Devin Ryan

Expect to see others follow here. It will be fascinating to track how much agents transact relative to human directed. 10x, 100x, more? Seems like we are on the precipice of a material multiplier to trading volume while driving better investor outcomes if done well. A win for brokerages + win for retail investors.

Prediction markets are about to deliver another record month with > $20B in trading volume, up double-digit % from February and >10x YoY. However, that’s a fraction of more developed asset classes like U.S. equities (>$10T / month) and even crypto (>$5T / month). Institutions aren’t gearing up to enter and exchanges aren’t experiencing substantial valuation increases because people are just excited about the “Sports” market IMO. I think sports are just the gateway, providing enough revenue to bring in real capital and infrastructure for a bigger vision. I think we have a different take on this than some…I expect prediction markets / event contacts to become an “asset class” that can add a lot of value - and we are only maybe in the bottom of the first inning. $HOOD $COIN @Kalshi @Polymarket Good to discuss with @joshuahlipton at Yahoo! Finance. Full interview can be watched here. youtube.com/live/N7V26B9i0…

@Nish1Nish Appreciate it!

@devinpryan Finally got around to reading this and certainly worth the wait. Thanks Devin.

Spending a lot of time reading about and trying to think about what the future looks like as AI develops and is more deeply integrated into the economy. Covering financials & fintech for over two decades, most big evolution points have been intuitive, and if you were wrong on the order magnitude or speed of impact, you could readjust your position as they played out over years if not a decade+. Today feels different given the rate and magnitude of change which is exciting but also a bit unsettling (seeing this play out in markets trying to assess winners & losers). To be clear, I expect certain areas of financial services to evolve slower than probably the consensus view, these are areas that are non-commoditized and heavily human - human with strong inertia. Like financial advice to retail investors (wealth management) or to corporations (M&A advisory). These businesses can even benefit in the near-term IMO despite some perceptions of the opposite. That said, I’m envisioning a world where a lot changes across financial services and not too far into the future. Trying to think outside the box but the marriage of scaling AI + blockchain should only get more attention given the technologies go hand in hand. Here's how I’m currently thinking about it, open to feedback/pushback. First concept to establish. Every transformational shift in economic history was about breaking some constraint everyone assumed was permanent. For example, the printing press broke the constraint on who controlled knowledge. Literacy went from under 10% to over 80% globally. The Reformation, the Scientific Revolution, modern democracy, etc - none of it happens without it. Electricity broke the constraint on what civilization could run on and moved the economy to 24/7. Medicine, food, shelter, communication, manufacturing - every system that powers modern life depends on it. GDP per capita roughly doubled in 50 years and that probably understates what actually changed.​​​​​​​​​​​​​​​​ The internet broke the constraint on access. To information, to markets, to each other - instantly, globally, for almost nothing. $30 trillion in new economic value in 25 years. Most of the companies that created it weren’t conceivable before it existed. In every case the world reorganized entirely around the new capability, and then, innovation quickly ensued that was not even conceived previously. There was a multiplier. I think the constraint being broken right now with AI is human attention (and its derivative…a multiplier on innovation like nothing ever experienced). There are 8 billion people on earth. At any moment maybe 2 billion are working? Of those, a small fraction are doing things that genuinely require human judgment. Everything else is bottlenecked by how many people are awake, focused, and available. Agents remove that bottleneck entirely. And theoretically, there's no ceiling if we build the infrastructure (and regulation allows this world to accelerate, where I already see some roadblocks in financial services like around fiduciary responsibilities, data privacy, but those should get figured out). There’s no reason this stops at 8 billion agents. Why not 80 billion. Why not 800 billion. Each one running 24/7, improving, building on what the others learned, no off switch, no weekends. For the first time in history, productive capacity doesn't have to be tied to the number of humans alive. I think this concept is difficult for most to digest as it’s not natural, but when that sinks in, imagining what can be possible becomes almost limitless (listening to Elon talk about a future of "sustainable abundance" driven by AI and humanoid robotics takes this concept to another level). The part I think is underappreciated is this evolution isn't just about automating existing tasks. When you take the governor off, the amount of innovation itself expands…potentially exponentially, and that’s what I’m trying to map out. 1/3

The market is dealing with a lot right now in the near term but IMO probably the single biggest story when we look back in 18 months is the speed of agentic evolution. When 1B+ agents are trading, making payments, micro-lending & borrowing 24/7/365 - much of it on blockchain rails - financial transaction volume will stop growing linearly, it should scale exponentially. There’s an arms race underway to build & own the infrastructure layer. Understanding the winners and losers here will be critically important.

Tough right? I like availability. Previously had “capacity” - which probably implicitly encompasses availability, aptitude, and bandwidth simultaneously. “Human capacity” is the governor as it captures time constraints, skill constraints, and the inability to run parallel experiments. But multiplier is really the framing. AI doesn’t just remove a constraint, it multiplies what was previously fixed.

