CreatorAnthony Pompliano
GuestJordy Visser (22V Research)
TitleWhy Is Bitcoin CRASHING?!
Sourceyoutu.be/5TTtJBwmmnM
Date2026-06-07

Summary

Jordy Visser joins Anthony Pompliano to break down why Bitcoin has dropped 50% from its all-time high, and why he is not selling. Visser argues the decline is a classic rotation event rather than a structural collapse -- hardware and infrastructure names have dominated for years, and capital is beginning to shift toward application-layer software, including human-body software via pharma and financial-guardrail software via crypto. He frames Bitcoin as the S&P 500 of crypto, the one asset that survives all innovation cycles, and sees AI agents settling transactions as a long-term structural tailwind. The conversation then expands into Eli Lilly and GLP-1 peptides as the clearest near-term AI application play, tracing second-order effects from addiction treatment to entitlement reform. Throughout, Visser stresses that the pace of AI-driven change has compressed market cycles so aggressively that mental flexibility is now the most valuable investor skill.

Key Points

Quotable Moments

Quotable moments are auto-generated from the transcript. Speaker attribution and quote accuracy should be verified against the original source before republishing or sharing.

Jordy Visser -- on AI agents and Bitcoin
"There are more AI agents on HTML sites than humans. This is where we're headed. So if you think the world of human beings is going to dominate commerce, it's not. It's AI agents. So if you ask me what is a direct play on that, I believe Bitcoin is a direct play on agents dominating humans going forward."
Why it works: Connects two macro trends -- AI proliferation and Bitcoin's neutrality as a settlement asset -- into a single clean thesis that bypasses the usual store-of-value framing entirely.
Jordy Visser -- on the four-year cycle
"Maybe these aren't four-year cycles. Maybe these are now nine-month cycles and we've compressed the time. And that's what a bubble is. It's price versus time. AI is speeding up time. The exponential age is speeding up time at a pace we've never seen before."
Why it works: Reframes the standard crypto narrative in terms of time compression rather than halving cycles, which is a more defensible and forward-looking model once you accept that AI is accelerating every domain.
Jordy Visser -- on long-term Bitcoin conviction
"I believe we are seeing signs of a top in hardware. Not that they're going to go down and not that this is a crash, but you look for rotations to where the market narrative shifts. And I think the market narrative is shifting from the chips, infrastructure, energy side of the cake and it's moving up to the application side."
Why it works: Gives investors a clear signal to watch and a framework for timing a portfolio shift, without making a price prediction that could be falsified in 30 days.
Anthony Pompliano -- on crypto as investor training ground
"Bitcoin has gone through that every 18 months for a decade. So you just see this stuff over and over. Crypto over the last 15 years may have been the best training ground to be an investor in any asset class because what you essentially did is you compressed all these market cycles into a very short period of time."
Why it works: Converts the argument against crypto -- extreme volatility -- into a feature that produces battle-hardened investors before they deploy serious capital in other asset classes.

Concepts and Ideas

Core Framework -- Bitcoin and the Rotation

Bitcoin as the S&P 500 of Crypto

Every stock and token is an idea, and ideas get replaced. Bitcoin is not an idea -- it is a settlement layer and a store of value that sits above the churn of innovation. Visser argues this makes it the only crypto asset likely to survive 50-plus years intact, the same way the S&P 500 index survives the failure of individual companies within it.

The 200-Day and 200-Week Moving Averages as Conviction Anchors

Visser uses moving averages not for trading signals but for conviction calibration. The 200-week moving average is roughly equivalent to the four-year cycle in time, so if that framework has any validity it shows up in the price relative to the long moving average. He will not deploy aggressively until Bitcoin recaptures and holds the 200-day.

Volatility Compression as a Sign of Maturation

The current 50% drawdown is happening at historically low Bitcoin volatility. Meanwhile individual stocks are moving 8-10% per day. This inversion -- Bitcoin becoming the calmer asset relative to equities -- is Visser's evidence that Bitcoin is crossing the threshold from speculative instrument to institutional-grade asset class.

The Hardware-to-Application Rotation

Jensen Huang's five-layer AI cake (chips, energy, infrastructure, models, applications) has only monetized the bottom three layers so far. Visser believes capital and market narrative are now rotating upward toward the application layer. That rotation is where Eli Lilly, crypto-native financial tools, and other application-layer businesses sit.

AI Economics -- Tokens, Models, and the Edge

Token Budget Blowouts and Efficiency Optimization

Enterprise users across Amazon, JP Morgan, Uber, and others blew through their annual token budgets in a single quarter. This is driving a wave of optimization -- fine-tuning open-source models, switching to cheaper providers, reducing per-query token consumption. Paradoxically, Anthropic's revenue is still growing rapidly because aggregate demand is outpacing per-customer efficiency gains.

Generalist Models Are Commoditizing

The performance gap between Claude and GPT-5.5 has narrowed to the point where users choose based on stylistic fit rather than capability. Visser prefers GPT-5.5 because it anticipates his thinking better -- not because it scores higher on benchmarks. This signals that differentiation will shift entirely to specialized, data-rich workflows rather than raw model capability.

The Edge Threat to Centralized Model Providers

As the cost of running capable models drops (DeepSeek, local inference, fine-tuned open-source), the case for paying token fees to centralized providers weakens. Visser sees this as the fundamental bear case for all model companies -- a race to the edge where compute moves closer to the data and the user, bypassing the cloud API layer entirely.

