Stable Takes Episode 2: Tokenization, 24/7 markets and stablecoin wallets

Stable Takes is a weekly mini-podcast from Tempo on stablecoins, tokenization, and agentic payments. Watch Episode 2 above or read the transcript below.

Tempo's Chief Business Officer, Dan Romero, and Head of Market Development, Simon Taylor, discuss the week's tokenization news, including the OKX and ICE joint venture, and why new regulatory guidance is pulling TradFi toward tokenized stocks. They cover Citrini Research's case for a 24/7 internet economy built for agents, when finance will let agents move from reading to writing, and the new personal agent protocol. They close on stablecoin balances, wallets as the credential store for agents and how protocols like MPP open agentic commerce to new neobanks.

This conversation is lightly edited for clarity.

Dan Romero: How are you?

Simon Taylor: I'm good, man. Happy Friday. How are you?

Dan Romero: Good. I was in California this week with the team offsite at an undisclosed location.

Simon Taylor: I know. FOMO was off the charts. I couldn't be away from the kids for another consecutive week. The international world of conferences keeps pulling me away from Team Tempo. But I'll be back with you guys very soon.

Dan Romero: And we have Money20/20 coming up, right?

Simon Taylor: Yeah, I'll be setting off next weekend, so it's coming quick. One more week to go and then the zoo begins in Vegas. And it's funny, the payments conference that was the old-school payments conference is going to be very stablecoin-heavy this year, I think. It's funny how those two things are becoming one and the same.

Dan Romero: Top of the news this week, I think, is tokenization. We were chatting before this, but there's the OKX ICE announcement and then the Securitize announcement, and beyond that, everything that's been happening recently. I'd be curious what your take on these announcements is. Do they matter?

Simon Taylor: Yes. And they follow the SEC's innovation exemption. Suddenly you're seeing TradFi firms that had these "in case of emergency, break glass, this is what we do" envelopes, and suddenly those strategies are going live. I think what that's going to mean is these tokenized stocks coming to the New York Stock Exchange, and potentially to NASDAQ as well. Tokenization is going to get very real very quickly.

One thing I did notice is that this is a joint venture called OKX ICE, which vaguely sounds like a '90s rap group. But it's also a great way for an incumbent to do innovation, right? Go find a big partner who knows what they're doing at the innovation edge, fund a joint venture with them and have that do the thing, and you could always acquire it later and bring it back in. But I didn't see the Securitize thing. What's your take on that?

Dan Romero: Well, I think it's just generally the trend now. And maybe to reset folks: CLARITY didn't pass. I think a lot of folks were hoping that with CLARITY passing, and then the eventual rulemaking that would come out of it, you'd have a long-term framework for how to do some of this stuff.

Even prior to CLARITY, you had people, especially outside the US, doing their version of it. If we were to rerun stablecoins: stablecoins started outside the US and then worked their way into the US in a gray area, and then eventually you had GENIUS, and now they're completely above board from a regulatory standpoint and TradFi is interested. I think we're speedrunning that with tokenized stocks, and given that CLARITY didn't pass, both the SEC and the CFTC have started to release guidance.

Now, I think it's important for people to orient that this can change. Imagine you have a new administration in 2028, with a different set of policies and views on crypto. You could have a change in policy with the stroke of a pen, which I think would be pretty unfortunate.

But I think the calculation here is, one, you directly have the SEC and the CFTC putting out clear guidance. To go back to 2017 and imagine this is what we'd actually be getting from regulators would be a mind-blown experience. But more importantly, if you have it out in the wild for two years, and the total market cap, or TVL, whatever term you want to use, grows significantly, and there are mainstream consumers, especially in the US, using tokenized products, whether they know it or not, then I think it's going to get hard to unwind. It's legal within the current regulatory framework, so even if you wanted to go change it, I don't think it's going to be, by fiat, "okay, we're banning this." But you may have some differences that emerge.

