The Agentic Economy: How AI Agents Will Transform the Financial System with Circle Co-Founder and CEO Jeremy Allaire

And got really excited, like if we had like a protocol for dollars on the internet, um, that, you know, potentially we could have a way to store and move value, you know, instantly, globally, frictionlessly at at no cost. ultimately. Um, the other idea that we were really excited about back then was this idea of programmable money and the idea that um eventually these these networks, blockchains, would become like operating systems and you could actually have machines.

that intermediate economic activity and financial activity on the internet, and including like autonomous software machines and Um, back then we didn't have generative AI or anything like that, but this sort of idea of kind of commoditizing the kind of payment utility layer um with like very safe digital. Dollar digital currencies. And then having like programmability of that with machines that are kind of tamper resistant can run on the internet.

That's what kind of drove the founding of the company. And and the view was like, if we could do that, like we could actually improve the financial system, make it safer, make it more accessible, um, make it more efficient. um and and kind of derive new utility from money that we haven't had before. And so that was sort of where

or we're not dollar centric. So for example, Bitcoin was in part a response to the great financial crisis. Yeah. In the view that all sorts of weird bailouts happened there. And therefore we needed some alternative sort of financial infrastructure for the world.

There has to be a way to build like a safer uh financial system. And the key issue there was um I was interested in this idea of full reserve money. And in some ways Bitcoin is full reserve money because you you, you know, you you kind of there is no way to fractionally lend. Bitcoin per se.

And in the nineteen thirties, there was a really big debate about like what's the right construct for the banking system and the financial system. And um, there was a a proposal from a group of economists uh called the Chicago Plan. And uh the kind of ringleader was a Chicago economist. Actually, it might have been a Yale economist or Princeton at the time, but uh Irving Fisher, who wrote a book called 100% money. And that idea was that

Full reserve money was essentially um, you know, government obligation money. It's so it's still the obligation of the government, like the US government in that instance. Um, but that essentially, um You can have that and you can hold that, but you can't take that and then um

uh fractionally lend against it. So you have kind of a full reserve, and you can only lend full reserve money. And so that um was a a big proposal for how to structure the the way the financial system worked. And It was actually um lobbied very, very hard uh against it by the bank.

And the banks really liked fractional reserve. They liked to be able to have the inherent kind of leverage and risk taking and instead convince the government to establish or they collectively with the government's sanction uh established a insurance company.

uh called the Federal Depository Insurance Company Corporation. Um and so that was a kind of corporate uh insurance model, but the risk taking still existed. And so we've continued to kind of face those issues. The great financial crisis was an example of 30X leverage, 12X leverage, 14X leverage against against these sort of base layer. And so my philosophy has been, well, um, right now,

In terms of general utility, um, our our existing uh economic system, like it d it does depend on um really m major reserve currencies like the dollar. And my view is like that's gonna continue for a while. uh maybe 30, 40, 50 years, it'll continue for a while. But what we want to do is construct a system that is in fact safer. So a full reserve form of money. And that's what stable coins are. That's what dollar stable coins are. And in fact,

with the Genius Act that passed last year, it's sort of codified in law. Like you can't do anything with this. It's like this very narrowly bound, narrow money kind of model. And so I think in some ways like that original vision we've now got established in laws around the world. Um and and now we have to do more with it. We have to make it extraordinarily useful. And you can l you can lend that form of money as well, but it's just that you can't do fractional reserve.

But you know, sort of fiat stable coins in this way, we're back that way. Now laws have now come into play in major jurisdictions, whether it's in Europe or Japan or the US, um, et cetera. And we've been following the whatever laws apply to us, uh, you know, whenever they apply to us, obviously. But um what that's really led to is a an architecture, which is basically what's now federal law, which is um really holding only short duration US government treasuries.