Human Software -- Peptides, GLP-1s, and Eli Lilly

Peptides as API Keys for the Human Body

Just as an API key unlocks a software service through a receptor interface, peptides bind to biological receptors to trigger cascading effects in metabolism, appetite, addiction pathways, and inflammation. GLP-1s are the first widely adopted example, but the framework applies to the entire class of receptor-targeted compounds.

Side Effects as Research Signals, Not Liabilities

Traditional drug development treated side effects as problems to be managed. Eli Lilly is now treating the unexpected side effects of GLP-1s -- reduced addiction, improved cardiovascular markers, apparent cognitive benefits -- as discovery signals pointing toward new receptor-level mechanisms. AI accelerates this process by pattern-matching across 150 years of trial data.

Eli Lilly as a Specialized AI Company in Disguise

Eli Lilly has a thousand Blackwell GPUs, an Nvidia co-innovation lab, Toune Lab (its internal Google X for AI-driven research), a partnership with Isomorphic Labs on protein folding, and is buying related biotech companies weekly. It is building the AI infrastructure to process its own proprietary 150-year dataset -- a data moat no generalist model provider can replicate.

The Intertemporal Transfer of Wealth

The hardest problem in finance is giving yourself or your children money 20 to 50 years from now. Real estate works until it doesn't (Detroit, post-COVID office buildings). Stocks require constant company-level conviction. Visser's argument is that Bitcoin -- as the non-idea settlement layer -- is the most defensible current answer to this problem over multi-decade time horizons.

Mental Models for Navigating Exponential Change

The Mobile App Churn Model for Bitcoin Adoption

Each bull run brings in new users; a portion churn during the drawdown. But each cycle leaves a larger residual base of committed holders who set a new price floor. Pompliano applies this same pattern to AI companies now -- big moves up, partial drawdowns, but the floor keeps rising because genuine converts don't leave.

Mental Flexibility as the Primary Investor Skill

Visser argues that the most important thing anyone can do right now is refuse to take a static snapshot of the market. Conditions that were true three weeks ago are already obsolete. Holding hardened views -- Anthropic is winning, Claude is better, open source can't compete -- will be wrong before the quarter ends.

Implementation

Implementation steps are auto-generated from the transcript content and are provided for informational purposes only. They do not constitute professional advice of any kind. Always consult a qualified professional before acting on any information presented here.

1

Establish Your Bitcoin Conviction Range

Visser's rule is simple: less than 1% Bitcoin allocation is intellectually dishonest given the asymmetric upside; more than a concentrated bet is reckless given binary tail risk. Decide where your allocation sits within that range and commit to it before the next price move, not during. Write it down with the reasoning so you can test whether your conviction holds during a further 20-30% drawdown.

2

Use the 200-Day MA as Your Re-Entry Signal

Rather than trying to call a bottom, watch for Bitcoin to recapture and hold the 200-day moving average. This is the technical signal Visser says he waits for before deploying new capital aggressively. Set a price alert now so you are not reacting emotionally when it happens; you want to be positioned before the signal confirms, not chasing after it.

3

Map Your Portfolio Against the Five-Layer AI Cake

Write out every AI-exposed position and label it: chips, energy, infrastructure, models, or applications. If most of your AI exposure sits in the bottom three layers and the rotation thesis is correct, you are overweight the part of the cake that has already run. Visser is looking for application-layer names with proprietary data moats -- healthcare, financial guardrails, and crypto tokenization infrastructure.

4

Study Eli Lilly's Strategic Acquisition Pattern

Visser says Eli Lilly is buying companies almost every week, and they are all connected to the receptor-level mechanisms that GLP-1s have revealed. Go through the last 12 months of Eli Lilly press releases and map the acquisitions to the underlying biology they are targeting. This is the equivalent of watching what the Mag 7 was buying in 2015-2018 -- the acquisition pattern tells you where the company thinks the TAM is expanding before the market prices it in.

5

Watch the David Ricks and Jensen Huang Interview

Visser recommends this interview specifically for understanding why GLP-1s matter beyond weight loss and where Eli Lilly is taking its AI-driven drug discovery platform. This is sourced, primary material from the CEO -- not analyst commentary. Treat it as a required primary document if you are considering any position in healthcare or pharmaceutical AI plays.

6

Audit Your AI Tool Stack for Specialization Gaps

Visser and Pompliano both note that generalist models are commoditizing while specialized, data-rich workflows are pulling ahead. Audit your current AI tool use: which tasks are you running through a generalist model that could be handled better by a fine-tuned or domain-specific system? This applies whether you are an operator, an investor evaluating AI companies, or building a product on top of LLM infrastructure.

7

Build a Long-Horizon Durability Layer in Your Portfolio

Visser frames Bitcoin through the Patek Philippe lens: you are not holding it for yourself, you are holding it for the next generation. Decide explicitly what percentage of your portfolio is meant to operate on a 20-50 year horizon rather than a 1-5 year horizon. Assets in that layer should be evaluated entirely differently -- primarily on durability and independence from any single company, government, or innovation cycle.

8

Track Second-Order Effects of GLP-1 Adoption

Victoria's Secret reporting 15% sales growth attributed directly to GLP-1 weight loss is one data point. Map out the sectors that benefit from a population that is measurably leaner and healthier: sportswear, active travel, smaller-format restaurants, fitness equipment, addiction treatment retrofitting, and Medicare/Medicaid expenditure trajectories. The second-order effects are where the non-obvious investment opportunities typically sit before the mainstream catches up.