That said, I think everyone's in a race to go do this. Obviously Robinhood and Coinbase are very focused on doing it, and then you have companies like Securitize that have been at the forefront of doing this before there was guidance. And what's surprising to me is ICE. Complete throwback: ICE invested in Coinbase's Series C in 2014. My neighbor used to be the president of the New York Stock Exchange; he's now the CEO of Bullish, Tom Farley. But to see them actually doing something here at the starting line is a pretty bullish signal for tokenization generally, right? Usually you'd see the startups do it and the incumbents would be a few years behind. But for ICE to be doing something in the space with OKX, which is as crypto-native as it gets, I think it bodes well for tokenization as an overall trend.

Simon Taylor: Massively. You probably saw the Citrini Research thing that's all over the timeline. They just did a brand-new piece about how tokenization will inevitably lead to a 24/7 online internet economy, which just happens to be perfect for agents. Agents aren't really compatible with KYC and banking that closes at weekends and other stuff that's not built for them, and you've got this other thing that's just much more machine-readable.

The big thing they point out, the pivot point in their research piece, was what they called the Hyperliquid moment. If you wind the clock back three or four months, during the conflict with Iran, there was a weekend in which Hyperliquid was giving much more accurate oil pricing data than you could get while the markets were closed. And everybody on Wall Street sat up and took notice.

I've been in calls with most of the big-name incumbent financial institutions: buy-side and sell-side firms, custodians, financial market infrastructures, exchanges. That came up consistently in those conversations. And I think it's so interesting that this wave of tokenization, this wave of legitimacy for it, has hit at the same time as this market demand for 24/7.

I was speaking to the head of product at a big custodian bank earlier this week for the Tokenized podcast; that episode will drop soon. She was saying to me, "We've got this two-speed thing that we need to do now, because we're moving to 24/5 in the traditional markets, since we know we need to get to 24/7. We've got tokenization that can do 24/7, but I still need to figure out how it connects back to the old legacy rails, because at the moment it's more of a digital twin than it is unique."

But if you wind the clock forward 10 years, you can see how more and more volume goes toward tokenization. And if you're building a new business tomorrow and all of the stocks are tokens, why wouldn't you just start there? I don't know if you saw that Citrini Research paper and bought the thesis, and where it stands with AI.

Dan Romero: So I think two things there. One, I agree with you. If you're starting a new fintech, whether that's a consumer app, a business-oriented app, anything fintech, stablecoin rails can provide you global access instantly. You probably do add fiat as well, but just to get started, you can run really fast with a product in market on stablecoins, with the ability to link a debit card to it and an earn program. We see that at Tempo: that's basically what everyone wants to use stablecoins for. It's this modular financial stack that you're able to boot up pretty quickly.

And I think tokenization will be the same thing. A lot of fintech apps traditionally looked more like a bank account, and then stocks would be a completely different thing. In a world where you're effectively crypto-native under the hood, even if you're abstracting that away and just saying "digital dollars" with stablecoins, we know the ability to rapidly add a set of financial services on top of one stack is going to work. Just think of the number of new startups that are going to get started over the coming years on that alone.

And then agent-first. I think we talked a little bit last week about Amazon blocking agents, and I'd be curious how banks are going to do this. To the point from last week, some banks will lean into it, and I think a lot of banks will say, "Whoa, we don't want agents," in the same way that when Plaid first came out, they said, "Hey, this is our data." Even though the consumer wanted to use it, they didn't like the leakage. But I think the writing is on the wall: consumers are going to want to be able to interact with whatever set of services they use in their life through agents. So I think these new services that can rapidly add a set of financial services that are programmable, so totally agent-native, will be really important.

I think the Citrini piece is interesting because it tried to say, if you want to get exposure to the agent economy, here's a set of public stocks and here's a set of cryptocurrencies. That's interesting for people in crypto who've been through a winter and are wondering whether people are even interested in crypto anymore, with AI on top of everyone's mind. And then obviously there's a component of, okay, OpenAI and Anthropic are going to be public, SpaceX is public. So there's going to be a set of public equities that are maybe less fintech-oriented, but because they're going to control the primary consumer interface, or a business's interface, with the agents, they're going to be able to route transactions within this agentic economy.

