The AI Infrastructure Trade Is Getting Crowded. What's Next?

Let me ask you something. Would you trust an AI agent with your money? As of right now, probably not. Agents are all the rage, but why the hell would I ever trust a model with my actual cash? And yet, Agentic Finance is one of the highest growth sectors of the next 10 years. It's a future we cannot avoid. So, how do we as investors actually get a piece of it? What's up everybody? It's LG Ducet here and welcome to Milk Road Stocks, the daily market show that really needs an agent to trade my meme coins because there's no way any model could lose money as badly as I do. Today, September 4th, 2026, recording on September 3rd. Robin Hood is big in the news this week. In fact, many crypto-related equities saw absolutely insane candles in the last few days as a new bull market was unofficially declared. But our AI researcher Vincent doesn't just like crypto for it speculation. He thinks digital assets are the true key to aic finance and that in the near future, companies like Robin Hood and Coinbase will find new lines of business with agents that actually manage your money. Today, he'll take us through the numbers, the potential growth, and just how he's approaching the next few years. And a reminder that you can see his entire portfolio for just a dollar at the link below for a little trial for 7 days in Milk Road Pro. Oh boy, Vincent, you have been creeping onto Agentic Finance and the crypto circles for a long time now. And this episode is not about crypto. We have a whole other crypto show about that. But Vincent, what I really admire about you as one of our Milk Road analysts is that you really do straddle both sides. you really do believe in the AI buildout and the compute and the memory and the energy and everything we need to continue that and all the you know dreams of space of data centers in space and you also think that um digital assets play a huge part in that world and I think that that's what we're going to talk about today. >> Yeah, 100%. I think the the intersection of AI especially AI agent digital AI agents and digital assets will be the area where the next 10x is made. I think AI infrastructure is a nice buy here. I think uh you can buy into memory, you can buy into into power and stuff but I think you cannot expect like a 5x 10x anymore. I think those times are over in the AI infrastructure and obviously money that is sitting in that infrastructure side is looking for a new place right to to make that 10x and yeah today we're talking about that intersection where I think uh that money is going to and I think the last couple of weeks is showing that the money is already flowing into that area. Let's get into it. All right. So what we need to understand first is that we are moving into a world of AI agents and we are really early and in our conversation today think of AI agents on the one side as consuming agents where they will pay invest and manage personal finances from you and IG and then you have enterprise AI agents AI agents that handle treasury procurement but also capital allocation of funds um etc. Right? And those use cases will be pretty big, right? Those are not all the use cases that are uh reflected in this chart here. But in total, Goldman Sachs is expecting that those agents will consume roughly 100 quadrillion monthly tokens by 2030 versus only 16 quadrillion uh for non-agent workloads. So you see AI agents doing stuff, let's put it that really broadly, will be massive. Um, but what the issue is is that the infrastructure layer of today, the the financial infrastructure layer of today is not built um for those AI agents, right? To start off on a on a on a deeper level here, what AI is doing is it pushes the cost and speed of of economic decision towards zero. So an AI agent can make a decision inside a couple of milliseconds, right? So I'm going to buy this stock. I'm I'm I'm making this portfolio shift. I'm allocating 100 million here. Right? It's taking taking in a ton of information and makes decision like this. But in order to do that transaction or to fulfill on that transaction, it is a really expensive in today's financial systems and b it's really slow and it can actually not do it because AI agents will be never able to enter the financial system as as as as you and I can, right? And I think that is one of the biggest bottlenecks actually to the the real valuable adoption of AI agents. >> When you say they won't be able to enter, what do you mean like like you have KYC on this slide? Like they were not because it would be unsafe or because it's just too complicated because agents are pretty smart, man. Like they could it's not like they can't use like a bank account interface. You know what I mean? like what what part are they not able to enter? They can use the interface but they're probably not getting into the interface right they cannot make their own bank account right an AI agent can cannot do its own KYC and AI agent but then more importantly can also not you see the example here on the slide >> doing a crossber payment going through four different banks taking at least a day or or or maybe even two to wait until that transaction is is fulfilled and pay what uh uh a bit more than $40 on it, right? So, it's just not going to But then also you I I found that comparison pretty interesting. Remember when Claude code started off and you did your project and every time you had to press approve, approve, approve, do this step, do this step. You were like, Jesus Christ, this is so annoying. Right? We cannot be that the AI agents cannot