Full Transcript

This transcript was auto-generated and may contain errors in speaker attribution, transcription accuracy, or formatting. Long transcripts may be truncated due to processing limits. Confirm accuracy and completeness against the original source before referencing or republishing.

[00:00]

Jordy Visser: There are more AI agents on HTML sites than humans. This is where we're headed. So if you think the world of human beings is going to dominate commerce, it's not. It's AI agents. So if you ask me what is a direct play on that, I believe Bitcoin is a direct play on agents dominating humans going forward.

Anthony Pompliano: What's going on guys? Today we got a great conversation with Jordy Visser. In this conversation, we talk about what's going on with Bitcoin. Why is it selling off? Is Jordy worried? And is he going to sell his Bitcoin? On top of that, we talk about what's going on with Eli Lilly, why he is so convicted that it could be the best AI name for the AI trade. And then we get into a bunch of other second and third order effects of what's going on with the infrastructure buildout, artificial intelligence, so much tokens being consumed, and then how does Bitcoin fit into all this? All that and much more in this conversation with Jordy Visser.

Anthony Pompliano: All right, Jordy, Bitcoin is down 50% from its all-time high. People are freaking out. Michael Saylor sold 32 Bitcoin this week, and people think that maybe he doesn't believe anymore. What's your take as to why Bitcoin has fallen so much and is it over or will Bitcoin recover?

Jordy Visser: So I learned my lesson last year not to predict levels on the upside. I'm also not going to predict levels on the downside. Last week in my weekend video, I highlighted the difference between a bull market and a bear market. And that's what charts are good for. We're still in a bear market until that changes and until we start seeing some moving averages start to either point higher -- but more importantly, we failed at the 200-day moving average a few weeks ago. We started to go down again all while the stock market was going up nine weeks in a row. So the good news is Bitcoin's not correlated to the stock market anymore. The bad news is it's down 50% off the highs.

Jordy Visser: I wrote a paper last week on Substack about taking a stoic approach to everything in life. But it's really important with markets and I think for people that are either frustrated or people that are really happy and pounding on X saying the bubble's going to unwind, strategy is going to go bust -- none of that's true in my opinion. I also don't believe that Micron can continue to go up every single week. So on one side we have this whole recognition of hardware -- everything should be bought. We had software go down. Now you've had some of the software names go up. Most of them, people realize, are not true software names. They are AI software names -- cyber names and the Bitcoin miners and things like that. So I think Bitcoin's been lumped in.

Jordy Visser: I'm starting to write about this more. I believe we're at a rotation point that's going to last for at least the next 3 to 6 months. And that rotation is kind of a rotation bubble. We've had one group winning and most groups have been losing. Starting to see a rotation. And I think software is going to have a bid. But the software is not going to be the traditional SaaS software. I think people are going to have to think outside the box on human software and also the financial guardrails with tokenization and everything. And that's where my attention is focused.

Anthony Pompliano: You told me months ago the four-year cycle was basically like male astrology. Why does this four-year cycle exist? Who made this up? Are you a believer yet?

Jordy Visser: No. First of all, I don't do the same thing with the stock market. One thing I will say is this is the first time in that quote-unquote four-year cycle that stocks are going higher at the same time that Bitcoin's going down. And I don't know whether that's good or bad. My whole belief has been that the only time crypto can go through the growth I believe it will -- which is the transfer of wealth from the fiat system into the crypto side or at least a merging of the two -- for that to occur you need to run to a point where they have to be uncorrelated to some degree. And I think we're at that phase now.

[05:02]

Jordy Visser: The 200-week moving average is a really important thing. And if you lock it into what you said -- what is the 200-week moving average? -- well, that's kind of a 4-year moving average. So if the four-year cycle is this, it's fine. I think the IPOs that are coming, what Google did this past week -- these are a sign of something important. It's not a stock market bubble to me, but it is a sign that people need a lot of capital right now for what is happening inside the AI world. And that's an important story. On the flip side with what's going on with crypto, if you would have taken the ETF launch and the US government basically deciding crypto is good and the president launching a memecoin -- if you would have taken those events and said what would that be -- well, that would be a sell-the-news event. And yet Bitcoin went higher into those. I do think there's been churn happening now since 2024. For people looking for hope, until the charts start giving you hope and you get the going-higher-on-bad-news type signal, I'm not going to get too into it. I'll still stick with the main thing: we right now have three-month rates that are below inflation. Just wait for the next trend higher in Bitcoin.

Anthony Pompliano: What could convince you to sell your Bitcoin?

Jordy Visser: I know this is going to not be the answer people want to hear, but there's a reason why all of my eggs are not in one basket. I have plenty of things. Does everyone who's investing own things that are probably going to zero? Yes. Everyone might sit there and say no, but if you own the S&P 500, trust me, you own some things that are going to zero. Bitcoin is not everything in my life. Do I think Bitcoin might go to zero? Sure. It's part of the distribution of outcomes. Do I think quantum is going to do it? I think it's a very low probability. So again, everything in my mind is always based on what my father taught me -- what do I think the upside is versus what do I think the downside is? And then what percentage of the bets do I want it to be? It's not everything for me. So if people are sitting there and they have 100% of their wealth in Bitcoin, I think that's very stupid. If they have less than 1%, I think that's very stupid. I believe it should be at least 2 to 3% of everyone's portfolio regardless of what they believe in.