I think it's interesting. I didn't buy his basket, and maybe they had a position in it to begin with. But I do think it's an interesting exercise for anyone asking, "If I believe the agentic economy is going to grow, how do I get exposure to it, and what's the right weighting?" I don't have the answer for that. But I do think about it in terms of: there are going to be the fintechs that benefit from this, there's going to be some amount of cryptocurrency if you buy that, and then obviously the companies that actually own the agents are going to have a lot of, to use a Stratechery term, aggregator power in terms of where they can route.

Simon Taylor: I'm so interested in who wins, because I agree with you. I thought the basket was almost a finger-in-the-air piece, and I don't do financial advice. I don't know if I'd agree or disagree with it. But there are companies leaning in. Robinhood has its own agents, but they'll let you connect to their MCP server. I've seen that OnePay by Walmart does that now, and then you can use Plaid, or there are the direct integrations.

So there's this interesting tipping point in finance in particular, which is when you let the agent do more than read-only and let it start writing. We're starting to get to that tipping point. Usually the agent can write, but it's inside a little sandbox. And when it's writing inside that sandbox, it's touching TradFi rails inside a secure login sandbox.

I've always struggled with the lack of intellectual honesty that goes, "Oh, well, this will obviously live on DeFi rails." I'm like, I don't know. I can definitely see the benefits, especially for machine-to-machine payments that are ultra-low latency and need to be really fast. We're already seeing that, right? When performance really matters, which it does to machines, I think stablecoins win. But a developer who's just giving an agent credentials is probably going to be better off with browser use on the internet with a card for the foreseeable future, buying most stuff and things in the real world. So I don't think you can say cleanly who wins or who doesn't. It's more of an attitude thing: who gets to lean in a little bit more?

And then what does the front door look like? I don't know if you saw this piece from Sierra, and I think it was Instinct and Muse and Stripe and Shopify and a couple of others. They announced the personal agent protocol, which, by the way, I love the abbreviation of: PAP. It sounds like you're knocking on a door or honking a horn to get in.

But it's really simple. It's a really simple way for a business to say, "Okay, here's my MCP server, here's my website, here's my API. If you want credentials, then I want to see the consumer's credential, and we're going to do a little handshake through OAuth, and then you're going to come in the front door and I'm going to give you scoped access to everything behind that." And it's weird, because that's not KYC, it's not KYA, it's just a little OAuth handshake. But it's the business saying, "If the consumer can OAuth through to me, then I'm going to give you access to those underlying rails."

It feels like such an interesting little moment, because Muse and Instinct in particular are in it. With those two personal agents, all right, it's a bubble; most people out in the real world, outside of the timeline, have never heard of them. But if there's any area of traction for personal agents, it's there. And if there's anything agents have been getting stuck on, it's that the front door isn't quite there. It's computer use at best. So I'm interested in your take on this one.

Dan Romero: I think it's extremely interesting, especially given the players. And what would really dial it up is if you get OpenAI or Anthropic also jumping in. Then effectively you're going to have a de facto way in, in the same way that MCP organically emerged and then everyone was using it, just by virtue of the distribution on the personal agent side of things. You have Meta, and Instinct, a startup but growing extremely fast, and then you have OpenAI and Anthropic, and you can even throw Google and SpaceX in there. If you get the core consumer agent platforms all saying, "Hey, we're going to adopt this," as, to your point, a lightweight way of authenticating to say, "Hey, this is an agent from a known company, and we're not going to abuse the system," then I think that only accelerates the use of agentic commerce. Whereas now, going back to the Amazon example, everything is a bit of computer use: how do I avoid bot detection? Instead, you embrace the bots for when you're doing a bot transaction.

And then what I think gets interesting is if these consumer platforms, who have the direct relationship with the consumer, start doing transactions. We've heard this from Mark Zuckerberg; I think he's said on a couple of podcasts that one of the ways they're thinking about monetizing Muse is to actually be in the transaction flow. That becomes a very interesting conversation in a world where, going back to how fast you can build a neobank, stablecoins start to be interesting. "I want to go build a stack where I can have the consumer hold a balance, because that makes them more likely to do a transaction. And if that transaction happens, I potentially control more of the economics, as the harness, as the direct relationship with the consumer."