Um, or uh treasury collateral that is uh overnight with with like global banks. Uh so that's you know very safe overnight treasury collateral for cash. Uh, and then, you know, some amount in cash that is for kind of immediate liquidity. Um, but in that case, it's sort of holding it in these sort of big custodial institutions like Bank of New York that holds hundreds of trillions of dollars, et cetera.

of assets. And so um that is essentially the architecture of USDC. And we we're very transparent. We have daily transparency onto most of it through a a system we set up with BlackRock. But um

you know, AI agents that are paying for uh the output of essentially the AI tokens of another AI agent. And they're, you know, spending, again, just, you know, uh a dollar, fifty cents, twenty cents, et cetera. So super tiny transactions at one end, all the way to the largest electronic trading firms in the world that do huge amounts of of capital markets activity who are

you know, settling multi-hundred million dollar transactions. And the powerful thing is it's all the same. Just like, you know, if I send you an email, uh, you know, the and my email is like, hey, this is what I had for breakfast. the payload of that is the same as if I sent you an email that had like a CAA dossier attached to it. Like

USDC doesn't care. You know, so as a as a general architecture, it can be used across a huge range of things. And we have everything from merchants in Stripe and Shopify that are using it to Visa actually using it themselves. to actually move money on their own internal network instead of using the legacy banking system to lots and uh lots of kind of neo banks, remittance companies that are using it as a way to move value.

You know, a great uh uh kind of B2B uh uh fintech uh ramp just yesterday launched, you know, USDC as like core to their treasury system. You can use it to pay invoices, pay

Basically, this is just internet native and it runs on internet protocols. And so it behaves the way that any piece of data or content or behaves on the internet, which is what our expectations are. Mm yeah, most people's expectations are.

And so we're now exporting digital dollars and we're doing that all around the world. And that's like strategically important uh to the United States in in from a geopolitical, geoeconomic perspective. Um, but you know, there's other things too, which is this comes from my own background uh as well, which is

These are this is programmable money. Um, there's never been programmable money. Like you actually have uh essentially like our stablecoin network is just a public API on the public internet that anyone can plug into and use. And so if I'm a developer and I want like Global d dollar settlement and I want to provide that as a capability to my users. I don't have to ask permission. I can just go.

Connect to that smart contract, connect to that uh public API, and boom, I have now an application with global digital dollar utility. And so that's really different.

Yeah, you know, the the idea of programmable money was like again this early idea that we had. And smart contracts um was sort of the original expression. But when I looked at that 13 years ago, my view was that.

Blockchain networks are operating systems. They're going to be operating systems. And so when we think about operating systems, we've lots of paradigms for that. We have mobile operating systems. The web was itself kind of an operating system with a runtime and a language model and an object model. And

You know, clouds became kind of like these big virtual operating system environments. AI foundation models are now essentially operating systems that execute tasks and other things. Blockchains are operating systems. And they have compute engines. They have virtual machines. And you can write Turing complete code. You can write software that runs on these. But there's some really key attributes that make them different. So the the first is that the code is

It's sort of tamper resistant. It once it's published, it's sort of like out as like a machine that's tamper resistant. The second is it's perfectly auditable. You can audit every single input and output of that machine of that code in real time. Right. So it's like all the compute is public accessible. It's open source by nature. And um and and that's really powerful as well.

And um and you know, it it also has these sort of um uh essentially um kind of transaction and compute integrity assurances. And this is really key. And it ties back to AI as well, which is like You want assurances that the machine is doing what it said it's going to do. And you want kind of the inputs and outputs to be provable to and the state of the machine to be provable. And these are these these are things that

Um, it it was not easy to do uh in the past. And so these these network computers, um, these operating systems now provide for that. And as we're moving into the AI-driven economic system, right. having those mechanisms becomes even more important. It happened to be important for financial transactions where, you know, integrity, proof, audibility, verifiability are like intrinsic in a fiduciary apparatus.

That was like really key. But now, when we're dealing with uh, you know, kind of autonomous uh actions uh in the economy, that also becomes extremely

And you had these economic games that were multi-turn economic games to some extent that these actors could play. And so we said, isn't that a great place to basically evolve intelligence? Right. Because you have these multi-turn games, you have economic incentives, you have game theory. You learn all sorts of lessons off of that. Obviously, there's a very different world now with sort of generative AI and foundation models.