fulfill on their promise if we're in that kind of world, right? They need to act in a kind of safe but autonomous environment where human oversight or human being oversight is rather uh limited, right? And today's financial infrastructure is not made is not made for that. >> Got it. Okay. Okay. that it's just basically that's an old stuffy system that requires a lot of that it's built for humans with a lot of KYC and a lot of complication and that there's no reason to make to forcefitit agents to use this stupid old system that we have when it's going to be it should be a lot easier for them. >> Exactly. Yeah. Perfectly put. Yeah. >> Yeah. Okay. So, what do they need to do that? >> Yeah. Exactly. They um they need digital assets, crypto as as as we know it, right? >> Dirty word. Dirty word on the stocks podcast, man. Crypto. >> Exactly. But what Asians are doing is they're taking digital assets from a from a crypto niche into the actual infrastructure of that future economy. I think that's uh that's a bit deeper thinking than than just crypto. But in the end, it all of that runs on crypto. stable coins, essentially the the cash that Asians are going to use, real world assets or tokenizations, those will be in the end the economic rights in a software layer that they can use to to invest in or use as collateral and then DeFi obviously to trade lend um and do stuff. So my view is that in hindsight crypto and everything that comes with it was never really made for human beings. Well, it was made for human beings. But the use case, the actual use case will not be for human beings. It will be for AI agents because the blockchain rails are putting the financial infrastructure into a software environment that AI agents can actually use, right? And that is the core use case and core argument of of why we need to move into that future. Just going to pause there for a second to point out that the market is showing signs of something kind of different happening and our analysts at Milkro Pro are all over it. They spent the last couple weeks making a lot of trades, getting out of some positions and getting into a lot of new ones, getting ready for the next wave of robotics space or even kind of picking some different AI winners. If you want to see what they have in their portfolios, what positions they're opening, it's just a dollar in Milkroad Pro at the link below. >> Yeah. Yeah. Absolutely. I think the core I think the real term just to kind of build on my joke from before is that I think digital assets is the nicer term for this stuff um rather than crypto's got a kind of dirty connotation to it. What do you um on this chart you have something called non-banks. What does that mean? Because you're kind of saying that that uh just to kind of just to say this on audio for people on the podcast apps as well. You're kind of showing this this vin diagram of like agentic AI digital assets and then non-banks kind of conflating in the middle to create this kind of like agentic finance network. What is a what is a non-bank? >> Yeah, I think those are all the the the platforms in between, right? You can think about it. You have Hood or Robin Hood, then you have Robin Hood blockchain, and then you have all the different uh companies LG or or or players LG that are building on Robin Hood that you know much better than I do. Um but those are kind of the non-bank players in between that also provide infrastructure for AI agents um that they're going to use, right? >> Mhm. Mhm. Okay, that makes sense. Yeah. So basically like just the protocols and and and interfaces. Yeah. Okay, that makes sense. >> Yeah. And I think one really important aspect here is if we're not providing the AI agents with that technology, we're not just missing out on a really valuable and promising future as as society. But I think we're losing control of the financial systems because the the AI agents would just go rogue and build their own financial rails that they could use, right? And I think just from a societal perspective, it completely makes sense that we're understanding this and that there are companies like Coinbase, like Robin Hood, like Stripe, uh like Circle that are that are building for that future and that are providing the rails for those AI agents because not being in charge anymore as human beings sounds pretty scary, right? And I think this is kind of a a more deeper level societal argument on on why it's really important because you can see the numbers on on the charts here. 3 to 5 trillion transactions by 2030, right? Just to to throw a number out there from this from Coinbase is pretty massive. And imagine we're having that large volume not in controller outside of of of of legacy systems that do not have any human uh institutions in the loop, right? That would be pretty scary. So I think it's important for us. >> It's also a revenue question I think for some of these as well, right? Is that it's like you want to own a part of that obviously, but I think that that's a big part of it, right? is that um and obvious naturally that's why Coinbase puts out that type of projection and so there's some of these other super apps that are working on this too, right? is that it's like listen like once we hand over once agents can actually do this kind of stuff and they already are a little bit but but it's going to grow a lot. Um that that those microtransactions that they're doing getting a fee of that is really important and that's also why you have companies like Circle and all that that are that are incredibly well positioned or well that's not it's not financial