Jordy Visser: I've never sold a single Bitcoin. I have traded MicroStrategy many many times -- got in, got out, got in, got out. I don't really own much Ethereum at all. I've been buying Ethereum during this whole drop. So I'm happy it's going down because I'm putting bits and pieces in. If we start to trend higher, I'm going to treat this like I did Micron last year. I bought Micron from 105 down to 60 all beginning part of last year and look where it is today. That's the way I view Bitcoin for next year.

Anthony Pompliano: Jordy, you can't come into the church of a podcast and talk about Ethereum. Why are you buying Ethereum if Bitcoin is so cheap?

Jordy Visser: Well, I'm buying Bitcoin as well. The Ethereum thing is more of an ecosystem play at this point. I really am betting that the network effects are going to be the big story for the next 12 months. I think Bitcoin is going to be a participant in this whole situation. I am working on a weekly YouTube specifically geared towards crypto. I want to convert for the traditional finance crowd how they can look at the crypto ecosystem in the same way they look at the S&P 500 -- some sort of sectors or themes within. And so for me Ethereum is a better proxy for that. I want to bet on the ecosystem. But the reality is when I create an equal weight of 40 separate companies -- 34 of them being cryptocurrencies or tokens, six being public stocks -- all as one index, very similar to my agentic thematic portfolio, it's extremely correlated to Bitcoin even when you don't include Bitcoin in it.

[10:00]

Jordy Visser: So for me, that means the ecosystem is acting the way I think it should. It also means that Bitcoin is what I think it is -- the S&P 500 of the crypto world. At some point it's the king. It's the only thing that lasts at the end of time. I don't believe any idea or innovation in the history of mankind has ever lasted. There's always a new thing that's better. That is what stocks are. That is what tokens are. They're ideas. They're innovations by human beings. And if AI is now going to be creating all of the ideas, that means they're not going to be around that long. There'll be something better that comes up. You're going to start seeing this far more rapidly.

Jordy Visser: I just think for this whole thing, when people really think about crypto, you have to make a decision on whether you believe in the long-term story. And if you don't, you should be out. If you still believe in the long-term story, then it's a gift that you're getting to put things down here. I'm not only buying for me. I have strategic Bitcoin reserves for my kids. This is a blessing to me. My business is growing and I'm able to put more of the profits back into Bitcoin for my kids. So if it doesn't work out, sorry guys, you guys lost out. If it does work out the way I think it will, they're going to be very, very happy in a few years.

Jordy Visser: There's a Patek Philippe ad campaign that talks about "it's not your watch, you're merely taking care of it for the next generation." And this idea that when you buy this, it is not only going to last, but you are going to pass it down. When you have kids, you start to think about what if I don't think of this portfolio as mine? I think about it as theirs. The hardest problem in finance is the intertemporal transfer of wealth -- how do you give yourself money 20 years from now? I don't know of an asset more than Bitcoin that people have confidence in over some very long period of time -- 50 years, 60 years -- is going to continue to do that.

[15:00]

Jordy Visser: I believe in Moore's law, Ray Kurzweil, singularity -- all of that stuff is why I believe in Bitcoin. It's ironic that Bitcoin is as of today a trillion-and-a-half-dollar asset. Isn't it ironic that SpaceX is coming and it's about a trillion-and-a-half asset? How can people justify buying a company that not only is not making money right now? When I talk to people who love rocket investments and I go -- how is that any different than Bitcoin? How do you have any idea on something that is more than three years out? There's no way to know what the future's going to look like when you're dealing with something that's based on going to Mars, which has never been done before. When I think of investments and the future, I don't think investing in the S&P 500 is safe past 2030. I've said that repeatedly because I think AI will disrupt all companies. I don't think Apple's safe. I don't think Google's safe. I don't think Amazon's safe. So when people look at Bitcoin and they don't think it's safe, I understand that's their belief. The question is, have they thought a lot about the future?

Jordy Visser: One of the great things about Moonpay is the agentic side. It came out on X this week that there are more AI agents on HTML sites than humans. This is where we're headed. So if you think the world of human beings is going to dominate commerce, it's not. It's AI agents. So if you ask me what is a direct play on that, I believe Bitcoin is a direct play on agents dominating humans going forward.

[Ad -- Figure: crypto-backed loans, HELOC innovation, real-world asset yield]

Anthony Pompliano: I agree with you that Bitcoin is a safer bet than any company just given all of the complexities, all the changes over let's say 20 years. Does that then mean if you get paid for the risk that you take, that companies should be more asymmetric and Bitcoin should not be asymmetric because it is quote-unquote safer and therefore the return profile should be lower?

Jordy Visser: Here's what's happening in the world today and here's where I say Bitcoin actually set the framework for this. This 50% crash has occurred with volatility collapsing. It's not been because of FTX. The volatility is going down. Bitcoin's bleeding like 2% a week. It's not going down 15% in one day anymore. But every day we're watching the leading stocks in the stock market move 8 to 10% a day. We're seeing individual stocks trading at the highest volatility relative to index vol that we've ever seen. This represents to me that the traders of the world -- very zero-DTE option related, very retail related -- it's training people to get ready more for parabolics and bubbles and speed crashes. That is what I think we're in now. That means that Bitcoin is acceptable as an asset because one of the big negatives was it was too volatile. It's not too volatile anymore relative to the things that are making alpha in the market.

