Why AI Agents Will Need Their Own Bank Accounts — Lincoln Murr, Coinbase

Thank you all for attending. My name is Lincoln. I'm the head of AI product at Coinbase, and today we'll be talking a little bit about how we're building the agentic financial account. And I know what you're probably thinking, Coinbase, crypto, AI, lot of buzzwords. But my hope is by the end of this conversation, you have an understanding of what agentic finance is and that it's a very real thing that's happening. You understand a little bit about what we've learned through this process of building this over the past two years, and hopefully can take that and, and distill some

insights from it that are helpful for you. And I'm gonna talk about, a little bit about where I think we're heading with this whole narrative and how agentic finance will evolve over time. Cool. So to understand first, uh, where, where I'm going with this, I think it's important to look at what is agentic finance. And in my mind, agentic finance is anything that has to do with an agent that can manage money or has access to a financial account. Kind of obvious, but what does that actually mean? There's really two use cases that we're seeing. It's agents that need to pay

with money and agents that need to grow money. So the grow is a little bit more of a nascent category. So I'll start by diving into the payments piece. And to do that, I think it makes sense to start around two years ago with our first exploration into agentic finance, which was this tool we created called AgentKit. The whole idea was prior to MCPs or really any sort of standardized AI tooling, we, uh, used the LangChain framework and added a wallet to that, and we gave AI access to money, in crypto particularly. And at the time, this was

a pretty novel thing, and I remember it was like two AM, I was typing into my agent, like, "Hey, can you send money somewhere?" And it did it, and I was like, "Whoa." These days, that's not impressive at all, but at the time it was pretty fun. But as we gave this to our developers, we noticed that there is one core problem. We gave an AI access to money, but we didn't really give it anything to pay for. This was a bit of an error on our part, and it's something that we wanted to solve, and we were able to do it by standing on the shoulders of giants. How many of you are familiar with HTTP status

codes? Hopefully... Yeah, most of the room. Okay. So for those of you who aren't, it's basically a standard that defines how websites communicate and how, uh, information is transmitted across the internet. One very popular one is called 404. It means there's an error. 200 means that there's a success. But there's one that actually has been in the internet spec for decades that nobody really ever implemented. And in fact, it was called the original sin of the internet by Marc Andreessen that we never properly

implemented this. For those of you who know a little bit about what we work on, you would know that this is the HTTP status code 402, which means payment required. And the official definition of this in the internet spec is reserved for future use. Pretty generic. But we have a great opportunity here now with the inflection points of AI agents needing a clean, fast, efficient way to transfer money and engage in machine-to-machine commerce and, uh, stablecoins as a cheap, fast, efficient payment rail for us to build x402. It's our

standard for internet-native payments. This says it's new. I guess it's still relatively new, but it came out in around last May. And since then, we've seen some very strong adoption. Honestly, much stronger than I ever thought we would. When it came out in May, our goal was for by the end of the year to have twenty-five thousand transactions made by agents. At the end of the year, we were actually seeing that every seventeen seconds. And these days we're seeing pretty incredible numbers across the board with tens of millions of volume and hundreds of millions of transactions occurring every thirty days. And if you'll notice from this list, you'll see that it's more than just

crypto companies. You see Adyen, Amex, uh, PayPal, Visa, MasterCard, Stripe, all these big names in payments who are engaging with x402 because we're building it as an open standard where anyone can use this how they want to create a way for machine-to-machine payments to happen. So I know there's been a lot of traction and comments around agentic commerce recently, especially with Muse and GrokBot and Stripe Link and this whole Amazon blocking agents com- controversy. And I think that's

all a very straightforward use case. And I could talk about agentic payments all day. But there's one that I wanna touch on that I think is particularly relevant to micro-transactions, which is something I'm very excited about with this whole agentic payments narrative. So if you all haven't noticed, the way that we've interacted across the internet has changed over the past couple of years. Now we're primarily doing it through chat experiences or Google's AI overview instead of going to the actual websites themselves. And while this is all fine and dandy for the consumer, it's resulted in a bit of a problem for these

publishers who have previously relied on ads revenue and human eyeballs to generate money for publishing their content. It feels a little bit like we're in the Limewire era of the internet, where you can scrape all this content for completely free, and there's nothing really to be done about it. And recently we saw that bot traffic has now eclipsed human traffic on the internet for the first time, and this number is only going to increase as we see more people using AI and these traditional web services and, and websites be disintermediated. So how do we solve this problem?