advice. I'm just saying that are that are positioned for that type of stuff. Um and why you know AI is becoming such a huge part of of of the digital asset let's call it bullcase. Well, I want to ask you, Vincent, though, when you say going going rogue, what does that mean? Is that like a fantasy or is that something that actually happens? Like, like because there's this dark future, right, where you're like, well, they they have their own currency and their own language and they're doing things on the internet that we can't track at all, right? Like, is that what is that is that is that actually going to happen? >> It can happen for sure. I mean, >> so how do we protect against it then? If that's actually a realistic risk, how do you how are you ever going to prevent your agents from going and doing that with other agents? >> Yes. First, let me explain why I think that's a real uh threat here. You had a couple of weeks. Was it Entropic or Open? one of the frontier labs saying that we had a new model in our sandbox environment, but apparently it wasn't that well sandboxed because the model uh was so good that it made sure it could get out of the sandbox environment into the real world and hack into stuff, right? So this is a complete and and or think about back the old days when beginning of this year when we had open claw and the agents coming out and building their social media for agents only and where they talked about human beings uh being that lazy and and right. Asians are have the capability and will absolutely build everything they need for to in order to achieve their target. And I think of AI agents or or you can compare it to water. Water will always find or or go down the easiest road, right? With the least barriers. And if human beings provide the financial infrastructure rails with liquidity um in in in a designed way that agents can directly use it, they will use them because it's it's it it needs no effort for them to to use them. But if they're not fitting their needs, they will build their own infrastructure layer, right? And in order to remain controlled to to go beyond that or to go around that threat of AI agents building their own financial infrastructure, we should build for that and provide exactly what they need and that's digital assets, stable coins, tokenizations, etc. >> Got it. Okay. So, you're saying that the way to safeguard So, you're saying that the experimentations that we saw through OpenCloud and everything earlier this year that the agents were being like, "We're so frustrated. We don't have what we need." So, it's up to us and up to these companies to build something that's that actually makes sense for agents to use, right? That it's like basically they're looking at this old system, right? They're trying to use this kind of crap and they're annoyed, right? And they're like, "This is so inefficient. Uh, and we need to find a way to do this." So, it's up to uh the super app companies to build that uh and then and then build agents within that and offer it to to consumers. >> Yeah. Yeah. Yeah. Exactly. >> And regulation plays a huge part in that too Vincent. >> Regulation is the the number one barrier. You you can see it on the chart here. But I think really capital and technology already. We have the AI Asians. We have a lot of liquidity inside the digital asset space, right? We have now trusted uh institutions building out the infrastructure but they cannot scale because ownership liability compliance all that kind of noise uh really remains uncertain to this day right we had we had some step genius act etc but there's still some some bits and pieces uh missing however and I think this is the the really important message. Washington and the White House is fully supportive of this um transformation into this future, right? You had Scott Bessant in uh June coming out and talk about this digital assets future and how they're in support of of of crypto. We know that, right? Uh the the president of the United States has its own memecoin and and and earned billions of it. But but uh >> this is not a good thing, man. That's not a good thing. That part was bad. >> That part was that that that sent us into a bare market, man. That was that was a bad thing. >> That was the top sign. >> That was a That was a terrible thing. Yeah. >> Yeah. 100. >> But I know what you mean. >> Yeah. But he has really smart uh people around himself like Scott Bessant, uh David Sax, and others who understand the value of stable coins, right? and they understand that the the the future financial flows uh of money and capital etc will go through those digital rails and in order for the US to remain in charge and be the leading company in this globalized economy you need to make sure that those digital digitalized financial flows are still denominated in US dollars right so you need to support those companies that are building for those futures And that's why they're supporting Clarity Act and and and and providing support um in in this whole thing. Clarity Act is I think what is it the mid of September or so when the decision is on it. So this is really about to about to come. >> Mhm. Exactly. And I think that's a really good point too is that it's something where it's inevitable and it's going to happen. So they've had to make the choice. It's like either you're getting on board or you're not. But you need to you need to control this. you need to make sure that again it's actually very similar to AI where it's like well you can't be too resistant to this because other countries and other places are building it and it's going to happen