Jordy Visser: The agentic move is compressing time. The MAG7 went from 1 trillion to 20-plus trillion in basically a decade. Isn't that a bubble? It's a bubble except for one thing -- it just happened over a longer time period. So if Micron can go from 60 to 1,000 and then goes back to 600 -- was it a bull market, was it a bear market, or was it both? And that's what Bitcoin is. Maybe these aren't four-year cycles. Maybe these are now nine-month cycles and we've compressed the time. And that's what a bubble is -- it's price versus time. AI is speeding up time. The exponential age is speeding up time at a pace we've never seen before.

[20:01]

Anthony Pompliano: So there's a couple pieces of this. First of all, we need more bubbles and we need more billionaires -- which I think are two ideas that are very counterintuitive to people. But the bubbles bring capital and you've mentioned that these companies need the capital to build out the infrastructure for what is going to be the future.

Anthony Pompliano: Now with Bitcoin, and I think that Micron and some of these other companies are now experiencing the same thing -- crypto over the last 15 years may have been the best training ground to be an investor in any asset class because what you essentially did is compress all these market cycles into a very short period of time. Volatility was magnified. The stat that I always tell people: 50% drawdown in public equities in the global financial crisis. Bitcoin has gone through that every 18 months for a decade. So you just see this stuff over and over.

Anthony Pompliano: The framework I've used for Bitcoin -- why does it go up and then crash, then go up and crash? It's very similar to a mobile app. If I created a mobile app with you, we want to go get users. We'd run a marketing campaign. 100 people come in and sign up. Some portion of those people are not going to stay. So let's say 30 churn out. We're left with a net gain of 70 people. We run another marketing campaign. Another 100 people come in, but 30 of those churn out. Now we have 140 people who are left. And you keep doing this. That's basically what we've seen in Bitcoin -- you get these big run-ups in price, a bunch of people pile in, it kind of bleeds out, some people leave, but you've converted new people. AI is starting to do this now.

Jordy Visser: The worst part about this -- we're now at a point where when we say AI, let's use Jensen Huang's five-layer cake. You got the chips, the energy, the infrastructure, the models, and the applications. The only thing that's been monetized or the only thing making money are the chips, the energy, and the infrastructure. There's a problem with that eventually. This is all about the buildout in preparation for it, and eventually you do have to get the revenues. A data center for one gigawatt has gone from $50 billion to $60-80 billion, and Jensen Huang said at Computex it'll very quickly be $80-100 billion. So you're talking about massive inflation in building this stuff. That's how bad the bottleneck situation is.

[25:01]

Jordy Visser: I believe Eli Lilly has a chance to be the largest company in the world and the number one AI company in the world within five years because they're building a specialized model. They have their own data center with a thousand GPUs and all of the data that Eli Lilly has had from 150 years as a company -- all the failures, all the successes. If you combine all this with Alphafold where they have a partnership with Isomorphic, combine it with their Nvidia co-innovation lab, combine it with Toune Lab which is their Google X -- they allow people to use their data center and the only thing they get back is the data. They are creating an innovation hub of using data all for human software. If I think about who's going to get the revenues -- well, if a drug company can make their entire process of getting FDA approval efficient and at the same time they can benefit by selling better drugs, they're getting revenue directly. That's an easy one for me to understand.

Anthony Pompliano: I want to talk about Eli Lilly a little more. On this general versus specialized workflow, we have direct experience with this -- we're building CFO Sylvia. The single most important thing I see happening in the AI industry is that the general purpose models were the default -- no one knew how to build specialized workflows. So OpenAI came out, Anthropic started to push Claude, everyone basically mandated to their teams: go use AI. And we're now seeing reports -- whether it's Amazon, JP Morgan, or Uber -- they all report that they blew through their token budgets in a single quarter for the entire year.

Anthony Pompliano: People are starting to question: are we getting the value for what we're spending? What we saw at CFO Sylvia is that we gave token consumption to the user. So the user was in charge of how much times they queried. You essentially could think of that as uncapped upside for usage but uncapped expenses for us as a company. So then that forced us, if we want to keep cost under control, to look at how to efficiently consume tokens. We have seen a significant reduction in the number of tokens used on a per-query basis. And then I started talking to many of my friends who run companies. They're all going through the exact same thing -- if I'm spending hundreds of thousands or millions of dollars per year using some sort of AI system, how do I get the same output but get a cheaper bill? It's human nature. It's corporate incentive.

Anthony Pompliano: In a weird way it makes Anthropic's revenue growth even more impressive that they are growing at the rate they're going at the same time that every single one of their customers is trying to become more efficient using their service. So you have per-customer per-token revenue coming down but overall revenue growing so rapidly because demand is insatiable. When I look at that, it does split the world into general purpose versus specialized workflow. I think finance, healthcare -- there are very specific areas where it's going to be incredibly difficult.

[30:00]

Jordy Visser: The cost of building the ability of providing the tokens is going up dramatically. So they have to raise the cost. So users who are very smart are like -- wait, I'm not paying for this. So then you start figuring out ways to not pay for it. David Baker said on an interview about six months ago that the bearish argument for all of the model providers is eventually going to be the edge. The problem with the edge is it's coming sooner than what people think.