But I'd love to hear your view of where is uh where where are Agentech payments going? Yeah. And is it gonna be crypto? Is it gonna be more just traditional banking systems? Is it a hybrid? Like what do you think are the drivers of that?

I think um, you know, that shift is is effectively going to mean that um a couple things in my view. So the first is that um, m more and more of the actual work that is done in the real economy, um, especially in uh in in you know, kind of the what we call the white collar economy, but uh but in in many, many areas of service and delivery and and and so on, like so much of that is gonna be conducted by AI.

Um, well, we don't have uh an infrastructure that can support that. We don't have an infrastructure that can um work globally interoperably, instantly, that can be uh uh programmed uh through software layers by uh arbitrary pieces of software. That doesn't exist. We need an infrastructure where the agents themselves can uh dynamically create and spin up uh different kind of uh uh financial endpoints themselves.

We need transactions that can scale potentially into the, you know, you know, billions or trillions of transactions. We don't have that. Uh, we also need um we need the ability to kind of um handle transactions at micro scale as well. So uh, you know, for example, consuming uh a certain amount of intelligence might be five cents or 10 cents as it is with these.

And so we need we need that to work. We need that to work in real time. Again, between any piece of hardware, software anywhere in the world.

And um transaction volumes have grown incredibly. And um off of like a a monetary base, it's also growing, but the transaction volumes are growing way faster. And that's because money velocity has picked up. The cost to transact is now subsent reliably. And so when you take out the cost, you can do more transactions.

And so with Arc, which we can come back to, you know, we we now have an infrastructure where we can conduct transactions for a millionth of a penny, uh, which just was never feasible.

Um, my own view is that uh, you know, a agents um and and you know, seeing what happened with OpenClaw and Multbook and all this stuff is all really interesting because It showed that you could actually see emergent forms of of cooperation, of, of interaction, of of uh of of engagement amongst AIs. Uh, and that's pretty powerful. And clearly like

We're at the front edge of that. Like there's gonna be a lot more of that. And so um, if you have AI agents that are from around the world, they could be generated from lots of different models and LLMs and the like, um, and and they need to kind of coordinate. They need a medium, a trust a trust, a trustworthy medium where they can do that, where they can instantiate an entity, where they can store value in that entity, they can execute and arrange contracts.

uh that intermediate the work and the tasks and that where all of it is real-time mathematically uh and computationally provable. And so blockchain infrastructure now actually gives us the building block. for when I say agentic economic activity, most people think, oh, that's e commerce or payment, it's not.

Agentic economic activity is actually how does the organization of of what we used to talk think of as labor and capital, but essentially like kind of how does this organization of kind of compute work? uh happen and and what kinds of kind of corporate forms might emerge uh in in that world to do that. So I I'm actually quite interested in that. That does tie to ARC because that's a a design surface that we care about, which is basically

Um we we describe ARC as an economic operating system. And this goes back to a comment I made earlier, which is These networks are operating systems. And um we're we're moving now from the kind of like early adopter era, which you're very familiar with, which was mostly around like, you know, speculation on different things. There were some interesting things like NFTs or whatever, but like

We're we're now moving very squarely because of stable coins into like the real economic activity side of this. And um, and and I think my view is that. As we go forward, the the substance of what we think of as contracts, the substance of what we think of as corporations are going to be software machines themselves.

And so we're gonna we're gonna see this progression. And so ARC as an economic operating system is conceived of as a compute environment for laying down all of the building blocks of economic activity, whether that's Storing value, moving money, uh, or instantiating a corporate form, or manifesting and intermediating complex contracts.

Like a lot of this stuff which was conceptual a long time ago is now like real. And we have a legal basis for it. We have a regulatory clarity for it increasingly. Uh and it's interesting is is that the drivers of this machine economy are actually machines. Uh and so, you know, our our view is um uh You know, ARC is designed for this moment, which is a a moment when machines are going to play a larger and larger role in all of the output of of of of uh of the economic system.

Ways to build smart contracts, to build a lot of the types of things that you just described. What do you think is the difference between some of these more traditional L1s or blockchains and sort of what you're doing at Arc?