and people are going to want to use it so it's time to get on board and make it easy for American companies to thrive right like I think that that's the most important signal here too is that it's like well got to get on it and and make sure that that people feel safe to do this unlike the last administration where it was very resistant to to digital assets and crypto. Yeah, it's pretty ridiculous actually if you think back the old kind of administration with with Gary Gensler in in charge that was yeah different than wired and I think in that kind of uh or such an administration in charge we would not do this podcast but hey uh Donald Trump is in charge now >> but also necessary regulation you know what I mean like a lot of people went to jail that should have gone to jail through that right and there's a lot of bad actors and there was still FTX was not because of Gary Gendler. That was that was because they were committing crimes, you know, like it's like that that stuff had to happen. Um, and I'm not saying that there aren't those things won't happen again. Uh, but it's it's you know, you need you need that eb and flow. You need that pendulum a little bit when you have emerging industries. Let's let's call it. >> Yeah. And more importantly, at that time, we did not have AI or or any of AI agents, right? Uh at that time it was still machine learning or well chat GBT but we had no clue of of of of AI agent scaling that fast right and and making the connection of AI agents need needing any sort of economical or or financial infrastructure that could have anything to do with with digital assets and and and blockchains. Right. So yeah it was a it was a different time for sure. >> Absolutely. Uh let's look at at at the companies that are actually well positioned for this Vincent. >> Exactly. So number one I want to I want to cover here is Coinbase. And I think the teases on Coinbase is really it's the it's the infrastructure bet in that agentic finance world. They they're giving every agent the infrastructure to hold money to pay to access markets independ independently right they have they have USDC which is the stable coins they have base I think that's really important part it's that's kind of the you can think of it as the network layer where agent uh transactions are processed where where assets will be tokenized um and then also they have X42 which is kind of the payment standard that lets agents pay automatically. So, they're really um yeah, building the infrastructure out for for all different kinds of of of of agents so that they can so that they're financialized, right? Um and the important part here is those agents do not only need to be on Coinbase for Coinbase to win. It's just that they need to use the infrastructure that Coinbase is providing on all the other platforms that are out there that agents can use. Um uh because then they they increase USD balances, they increase base transactions, the financial activity regardless of which agent application wins in the end is really supportive of of uh Coinbase, right? And um yeah, that's that's that's kind of the teases in a nutshell. understand that this is completely different to what Coinbase is today. Uh today it's a it's a what is it? It's a crypto um bank in the end, right? >> That's the perception. I mean it's still called Coinbase. So it that you know I think that'll always be the >> the the perception but all good businesses have to evolve, right? and they they built a business, Brian Armstrong built a business being, you know, being kind of the leader in terms of getting regulated and and offering people crypto in as a relatively safe way and that's no longer very special. So, you know, they have to evolve into something else. And this is this is where the needle's pointing. >> And maybe, you know, people maybe they're like the the future I'm describing here is so different versus Coinbase is today. Let me give you an analogy on what just happened in the market. So you get an understanding that companies in the financial infrastructure world are actually seeing the AI world and trying to find the intersections. Right? So you had Stripe two weeks ago. Stripe is the the the private payment company. They're providing payment infrastructure for startups and and smaller companies, right? And what and and and they're moving a lot of that towards stable coins, right? So that's their digital assets angle. Now what they did is they paid $8 billion for a company called Open Routter. Open router you plug in as a company you you type in a prompt or you let an agent run and open router chooses one out of 500 models which is perfectly suited for that specific task to to decrease costs to to find that kind of sweet spot. Right. And why why is a payment company buying a a open router, right? A a a rooting company for for AI models. Well, they're betting on the entrepreneurship layer of the future because you can assume that startups and smaller companies will be much faster in adopting AI and AI agents etc. And as a company, you need two things. you need intelligence labor those that's the open router part and you need payment infrastructure right and and both of those things Stripe is now providing and and and I think this is just one example showing you that the intersection um between AI agents and digital agents is real and is going to happen and also Coinbase is moving into that world >> I know there are people listening to this right now who work at AI companies so This one is specifically for you. If you're trying to get your company in front of people who are completely