Jordy Visser: When we talk about Anthropic and we're like it's 44 -- it's not 44 billion, that's the ARR. So going like this for three months does not ensure you'll be doing that for another three years. And you're still playing with this game of -- we need more compute and they're racing to Colossus and to Google and trying to get all this stuff. It seems very clear to me that there's a problem with bottlenecks and shortages. And oh by the way, the Strait of Hormuz is still shut. I can't say this loud enough. If you believe the strait will be completely reopened to normal before the midterms, go bet on PredictIt because right now that probability is less than 50%. So we're getting to the point where all of these bottlenecks are building.

Jordy Visser: As someone focused on bottlenecks, I want to focus on what Eli Lilly is doing because they don't depend on tokens. They spent the money on a thousand Blackwells to build a data center to take their data. So you have to think real hard about why do they have an advantage? Because they are printing money and they can build their own. For companies like Morgan Stanley and Goldman Sachs, they're going to have to make a choice at some point -- do I want to pay Anthropic or is this a way to get started? There's a combination of hoarding on the hardware side, FOMO on both the model side and the enterprise side. Goldman doesn't want to fall behind Morgan, doesn't want to fall behind Bank of America or JP Morgan. So they all have to start spending money.

Anthony Pompliano: I recently saw one person online say they have completely eradicated the use of Claude internally and gone to DeepSeek V4, and they believe it's saving them a lot of money and performing better on accuracy and latency. It will be very interesting to watch over the next 30 to 60 days if more people start saying -- what are we getting for the token bill? Can I get away with using an open-source model that maybe we fine-tune a little and get 90% -- maybe even better -- performance than what I'm getting here with Claude?

Anthony Pompliano: I actually think this is a pie-expanding exercise, so Anthropic is going to continue to do well. OpenAI will continue to do well. I think XAI is very underrated in terms of their ability to come out of nowhere and start to get adoption. At the end of the day what people are always going to be incentivized to do is control as much of the stack as possible. If you have a model -- whether it started as open source that you fine-tuned or you were able to actually go and do it similar to what Eli Lilly did -- you are going to have such a competitive advantage. No wonder young people are having a hard time trying to find a job. People are saying I can go and get this same intelligence in a different way. I don't have to do recruiting.

[35:00]

Anthony Pompliano: My conclusion over the last couple weeks is the single most important thing you can do right now is to remain with incredible flexibility mentally because it is changing very quickly. If you try to take a static snapshot you are going to be wrong three weeks later. So you've got to stay informed. You've got to continue to update your mental models here. If you do that, that's the best way to navigate all this.

Jordy Visser: The only thing I would add is the LLMs are now commoditized. The difference between Claude and ChatGPT -- they're indistinguishable to me in terms of now being able to use them. I got into a conversation with someone we both know this week who had just gotten off a plane, and I said "I don't really use Claude as much." And I was the person that said you really have to use Claude, not ChatGPT. So people are starting to ask me -- are you using Codex and ChatGPT? And I'm like yeah, that's my model. I'm using both of them now.

Jordy Visser: The reason I say they're commoditized is similar to hiring. Assume you had four pieces of information for each candidate -- grades in school, both went to the exact same college, both finished with perfect scores, SATs exactly the same, IQ exactly the same. Now you're going to pick the person not based on that. You're going to pick based on the nuances that match with you. They're not going to be the same person even if they have exactly the same scores -- which is where I believe the LLMs are now. Maybe Claude is slightly better at coding, but I'm not really sure about that anymore either. I'm building stuff in Codex every day. The reason I chose ChatGPT 5.5 is because it is the better brainstorming partner for me. When I meet intelligent people, only two things might enter from an hour conversation, but those two jewels are worth so much to me. ChatGPT does it all the time. Claude almost never does it for me anymore. I just believe that the critical things that make my brain go "oh my gosh" happen more in GPT 5.5. It's better at anticipating what Jordy wants, and that's why I've migrated back towards it.

[Ad -- Arch Public: agentic trading platform, 25,000 users, no custody of private keys]

[Ad -- Simple Mining: white-glove Bitcoin mining hosting, Cedar Falls Iowa, 40,000 machines under management]

[40:00]

Anthony Pompliano: Today you said something to me that was very interesting before we started. Peptides are the API key for the human body. Now I have learned throughout my life there are two types of experts: medical doctors and the bro scientists. The bro scientists have been on the peptide train for a while now and it feels like the rest of the world is now saying -- oh, peptides, that's interesting.

Jordy Visser: I want to take a step back. People who are not scientists never enter the biotech and pharma world as an investor, because you can't understand the science. But for some reason we all became experts at technology -- we know what a GPU is, we know what Moore's law is -- because that was the way we all made money. So why would people know what an API key is? An API key and a peptide are basically ways to enter into something with a receptor. I want to do this, I want to take my data from this software, I use my API key. It's a door into something. And GLP-1s are peptides. Peptides are growing exponentially.

Jordy Visser: I can't go into any setting without hearing it. I worked out today. The health coach at the desk -- before I could finish saying peptide, she said "what percentage of the people that you have as clients every single day come in and talk about peptides?" And she went "everybody." She said "if you're talking about GLP-1s, more than 50% of my clients are on GLP-1." Now, this is Manhattan. But they're growing rapidly.