Perhaps availability. Humans are available a small amount of time. Set work hours, vacation time, sick time, and finally time we are "working" but zoned out. Agents (and robots - though they need to be charged) have nearly continuous uptime. Even then, the term feels constrained. Amplitude? We are only capable of so much, but AI is a force, time, and knowledge multiplier. This is probably my favorite. Aptitude? Also a good one, since you have creative people already building using AI when they were limited on coding skills or time. Those constraints have now been removed.

Depends how far out in the future you look. I’m pretty confident that the rate of change over the next decade will be dramatically more than the past decade - which also creates substantial variability based on the path it takes. I believe that the global standard of living should go up and economic growth will accelerate. Elon has laid out a vision of “amazing abundance” worth listening to - to me sounds too good to be true. Some benefits will accrue to all (raising living standards) but through prior paradigm shifts, an increasing portion of the financial benefits accrue to a smaller group. I’ll follow up with a deeper thread on what this could look like tangibly.

@devinpryan Legitimate question - What is the end game?

I love that you went to the word attention. I probably changed it three times. Open to ideas:. Was trying to convey that every one of those examples created a multiplier on what was previously possible but were multiplying human capacity/output. This is very different because we’re adding almost a limitless number of capable producers to the economy operating outside of human attention and limitations.

@devinpryan Great piece, though "attention" feels like too limited of a word. That said, I love the idea. And with robotics coming in a massive wave behind it, it's definitely the right idea.

@CaseyBrink2004 No doubt. Rate of change here is different than prior shifts, companies will separate over the next few years based on the decisions made today.

@devinpryan Definitely seeing a moment in history where the risk in financials of little to no action is far greater than the risk of adoption

3/3 Companies are largely being valued against the transaction volumes of today, not what happens when the agent economy is 10x, 100x, or 1000x the size of the human one. Companies will need to keep up and adapt, but I still expect the ones that see this coming will position to participate as they have over history. Appreciate there are probably still more questions than answers here. That said, even as much remains uncertain, the direction of travel is becoming more clear.

2/3 Think about what a governor actually does…it doesn't just slow the engine down, it prevents the engine from ever discovering what it's capable of. Right now human attention is the governor on the rate of new ideas. Every experiment that gets run, every hypothesis that gets tested, every drug candidate that gets screened requires a human to design it, resource it, and wait for results before designing the next one. When DeepMind's AlphaFold solved protein folding, a problem that had stumped biology for 50 years, in months, it wasn’t because it was smarter than the scientists but rather because it could run experiments at a scale no human team could attempt. Point is, we are hearing about these “step-function” breakthroughs at an increasing cadence and this is probably still the first inning. Now extend that to every field. Materials science. Climate modeling. Drug discovery. Financial market structure. The number of ideas that get attempted, not just the speed of executing known ideas, expands by orders of magnitude. I suspect most of the breakthroughs we're going to see in the next 20 years aren't on anyone's roadmap today because the experiments that would reveal them haven't been run yet. Agents will run them. Dario Amodei at Anthropic has been talking about AI compressing 50-100 years of scientific progress into a decade…this is what removing the governor looks like at scale. The part I'm most focused on professionally is the connection to financial services and fintech. Here's the problem agents have under the current construct. They can reason, plan, and execute but the moment they need to move money they hit a wall. Traditional financial rails like banks, SWIFT, and ACH were built around human authorization. Every move requires a human in the loop. Agents can't use them at their speed, across borders, at the transaction sizes that make agentic commerce work. My thesis has been that blockchain solves this, and I think we are finally seeing some tangible evidence of that ( $CRCL provided great perspective on their earnings call around actual agentic Stablecoin payments already occurring). An agent can hold USDC, execute a payment to another agent or human, confirm settlement, and move to the next task - autonomously, globally, at any hour, for fractions of a cent. $COIN launched AgentKit specifically for this use case, Stripe also shipped an agent payment toolkit. So, why does this matter, what could this mean for transaction volume? It’s very difficult to model out cleanly given the step function change coupled with a number of complex variables, but it’s clear it won’t be a linear extrapolation. A passive retail investor trades maybe twice a month. An agent managing the same account could trade 20 times a day, maybe more. Same customer, same capital. 300x the transaction volume. Multiply that across $HOOD 27 million accounts and the math changes fast. And that's just retail brokerage. B2B payments, cross-border settlement, consumer lending, corporate treasury, software procurement, digital advertising. Virtually every category where human initiation creates friction becomes a category where agents generate volume that didn't previously exist. The velocity of capital accelerates tremendously, which should be good for economic expansion, the question is who the benefit accrues to. Now that is still probably overly simplistic as an agenetic financial system will likely evolve differently. Will agents interact directly, self custody assets in digital wallets, operate outside of how we think about the current financial system? I think these questions come down to both infrastructure evolution and regulation, but expect a lot of change when zooming out over a decade or so.