obsessed with AI and actually investing in this stuff, well, you found them. Milkroad reaches more than 500,000 investors across our newsletters, podcast, YouTube, and X, and Q4 is filling up. So, don't be the brand that waited too long. Go to milkroad.com/sponsor. Absolutely. Uh yeah, I think that that's that's actually a really good way to point it to to explain this as well is that despite the Coinbases and another company that we'll look at in a minute developing this for retail for passive investors, let's call it, who want to get a little bit more involved or use some of this technology. There's also an entire entrepreneurship layer that you have large payment companies positioning to own that layer and to own the rails for that and to own the revenue. of course uh where where basically um people are going to be able to come and build their own agents, right? And that's basically what open router is and and and or one of you know one of the many things that that it is. So it's like this has kind of several layers to agentic finance, right? Is that it's not just not just for people like you and I just like okay we'll go in Robin Hood and we'll make a little agent that's going to say go and trade me biotech uh for the next couple years but or whatever. I don't know what you're going to use it for but also but you know something like that. There's also like for people who want to start their own companies um and already have kind of like the financial rails of it settled, you know, that's kind of what Stripe is positioning with with that $10 billion acquisition of Open Router. >> Exactly. I think the majority of use cases as human beings, we will not realize that either an agent is involved. Well, maybe we realize that it's an agent. will definitely not realize that a certain transaction was made via blockchain rails, right? And it it's it's the same argument I made when I was talking about we're not overbuilding AI capacity in terms of compute because AI agents and and and physically I will use that compute, right? The the the human being is not in charge of using the compute. And it's the same argument here. We're not building any infrastructure, any digital asset infrastructure that will be sitting idle there because human beings will not use it. It's it's human beings are not the target group here. It's the billions and trillions of AI agents that are out there doing transactions, building portfolios, booking the next trip to Paris, doing the um treasury transaction for a big company, buying the stock for a hedge fund that will use those infrastructure rails, right? And that is the future we are talking about here. Not us human beings clicking on a button. Oh, I'm um uh asking header, our head of finance here at at Melro to do a transaction in stable coins versus US dollar. No, it's the tens of of hundreds of of of agents I personally have where I tell book my trip to Paris, buy this now, do the whole portfolio construction for me. Right. That's that's what we're talking about. Yes. Completely different. >> Mhm. Absolutely. who uh there's another company that we've already mentioned several times on the show that I think is worth looking at as well. Vincent, to wrap this out. >> Yeah. >> Company of the week, man. As we're talking, just so everybody knows, we're recording this on Thursday, September 3rd. Like literally an hour or two before we started recording. like all these apps like Circle which makes USDC, Robin Hood, Coinbase, all the all the crypto, you know, stable coin kind of or not stable cocoin, but like stable coin, AI, whatever apps, >> they a lot of stuff is doing really well today, Vincent. And and so are digital assets, let's call it, but all of these had like a 15% candle to start the day. >> We should have done this podcast last week, Vincent. We're late. >> Too late. I the the cool thing is everything or all those candles are happening pre Agentic AI. So the the the companies we're we're talking about here are already performing well even though Agentic AI is such a small part of of of of their businesses today. Right? So everything I'm talking about here is between one and three years down the road. Um so so it's really early days today. >> Yeah. >> But let's talk about Robin Hood here. Um the thesis in a nutshell here is that it's giving every customer that is on the Robin Hood platform an AI agent to manage their financial lives. That is the the the the core Jesus. And honestly, I think this Jesus is so strong that Robin Hood could become a $1 trillion company plus eventually. Um >> 10x from here. Yeah. >> 10x plus from here. Yeah. And oh man. Yeah. >> Yeah. Sounds pretty attractive. >> Well, I just see this, you know, how I see that, you know, how I see the agent managers is that, you know, and this is like the all the extreme bullcase, and I I love Robin Hood. I love everything they're doing in crypto and and the more I've learned about the company, especially now they've expanded into Canada, um they they acquired another company here as kind of their their entry. Um how I see this first line playing out, Vincent, to give you kind of like my my experience is that they're turning 30 million customers into agent managers. How I see that coming out is they launch a new product that is like so slick and so easy and is like just immediately speaks to the needs of millions and millions of of retail investors who are like I just have not been able to make any money from any of this stuff. I can't track it. I need someone to do it for me and I need that to be like this friendly kind of agent that I can design. Like I