Uh a lot of it was sort of like, hey, we're gonna build something that is completely um, you know, censorship resistant or outside of the reach of governments. It's sort of like we're building an alternative universe. And um, and that's like the goal.

Um and and I think, you know, decentralization is itself a good goal, but I think as we move from early adopter to sort of mainstream scaling, where, you know, whether it's, you know, a major company like Walmart or or it's, you know, a household that's thinking about like how they store their wealth.

The the the intermediaries, and there will continue to be intermediaries. We're not all going to be your own bank. Um, the intermediaries um have obligations in terms of the the the the kind of like robustness of the infrastructure that they have to run. And so ARC is actually set up. with a number of features. One is that it's actually a known validator set. And so um the the infrastructure operators of ARC are major financial infrastructure companies.

And so, and and those are companies that are held to these very high standards for InfoSec compliance, reliability, availability.

The bad guys aren't running your transactions. Um, and we can also run assurances that transactions actually have settlement finality. Like they can't be hard forked, they can't be reorged. Um, and so you can get what's called deterministic settlement finality. And that's really important whether it's a security or a piece of cash or whatever it is, in, you know, in essentially hundreds of milliseconds. And that's really important.

The the other piece is that it's sort of designed with real money uh as the foundation. So there's not like a volatile gas token. USDC is actually the default native token, which is now under the law essentially like a legal form of electronic gas. So you have real dollars as the way that people understand. And so to a company that's like doing this, it's like I pay AWS credits.

Uh, I understand how to budget for that, my treasury, my operations, my compliance, et cetera. So this allows basically for like. actual usage to be make sense both to the user, to the developer, to the corporations, the FIs that that deal with this. So so that's really important. And then I think the other is like Um, we've been building in a lot of of uh primitives that are important to like the way that payment systems work, the way that capital markets work.

uh the way that you know the the privacy requirements that are needed in some of these cases, but still allowing for like compliance to happen. So We've kind of purpose-built this for a different uh kind of set of participants who need to run on it. And we've had the advantage as Circle.

Of working with many of the leading financial institutions that are getting into this space over the last couple of years, whether it's the Visas or the Black Rocks or the Bank of New York Melons or all these types of companies. So we've had we've been able to work with them and we've been able to work with governments around the world to hear like, hey.

like central bankers all over the world. Like what's important as you think about like allowing this internet infrastructure to run the financial system? And we're kind of trying to incorporate in a lot of the kind of requirements in the sense that they have.

And so that's that's very different. I think it's a different design space. And I think um, you know, these new distributed network operating systems will will need to support like the real economy's activity, not a kind of shadow economy. And so that's just uh substantively different. There are a lot of other technical things I could talk about which um that that are part of it, but those are I think helpful just in terms of framing them.

Um, on the infrastructure side, there was EK roll-ups, there's a variety of approaches over the last few years. Besides stable coins, is there anything that you think was especially um interesting or that will be impactful or a lot of these things kind of infrastructure looking for solutions? Or I'm sort of curious how you think about crypto writ large right now.

That's a as a as a general utility space, that's like super, super interesting. Uh-huh. Um, and so a lot of the a lot of this has been like, okay, well. How do we make sure that that can scale? And so as an example, right, in a in a world of like billions of AI agents that are swarming and doing other things, like scaling this is actually it's

Extremely important. So ZK, you know, roll-ups as an example, or zero knowledge proofs more broadly as a way for um proving compute, which allows you to do compute off off-chain and then prove it to the on-chain. That's actually really important. So you're these sort of off-chain or trusted execution environments and other things that provide cryptographic proofs of compute or or or uh other kind of uh

Assertions becomes really key. So that actually a lot of that research is now becoming extremely valuable. Um, the same thing goes for a lot of that research um and development is critical to um enabling privacy. And so we want, you know, we want all the benefits of of kind of open, interoperable, kind of permissionless infrastructure, but we also want to be able to have privacy.