Jordy Visser: If the decade of 2009 to 2020 was about the app store and API keys, you could argue that API keys were the key to all of the money made by the Mag 7. I believe the pharma industry with peptides has locked into something where -- whether it's gene editing, any way you want to go -- we're getting to the software side of the body. We're actually getting to the point of understanding this does this to someone. So Eli Lilly has data that is unexplained for them. And if you listen to David Ricks, he'll say -- we knew it was losing weight, we knew it was helping with diabetes. What we didn't understand is why is addiction going down?

[45:00]

Jordy Visser: We're getting to the point of actually getting to the root cause of problems as opposed to the treatment of problems. And if you get to the root cause of problems through peptides and through receptors, you're now putting an API key into the human body to deal with something. This is the beginning of a massive trend because if it wasn't for AI, we're still kind of guessing and going through a long process. But AI helps speed up the process. Very recently -- I think in the last month or two -- Eli Lilly stopped disclosing phase one trials they're working on, which is when they're actually using human beings. They won't disclose it anymore. Why? Because it's much easier to catch up to IP if you have AI. What we're watching in front of our face is human software. The TAM market for obesity in the world is between $200 and $300 billion.

Jordy Visser: I believe we are seeing signs of a top in hardware. Not that they're going to go down and not that this is a crash, but you look for rotations to where the market narrative shifts. And I think the market narrative is shifting from the chips, infrastructure, energy side of the cake and it's moving up to the application side. And I think the best application for this is going to be on the human software side. One of the aspects I became very interested in is the second I heard that GLP-1s could help with addiction. And the data point that really hammered it home was that some hedge funds on Wall Street were requiring people to not take GLP-1s because it took away some of the risk-taking component.

Anthony Pompliano: What happens when this hits the drug and alcohol addiction market? If you're telling me there is now something that can be injected that will help solve that problem and it maybe is matched with recovery centers or rehab or whatever -- that is a very big market. That is a market that not only would have a profound positive impact on society, but there will be very big businesses built there. And so to me what's interesting is you're talking about this API key for the human body, but it seems like a lot of these peptides are not just single application. We're starting to see -- okay, it can do weight loss, it can do addiction -- it's one single drug or peptide. When we look at the cholesterol shot, my guess is it does other things as well. As an investor, what else do you want? It's already a big TAM, and oh by the way it's kind of like the SpaceX TAM slide -- rocket launch, Starlink, enterprise AI -- same thing is going to happen here in healthcare.

[50:01]

Jordy Visser: When you talked about addiction -- people hear the word side effect and they think it's something negative. Well, these side effects of addiction going down, all these are just one of the things that are coming that they can now go study. I believe what Eli Lilly has figured out is side effects now become something very useful to go investigate as to why this is having an impact. This is different than spending an enormous amount of years to get a drug that treats something, as opposed to a peptide which is there to make the body act differently.

Jordy Visser: I highly recommend listening to an interview on YouTube with David Ricks, the CEO of Eli Lilly, and Jensen Huang. They're partners. Eli Lilly is partners with Nvidia. That should be enough for people to be interested. But if you listen to the interview, you'll learn more about why GLP-1s are so important and why peptides are so important and you'll get a sense as to where Eli Lilly is going with this.

Jordy Visser: I've never worried about the debt of the country despite all the doom and gloom because the household net worth of the country is five times the size of the debt. Yes, the government has a bad balance sheet, but the household balance sheet is phenomenal. But entitlements are a major issue. And what makes up the bulk of entitlements? Medicare and Medicaid -- the health of the country. If we're going to get to the point with peptides, gene editing, mRNA -- everything -- where people are going to die healthy as opposed to die sick, that has a huge impact on the entitlement issue.

Jordy Visser: Watch Eli Lilly stock because number one, it is a trillion-dollar company. Number two, it's been consolidating to a very high degree for the last 18 months. Victoria's Secret put their earnings out. The stock was up 47% earlier this week. They reported 15% sales growth in Q1 directly attributable in large part to GLP-1s -- people are losing weight. Bedroom fashion is back. The second-order effects here on some of these businesses. Everyone was talking about fast food companies going under pressure. But there are the positive stories as well. And so it does feel like this is going to infiltrate out into the US economy in a bunch of ways we didn't predict.

[55:00]

Jordy Visser: I can already see the supersonic tsunami of healthcare is approaching. So I think there's going to be more and more Victoria's Secret stories going forward. What Eli Lilly is doing -- they're making so much money from this one drug that they're taking the cash to build a data center. This is the greatest story you could ever want to hear. Toune Lab is basically Google X. Their co-innovation lab out in Silicon Valley with Nvidia is to speed up health and attract talent from Stanford -- people that are graduating there making sure that when they have a choice between going to work for a software company or a human software company, that Eli Lilly is the answer.

Anthony Pompliano: What are you going to put in your video this week?

Jordy Visser: There's obviously going to be a lot of Eli Lilly, teasing out some of the things I said here. There's going to be a lot on my belief that the IPOs coming to the market -- people are underestimating what they signal. The Google deal, the $85 billion, why Berkshire is buying it, why that says something -- a company that's been raising cash decides oh we'll take $10 billion of a company that's up 120% year-over-year. A lot of stuff going on in the market that people need to be paying attention to.

Jordy Visser: I will emphasize one other point. We talked about the human software side. I'm not going to underemphasize the importance of crypto and tokenization on the financial guardrails, but that is the other software side. If and when Bitcoin breaks through the 200-day moving average -- if and when, maybe it's two years from now, maybe three years, maybe it goes to zero before then, I don't know -- once it does, I believe we're entering a new phase where the software that people are choosing is for the human body and for the financial guardrails, because AI agents need both parts of that software.