So to steelman how stablecoins and crypto find their way in there: again, I think it's probably as infrastructure. I don't think the consumer is thinking, "Oh, I'm using this specific stablecoin on this chain." It's more, "This just worked for me. Maybe I got a discount, 1% cash back, for using the balance rather than a card." So I think that offers an opportunity for folks building with stablecoins: if you can make that balance attractive. There's probably some earn program there. And especially internationally, the ability to just hold your balance in dollars is a huge feature. So that would be my bull case for how stablecoins find their way in there.

But I do think the train is leaving the station right now, and making sure stablecoins can actually work and the UX is really great, especially abstracted away from the consumer, is pretty critical.

Simon Taylor: It's going to be key. And don't forget, with a stablecoin balance you can just stick a card on the front of it too. So there will be some places in which I want that stablecoin earn. I might even have merchants that can accept them directly, but more under the hood, right? It's not necessarily a different credential that's going in. And certainly the machine agents would probably be more likely to do that. And everywhere it's not accepted, well, here's a card, but it's attached to the stablecoin wallet.

And I think in all of this, the wallets become really key. I've noticed how quick out of the blocks Stripe, Shop Pay and PayPal have all been when it comes to Muse and Instinct integration. They're straight in there. And I think the wallet could, over time, become much more important as the credential store, with the ability to delegate authority to agents: what are they allowed to go do on your behalf? And if you add the personal agent protocol into that, could the wallet help direct some of that, or does it live inside the agent?

I think there are some UX things we still need to figure out, like that path between Dan and his agent and the delegated authority to do a variety of things. But this is a really great starting point for that.

And you might one day get the dream, if people are more and more interested in stuff that just works and you're building from scratch: "I've got tokens that move 24/7 and can show up in any wallet. I've got dollars that are 24/7 and can show up in any wallet, and I've got increasing adoption of them. My agent finds it a lot easier to move between these than it does the traditional stuff." And that, I think, brings you back to Citrini's point: you can actually draw a very logical path to how this happens. I just think it's still a little foggy in terms of how we get there.

Dan Romero: I think you said something key there. Obviously PayPal had its moment and its use cases, and it's still large as it relates to the internet. But most transactions, for e-commerce or however consumers do internet transactions, still involve you typing in a card at some point, whether that gets saved as a credential with Amazon or Shopify or someone else.

But I do think agents may prefer to use wallets in a world where the wallet is doing an OAuth-type connection, or using PAP and MPP, with this granular set of permissions and controls. It could very well be using a card as a payment method, but you've moved it up one layer, where the card is a source rather than the primary. And then the agent can say, "Well, I have a wallet that speaks the right set of protocols to get the job done," and I can have a different set of funding sources. The wallet itself can then compete to be the primary place for balances, offering an earn program and issuing a card against it.

And I think that is the advantage for neobanks, or any type of new consumer service or financial service built on top of stablecoins: it's modular and interoperable from day one. And all of these protocols, whether it's MPP or PAP, are going to play nicely with stablecoins. So if I'm some hot new neobank in a country in Latin America, I can now plug into this whole global e-commerce system. In the old world, you would have had to do extensive BD to reach the same level as PayPal. You might be dominant in one country, but getting into all potential merchants on the internet is hard. But if we've made everything interoperable with MPP, PAP and wallets, then you become a funding source, or you do it directly. And so now everyone can compete on just one question: am I offering the best primary place for a consumer to keep their balance and spend from?

Simon Taylor: And I'm so interested to see what the UX innovation looks like for that. I don't know that I've seen anybody do anything that's more than a better neobank. Better because it's built on stablecoins by default, because it lets me go more global, arguably a little less foggy in terms of what's under the hood than the old banking-as-a-service model, a little cleaner in a lot of ways and faster to market. That's been it, really: more global, more instant, cheaper, but with a card on the front of it.

Can you fundamentally reimagine the agent-first experience as a neobank and build something that's agent-native? I don't know. Maybe the next Y Combinator class will show us some stuff, but I'm excited to see what that could be. Cool. Good to see you.

Dan Romero: Good to see you. Bye.