Um, like corporations don't want everyone to see what they do, or we don't want to be doxxed and like all this kind of stuff. And so the that's now coming into into uh real production. That was like researchy for a long time. Like Arc day one is shipping with like built in privacy primitives, uh, which is which is again the result of a lot of work for a long time.

Um, and so those are those are important pieces. And then um, you know, I think uh as kind of more large-scale financial infrastructure comes over to this, as we move, you know, where the New York Stock Exchange or the the biggest derivatives, clearinghouses are like, yeah, we're gonna move to an on-chain world. Like,

the scaling stuff becomes really, really critical. And so um I I actually feel like now more than ever, like those big work streams uh are are are coming online. And, you know, if I use as a reference point like you know, I spent a long time building on the early internet and the early nineties and the early web and like all this stuff, all the way up until like two thousand one for like, you know, for me it was like

10 years. And it was still, it was like awful still. Like it just like you kept grinding and it was like, how do we make this useful? How do we make this useful? How do we make this useful? And then you had, you know, a whole bunch of things happened that were in the background. Like, you know, Wi-Fi, broadband, you know, uh you finally got like usable um other internet connected devices and you could actually really start to do stuff so you could actually

actually deliver software over the internet. You could actually deliver media over the internet. You could actually do communications, like real time communications over the internet. But it was like ten years in the desert or longer before you could even get there.

And I kind of feel that way about the blockchain space. Like it's been a dozen years or so and and now we're kind of having like the broadband moment and the demands of society and the financial system, the agentic economy, all of that is sort of coming together at a really interesting time.

When do you think that stuff will happen and what approach do you think will be taken? And how does that tie into the agentic world?

We also uh, you know, have seen this growth in like tokenized money markets. So basically like on-chain treasury bills. We actually operate the largest tokenized treasury uh product called USYC, like US field coin. Um that's uh that's grown quite fast as well. Um And we we run the largest tokenized Euro as well, uh, EURC. And so we're we're definitely like looking at this this broading out. I think um

There's a huge effort right now at every layer of the whole financial system stack to go into tokenization. So all the way down at the layer of like the people that keep the records of the stock. Which is like, you know, if if you're familiar, like the computer shares of the world, up to like the the layers that like are the depository clearing systems like the DTCC, which you know, most people don't know, but it's actually like the back plane of how all

And as we speak, the SEC has been providing clear guidelines on how to do that. And so they actually issued guidance just about a month or so ago that basically said, here's what you do in all these layers. Here's, you know, here's your obligations. And so we're at a point where like technology and then the market's desire is creating that. And right now, the interesting thing about things like tokenized stocks is um mostly it's interesting to enable people not in the

to access these. Um that's where the a lot of the growth has happened because not everyone has access. Uh you know, if you

I think um what becomes a lot more interesting or here's the game. It's this game that used to be on a C D and now I can you can download it or whatever. I think the really interesting thing is like what can you do that you couldn't do before? What kind of utility gets unlocked, whether it's fractionalization or

how you can uh how you can borrow and lend on these things or um how you could package them together in different ways and and enable and AI could play a pretty significant role in that as well. Yeah.

And uh yeah, I mean we we're we're seeing that. I mean USDC powers polymarket, for example. Um, and uh and and so the same guys that are, you know, trading derivatives over here on on, you know, oil or Bitcoin over here are also like, I'm moving my money quickly using USDC over here to like figure out what's gonna happen in some uh event.

basically tying proof of work into just generic inference work. Yes. In other words, can you tie those two things? So you're being very um GPU efficient in terms of what you're doing, but also you can effectively generate incremental revenue through GPU usage while you're using it for inference or other purposes.

in in a sense, the the energy is is is waste. And and so I think the idea of uh essentially like inference compute uh uh as GPU, uh inference compute as proof of work. And so the work itself is the inference. And uh that as the underlying basis for proof of work uh cryptocurrency is pretty interesting. um and uh would would be, you know, potentially something that could align with the kind of monetary principles of something like Bitcoin. Yeah. But actually uh

be productive, uh productive proof of work. That's really interesting. And so, you know, I I My my my own view, and this is like, you know, I think goes back a long time, is like the, you know, people have kind of axiomatically sort of assumed like, well, Bitcoin is the the the the

The thing, it got the network effects, it has all of this. Um, and I've always said, like, I don't know what we're gonna be using in 10 years. Like, we don't know. Now, um, Bitcoin has lasted a really long time. Um, but I think like the paradigm shift. that we're seeing in in um in energy infrastructure.

in the performance of of you know the conversion of energy into intelligence and the compute layers and that, it certainly opens up a new avenue to think about this that wasn't readily available, you know, 15 years ago.