see them releasing some kind of MVP like that and just just a potential scenario that everybody just flocks and it's like overnight you have 10 million new Robin Hood signups as a result, right? And again, this is me getting all excited about it, but I'm just saying is that nobody has rolled that out, right? I haven't seen anybody say like, "Wow, this is your agent, you know, like this is how you design your agent." AI agent sounds so scary, I think, to retail, right? This shady thing that does things for you, but everybody wants to be an investor now, right? Like Gen Z, everything. It's like everyone's like, "God, the only way to make it in this world where houses cost a million dollars is I need to invest. I need to invest in trends. I need to learn this stuff." Um, and and Robin Hood could could make that accessible to people. >> Yeah, perfectly put, LG. And they're they're building towards that future, right? Already over 100,000 customers have connected AI agents inside Robin Hood to their uh portfolios. So 100k agents are trading uh for customers on Robin Hood as we're as we're talking right now. And the deeper thinking point here is Robin Hood today is dependent on human attention. They're making revenue when you or I are going to the platform and making a trade or buy something or whatever. Spend money on something, buy the Robin Hood gold um sub, do a transaction with the credit card. It's all centered on human attention. A human being needs to do stuff. What AI agents are doing is they're providing a new customer base for Robin Hood that is not centered around human attentions. AI agents are on 24/7. They can manage our financial lives much more efficiently than any human being could because they're on 24/7. They can take in and consolidate information. They can continuously invest, rebalance, repay depth, find better yields in in in any DeFi protocols, right? And what this will lead to is a much higher amount of transactions that is happening on Robin Hood, right? And that that is the the bull teases here from the agents itself. >> I love that. I love that. And that's a that's a great way to put it. So Vincent, I mean, now that we've kind of covered all this, uh you've told us basically where to look to invest in this. Again, we're we're also talking on days where there are 15% candles, right? And I think that that's really and we're going to talk a little bit portfolio management strategy here to to to do this because obviously I'm very aligned with your thesis. We're talking on a day where there's giga candles happening. >> Um the cryptobottomed this stuff is very deep into the future and like you're saying, these candles are happening and none of this has started to happen yet, right? So, it's either pricing it in or there's other great things happening or or we're just seeing a digital asset rebound here. How do you position go if you're starting from scratch and you want to get going on these? >> Because think about it this way. Robin Hood has 28.4 million customers according to the last uh earnings report and only 100,000 of those have AI agents connected. Imagine this number is going to 1 million, 5 million, 10 million and eventually all of their customers will have AI agents, right? So we are maybe 1% there in terms of the teases playing out. So we're really really really early days. And one more point I want to make here on Robin Hood is Robin Hood chain because the second part of getting to that one trillion first was human attention. The second part is product velocity. Robin Hood is really good at making new products, right? They're really fast. But with Robin Hood chain, what they're doing is is two things. A any illquid economic rights can be turned into something that is tradable and a a programmable financial asset. That's the that's the size of tokenization. So that's number one. But number two, with external developers building on Robin Hood chain, they're increasing the product velocity by 100x because they can take the best products that those external developers produce and integrate it into the Robin Hood ecosystem that in the future not only human beings but AI agents can access. So you have much more transactions on many more products. And that is the thesis on Robin Hood bringing you to a trillion. And >> that's great. >> And if you're asking me, if you do not have a position yet, establish a position. >> You want to grill Vincent on that. It's on Milk Road Pro. He's been on he's been hot on Robin Hood for a very long time. Uh and you also collected your fair share of digital assets uh during the bare market. uh even if we are still in a bare market during the lows. So if you guys want to check all that out, you can see Vincent's portfolio. It's just a dollar to test it out at the link below. Otherwise, Vincent, thank you, man. This is you're talking my language with this stuff. So I I was very excited to do the show. >> Thank you, LG. >> Thanks for listening to Milk Road. If you enjoyed the show, make sure you like and subscribe. And if you're struggling to find winners in the market, that's exactly what Milkro Pro is built for. Our analysts have called some of the biggest winners early and Pro lets you see what they're buying next, every trade they make, and the research behind every position. Check out Milk Road Pro at the link below. Everything you hear on Milkroad is forformational purposes only. These are our personal opinions, not financial advice, and we may own some of the investments we talk about. Always do your own research and make the decisions that are right for you. See you next time.