There's a very compelling opportunity here that's just starting to play out, and it's, it's really coming to life over the past couple of months for websites to gate their content behind a very small micro-transaction specifically for agents. So the idea here would be if I publish a Substack and I don't want an agent reading it for completely free, I can charge them one ten-thousandth of a cent. It's pretty trivial to the user. It should ultimately be fully abstracted away in the same way that your AI subscription handled this for you. But it's a meaningful revenue driver for these companies, and we're

gonna start seeing a shift from this attention-based ads economy online to one that's primarily oriented around micro-transactions and, and utility. And if you don't believe me, I, I encourage you to look at AWS and Cloudflare. Both of them, they're these massive content delivery networks. Together, they control somewhere in the neighborhood of forty percent of all internet traffic, and they've both recently enabled x402 micro, uh, transactions to gate any website that their publisher is creating. And though this hasn't seen very strong adoption

just yet, it feels inevitable to me that over the coming months and years, this will be the new way that we have companies differentiate and continue to un-disintermediate themselves in some sense. So payments. Makes sense. We're seeing some pretty strong traction here. All is going well. But what about this whole grow with money thing? What could this possibly mean? Well, it's more agentic trading and investing. And the first time I heard about this, I thought it was the most ridiculous thing in the world. Why would I ever give my ChatGPT access to my bank account? Or why would I

ever tell Grok to make me money and actually expect it to work? But after digging into it a little bit more, we found really three core use cases that exist today. I'm sure there'll be more going forward. And the, the core differentiation between this whole Grok make me money thing and what's actually happening is the human just providing a teensy bit of agency. And what I mean by that is the, the human still sort of directing the agent to automate these tasks that would otherwise be very difficult. So say, for example, I just wanna have my

low risk, you know, eighty percent S&P, five percent Bitcoin, you know, fifteen percent T-bill account fully managed and handled for me by an agent. I can just tell it to do that, give it authorization with all the appropriate guardrails and whatnot, and let it run, and I don't have to think about my money. Similarly, we've seen autonomous execution be a pretty compelling use case. I'm gonna go into an example about what this means later on, but you can imagine if you're interested in trading the, the Strait of Hormuz news and what's happening with oil, you could either be sitting in front of your keyboard at three AM watching these charts and watching the

news, or you could tell the agent, "Hey, this is what I think is gonna happen in these different scenarios. If this happens, just make that trade." And the third one that's been particularly compelling is the idea of taking a high-level natural language idea and translating that into actual financial primitives in a, in a strategy. Invest in space, invest in quantum, give me a low-risk portfolio, invest like Warren Buffett. All of these things for people who have these ideas, but they don't necessarily have the time to execute on them, nor do they exactly know how, is a very compelling

opportunity for agents to come into the foray. So I'll give you one coherent example of something I was actually doing a couple weekends ago. So you, uh, you all w- may remember that Dario put out this post, We Must Pace the Frontier. This was over the weekend. This was, uh, about a month ago. And I saw on X that Elon said Dario is right. I was like, "Holy shit, that's the first time that these two have agreed on anything. That's probably gonna be market-impacting news come Monday." So I went into my agent and I said, "I think that blank stock," you might be able to guess what it is, but I'm just not

gonna, not gonna go there. "I think that blank stock is gonna go down." And because we're now entering this era of twenty four/seven markets, I was able to put a short on these positions over the weekend just by talking into my phone. Could I have done this by sitting in front of a computer and then pressing all the buttons manually? Yeah, totally. But having it completely handled for me was so easy, it let me go about my day and not really think about it. And since then, since this worked fairly well, I now have my agent searching over the weekend for any news that might be market-breaking or blockbusting or whatever it might be, and it lets me know anytime that

happens such that I can make a trade-off that going forward. And is this specific opportunity going to exist forever? Likely not. But the whole idea here is that it's so simple and so straightforward for me to just talk my ideas to life that there's a very compelling use case for, for agents around this whole thing. This is another way to look at it. This is kind of how we think about building this AI-enabled financial account, and it really, you know, breaks down into advanced traders and also more like retail consumers and, and what exactly we wanna do

to provide them with the best-in-class experience across, uh, across the spectrum of capabilities that we can offer. So all that being said, we've, we've been building this stuff for about two years now, and we've had a lot of takeaways and lessons that have come into play and I think are pretty broadly applicable to building for agents and building for this new agentic era. The first is that trust and autonomy are earned one step at a time. The most common question we get when we tell people that we're working on agentic finance is, what if the AI goes rogue?

Why would I give money to an AI agent? How can I trust that this thing is going to work? And it feels very analogous to me to the way that we've seen self-driving cars start to proliferate. A couple of years ago, you'd be insane to try to use lane assist to parallel park. But now you go to San Francisco, you jump in a Waymo, and it feels completely natural. And we're seeing the same level of progression with the way that people view AI having access to money and finances. It starts with complete hesitation and, and no real thought whatsoever to the concept.