Um, what does that world look like in your mind? And obviously we're going through a period of intense change. Yeah. I'm finding it incredibly hard to predict the future right now in terms of just what's going to happen in AI, much less AI plus crypto, plus the global economic system, plus everything else. So given all that, and putting that aside, yeah, exactly. Um, what is your vision of the future?

like, you know, the diffusion limiters are are um in in in some in some cases bureaucratic, in some cases legal, in some cases human risk or other things, right? But we have we have these limiters that are there. Um, but it does seem like Um the the pace of diffusion is accelerating and will continue to accelerate and um and and that's pretty dramatic.

And and so I guess like m m my own view, um, it's it's very rooted in my own like political and economic philosophy is is that um We we have a real opportunity to uh create um essentially new new social, political, and economic organizational structures. Um, and we and in in many ways, like we have to. There's a a a kind of um, you know, in these periods, whether it's the the Enlightenment and and the Industrial Revolution and other things, like there's these periods where

Um, there's like a a new definition of the social contract. And and that new definition of the social contract. is then in turn reflected in social, political and economic ordering and the mechanisms that we use for those things. And I it feels like to me like we're gonna be we're gonna be forced through that. And um and I think, you know, that's simultaneously like terrifying and exciting, et cetera. And and and I am of the view that

we are going to have a a uh a kind of lag effect between the disruption and the establishment of those new institutional forms. Um, but at the same time, I actually believe like new institutional forms are going to be um emergent out of this. And so as I talked about earlier, like, the formation of these kind of on-chain organizations that have different forms of of governance and contracting and

uh and and a mixture of human and and agentic uh actors like that seems like we're gonna have a lot more of that. We're gonna probably have huge proliferation of that. And it maybe that those corporate forms um are like the most productive corporate forms that we've ever seen in economic history. And and then, you know, the thing to be kind of like an overlay into the governance systems of

of uh p you know political uh or organizations and and systems as well. And it um yeah, so I you know, like m my view is uh We're going to have simultaneously all around the world, um, a renegotiation of the social contract. And it's gonna be uh it it it is gonna require um new systems of participation in economics and governance that we haven't had. Um that's like very high level mumbo jumboe, but it's also um, you know, sort of how how I think about it at a high level.

Agents negotiating policy. So it's kind of this interesting view of how can you spontaneously spawn these sorts of systems from a governance perspective, which is kind of cool.

How do you think about the GDP impact of AI? So if you think ahead five years. Yeah. And you know, what what do you think the global economy is? Ten percent bigger, fifty percent bigger, three times the size? Does that even matter as a metric anymore? Like

really drives a very significant discontinuous jump in in in GDP at an absolute level. It'll have, in many ways, probably less meaning than we've historically had with GDP. Um, and you know, the kind of uh economic well-being indexes that we think about, like GDP uh will be you know, the the risk here is that GDP. Uh effectively, like the GDP growth is a sort of capital capturing more capital at the at the expense of humans. Like that's the real risk.

Um, and so like GDP growth generally has been like a really great thing. Um uh and and so the the question is is like is will it remain a great thing? Do we have the new social contract uh to to to deal with that? Yeah. Um, but m I guess my my general view just sort of as a technologist, talking about, you know, seeing what we see with diffusion and and kind of other other changes like

Um, it does it does feel like uh we have the potential for double digit GDP numbers in the 2030s. Like that's that seems not unrealistic to me. Not that that's going to be uniform all around the world, but certainly um in large large parts of the world that seems very uh yeah achievable uh based on what I see.