want any more committees. That would be the pushback: we already have nine, and you want another one? We’re trying to get rid of committees.
That’s probably the most important one. At least re-architect the committees you have. Without deposits, you have no business.
I couldn’t agree more. More deposits.
On that note, today we launched Coinbax Deposits, a way for our banks and credit unions to start reclaiming deposits from the stablecoin world.
Congratulations. I had the pleasure of seeing it privately on the screen. Fantastic. Good job.
Thanks. This was fun.