Then people come around to the idea that, well, I mean, my AI is already chugging through credits. That's kind of a form of money. What's the big deal if I gave it a couple of dollars to pay for some micro-transactions? From there, you'll notice that these models are quite conservative and, and aligned to the, the user's needs, and they don't really do anything crazy. Anecdotally, over the past couple of years, we've never seen any situation where a model goes and, and blows up a user's entire account in a way that's completely, uh, un-unfixable. So then we get to the point of a user feeling

comfortable with the right guardrails and rules and safety policies and procedures in place to give their agent access to an isolated portfolio that they can attempt to grow however way they see fit. One way is trading, but another is by offering a service and seeing if that agent can go make money by offering whatever it's doing and getting paid for that via these x402 rails that we're talking about. And you can see how from there it just continues to, to go on and on. The other thing that it has been particularly interesting to see is that especially in financial services, AI and agents are not the

product. It's actually the enabler. And this is a screenshot from one of our internal AI products called Coinbase Advisor. It's actually an SEC-registered AI, uh, AI financial advisor. It's the, the first in the world of its kind. And what we noticed is that AI and this concept of an AI giving financial advice, even though we've had a team of asset managers who are tuning this model, that's not the value prop. The value prop here is that we can scale this advice to hundreds of millions of users, provide it for completely free, and give them the ability to talk to a trusted regulated entity

twenty four/seven. People actually don't care, don't even necessarily know or want to know that there's AI going on behind the scenes. And though it may be obvious with the chat button and whatnot, I think this is going to lead into more generative UI experiences that feel pretty traditional, where you have agents behind the scenes and whatever sort of interface you want on the, on the front. Another thing we saw too is this whole AI chat experience. Though it's great for these generic use cases, it's very obvious that there's this blank-canvas cold-start problem that happens, where users come in, and they have

absolutely no idea what they can do with this stuff. So we started using these widgets as a way to show the user proactive insights and, and opportunities that exist and, and what they can start using these for. And we've seen really strong conversion through this surface specifically because of the general nascency and lack of education around agentic finance and what these opportunities are. And I suspect over time, as people understand it, and you see this a lot too with the, you know, Google Gemini and OpenAI ads, they're all trying to subconsciously drill into users the

variety of use cases that you can take with these products. Lastly, all these next-gen agents, particularly GrokBot and Muse, they all have a microVM. It's like a little personal computer that they can run things on, have storage, have access to a file system. That's been huge for our development process. We're now able to create CLIs and MCPs to let users connect their Coinbase account into these agents. Nothing special. What is special, though, is that we get requests all the time for things like, "I wanna do paper trading. I don't

wanna give my agent any money first. I just want it to trade entirely on its own," or, "I wanna backtest or do all these different things." And prior to these agents being a thing, we would have to go to our internal brokerage team, we'd have to build within this random repository, all of these changes and, and big tech things would have to persist and distill and percolate and, and just a lot of work. But now we can just write a skill to write the paper trading code onto the user's machine. And while this probably isn't the most, uh, the, the, the most efficient way to do it, it allows us to

execute and build and iterate at the pace that our customers are demanding without having to worry about all of the implementation details and all of the dependencies and traditional software cycles that might come into play. So with that, I wanna leave you with a couple thoughts about where I think this space is going and why agentic finance is going to be such an interesting opportunity. My core thesis here is that financial independence leads to greater autonomy for these agents. And you can imagine that in the same way when a human has access to a card on the internet, they can do so much more than they can when they're just

browsing and don't have access to pay for any things. I think the same exact thing is true for agents, and we're going to get to a point where we can see these agents interacting over long periods of time, and maybe they're just bringing in humans at the edges when they have a specific regulated task that needs to be take place, like forming a company or getting a legal opinion or any one of these tasks that is, is going to maintain its, its human-oriented factor for the foreseeable future. There's a really interesting, uh, thing that we've been thinking about as a thought experiment, which is what happens when an AI is

able to sustain itself? You give it access to money, you give it access to a wallet, and you allow it to start earning. Can it start to pay for its own inference? Can it start to pay for its own compute and its own upkeep? And does this start to feel a little bit like a, like a proto-digital life form of sorts, where this AI is now able to just exist and subsist based on the, the, the quality of the services that it's providing? And though I think this is a long way away, I think it's very possible to see how this concept comes back to life. And within the crypto space specifically, there's this concept of a DAO. It's a D-A-O, decentralized autonomous organization.

And it's been an idea for, for almost a decade now, but there's never been really any interesting implementation where there's automation at the center and humans at the edges. But I think now is the time where this is finally going to become a thing. You have an AI that follows a set of rules and has access to a treasury to manage at that core, and it can hire out humans to do other tasks at the edges, but it's really this self-sustaining, uh, organization that, that uses AI and agentic finance to make that all come to life. Another example being a self-driving car that owns and run itself. Give it access to money, let it take people on taxi rides, use that money to

pay for its operation and maintenance and upkeep and whatnot. It, it's a very interesting thing and, you know, who knows how long it'll be until that comes to life, but it feels inevitable at some point. And the, the whole idea here being that agents are just gonna be start transacting at the speed of machine-to-machine commerce. They are going to be both paying and earning and building, and ultimately, I think it's gonna be the primary way that we see money move across the internet and at velocities which we've never really seen before. So that's all I had. Really appreciate everyone's time today. If you're interested in, in collaborating or, or jamming, I

would highly recommend you play around with x402 and the Coinbase MCP and all of our fun products. I promise you they're much easier than you think to get started with them. And, uh, yeah. Thank you all. Thank you all for your time.