Appreciate that Nick. Was the incentive on the IRA what got your attention or were you looking to move any way? I think your other point is a good one. I call that inertia - people trust their existing financial institution and it’s too much effort to move the money/account - so they often need a push (something changes that trust-level…saw a lot of this post-GFC, a bad experience, or a financial incentive that creates a catalyst). For frequent users, a superior UI/UX can be enough as that “value” can exceed better pricing. If you look at the incentives at firms like $HOOD, they are targeting them to attract the type of customer they want (e.g., low margin rate attracts active traders, so use that as the tip of the spear).

@devinpryan @JuanRodrig07 it was super easy to trasnfer my roth from etrade to Robinhood. problem is old people are such sticky customers, and money isn't an issue anymore for them. so why risk change

Curious how people think about transfer bonuses? Was that the primary catalyst? Seeing some pretty attractive offers in the market. My experience is people rarely wake up looking for a new bank account or broker unless something goes wrong - inertia is powerful in financial services. But a strong incentive can get attention and create a reason to move. If the economics work, it can also be a very efficient acquisition strategy. Win/win. I’m seeing this work out particularly well with neobanks and digital brokers, and I suspect the next step is much more personalized incentives, with AI optimizing personalized offers based on specific customer behavior and expected lifetime value. Will be interesting to watch how firms experiment.

Thanks Juan. Agree, it has to be enough to move the needle to compensate for the effort, which I think these firms are targeting. There are other types of offers for lower balances/first time accounts. It’s only becoming easier to transfer in and out with a couple clicks, which I think is another contributing factor to the amount of money in motion. Net new assets/deposit growth is probably the most important KPI that people think about for organic growth - firms are starting to separate on their strategies…expect a lot of innovation here.

I think it works when you want to consolidate everything in one platform. I transferred my crypto from Coinbase to Robinhood about a year ago when they were offering a 2% transfer bonus. Other than that, I don’t see a reason to do it unless you hold a significant amount of money and the transfer bonus is therefore a good amount of $.

@jonathan_mg27 @blakerking93 @brian_wright21 Jonathan, thanks for joining us again. Great discussion with Tony and Blake highlighting the magnitude of the opportunity forming and Galaxy’s ability to execute and thrive through complexity.

Great to kick off March with one of the busiest conference weeks of the year! Tony Paquette (CFO) and @blakerking93 (Head of Power) spoke at the Morgan Stanley TMT and Citizens Technology Conferences in San Francisco, and @brian_wright21 (Co-Head of Data Centers) spoke at the Jefferies Power Conference in NYC. We also participated in the Morgan Stanley Energy Conference (NYC) and Daiwa’s Investment Conference (Tokyo). Across the week, we hosted 40+ 1x1 and small group meetings with a broad mix of long-onlys, hedge funds, family offices, and both equity and debt investors. Key themes: Data Centers • ERCOT’s Large Load Interconnection Batch Study Workshop and implications for $GLXY and the broader market • Our appetite for behind-the-meter solutions • Progress on tenant discussions for our newly approved 830 MW Digital Assets • Crypto macro, recent price volatility and go-forward outlook • Clarity Act approval probability and the impact to $GLXY and the industry, broadly • How we're leveraging our blockchain infrastructure and technology stack to help bring more institutions onchain Next week, we’ll be at Cantor Fitzgerald’s Global Technology & Industrial Growth Conference (NYC), Canaccord’s Digital Assets Symposium (virtual), Wolfe's FinTech Forum (NYC) and Compass Point’s Powering AI Infrastructure Roundtable (Washington D.C.).

@Kross_Roads @stevenfiorillo @amitisinvesting Love it!

Heading back after the Robinhood event and hanging with a legend and great friend @amitisinvesting.

It’s a good point. This space is fiercely competitive, with many very strong offerings. Robinhood is growing by delivering a good experience + strong economic value - and then listening to customers around where they can do better or have holes to fill in and quickly work on delivering that. The way they are set up with a “GM model” - where specific GMs own each product - helps break down some of that bureaucracy, getting resources allocated efficiently and products to market quickly.

@devinpryan They recognize that their competition is complacent and will phase out as older generations pass Other flaunt new account openings by young people but don’t mention that they’re mostly inherited iras and converted custodial accounts

Lots in there Amit. Probably received the least direct attention but the “Families” account aggregation feature adds another deposit engine. With a household’s full financial picture (spouse, outside accounts, even potentially parents), coupled with the strong UX, incentives will likely be increasingly personalized to move assets - also another step to position for the “great wealth transfer.”