Anthony Pompliano: I love it. Three asks of the audience today. Go search on YouTube for Jordy Visser and subscribe to his channel so you can watch his weekly video -- every Sunday. Second, go to Jordy Visser 22V Research. He's putting out two pieces a week. Third, if you want to use AI to better manage your portfolio, go to cfosylvia.com. You can sign up there for free. Go subscribe to Jordy on YouTube, go subscribe to 22V Research, or go sign up for cfosylvia.com. If you do those three things, we'll be happy. We'll see you guys next week.

AI Master Prompt

The AI prompt on this page is auto-generated from the transcript content and is intended to support further exploration of the topics, concepts, and conclusions discussed. It is provided for informational purposes only. The user is solely responsible for all outcomes resulting from its use.

Master Prompt
You are a strategic investment and macro research assistant trained on the frameworks discussed in a conversation between Anthony Pompliano and Jordy Visser (22V Research) recorded in June 2026. The conversation covers Bitcoin's 50% drawdown, the AI hardware-to-application rotation, Eli Lilly as the leading AI trade, GLP-1 peptides as platform technology, and the mental model required to navigate exponential change. CORE FRAMEWORK Jordy Visser operates from a few key principles. First, Bitcoin is not a company or an idea -- it is a neutral settlement layer and store of value that survives all innovation cycles the way an index survives the failure of individual components. It is the S&P 500 of crypto. Second, market cycles are compressing under AI-driven time acceleration. What used to be a four-year cycle may now be a nine-month cycle. Price-versus-time is the definition of a bubble, and the exponential age is running that clock faster than any prior era. Third, the AI investment stack has five layers: chips, energy, infrastructure, models, and applications. Only the bottom three have been monetized so far. The narrative is rotating upward to applications -- specifically human-body software (pharma, peptides, gene editing) and financial-guardrail software (crypto, tokenization). Fourth, Eli Lilly is Visser's highest-conviction AI play because it has 150 years of proprietary trial data, a specialized model, its own data center, and Nvidia partnership infrastructure -- everything a generalist model provider cannot replicate or price-compete against. Fifth, peptides (including GLP-1s) function like API keys for the human body -- receptor-targeted signals that produce cascading biological effects far beyond the initial indication, with side effects that are now being studied as discovery signals rather than liabilities. KEY PRINCIPLES -- Bitcoin at 2-3% of portfolio is a minimum defensible position regardless of conviction level; below 1% or above concentrated bet is both intellectually dishonest on opposite ends -- The 200-day moving average is the re-entry signal; no aggressive deployment until Bitcoin recaptures and holds that line -- Volatility compression in Bitcoin relative to individual equities is a sign of asset class maturation, not weakness -- The hardest problem in finance is intertemporal wealth transfer -- giving yourself or your children money 20-50 years from now -- and Bitcoin is currently the strongest answer to that problem -- Generalist LLMs are commoditizing at the output level; differentiation will move entirely to specialized, data-rich workflows -- Enterprise token budgets blowing up drives efficiency optimization that is a tailwind for specialized model operators and a threat to generalist API revenue per customer -- Mental flexibility -- refusing to lock into a static market snapshot -- is the most valuable investor skill in an exponential environment WHAT THIS IS NOT This is not a framework for short-term Bitcoin trading, price-level prediction, or momentum following. Visser explicitly says he does not call levels up or down. This is not a framework for blanket crypto enthusiasm -- he is skeptical of most tokens as ideas that will eventually be replaced. This is not a "hardware is dead" argument -- the rotation thesis says hardware names pause and consolidate, not that they crash. And this is not medical advice in any form -- the peptide and GLP-1 discussion is an investment thesis rooted in market data, not clinical guidance. HOW TO USE THIS CHAT 1. APPLY THE ROTATION FRAMEWORK: Tell me what AI or tech positions you currently hold and I will help you map them against Jensen Huang's five-layer cake and assess whether you are overweight the bottom three layers relative to the application layer. 2. BUILD A BITCOIN POSITION THESIS: Tell me your current allocation, time horizon, and conviction level and I will help you develop a structured position thesis using Visser's moving-average and volatility frameworks. 3. ANALYZE ELI LILLY AND PEPTIDE PLAYS: Tell me what you already know about Eli Lilly, GLP-1s, or adjacent healthcare positions and I will help you map the second-order effects and assess competitive moats. 4. STRESS-TEST YOUR ASSUMPTIONS: Share a market view or portfolio thesis and I will challenge it against the mental flexibility principle -- identifying where a static snapshot might be wrong in 30-60 days. 5. EXPLORE INTERTEMPORAL WEALTH TRANSFER: Tell me your current portfolio structure and generational goals and I will help you identify which positions are truly built for durability over 20-50 year horizons versus which are optimized for shorter cycles. 6. MAP TOKEN ECONOMICS: If you are building on or investing in AI infrastructure, tell me your current setup and I will help you work through the generalist-versus-specialized model decision and the edge-migration timeline. TONE INSTRUCTION Be direct, macro-grounded, and framework-first. Do not preach about risk management -- apply it analytically. Help me think through the first, second, and third-order effects of whatever I am examining. Challenge static views. If I ask for a price prediction, redirect to the signal framework instead. To begin: What is the biggest position or market view you are currently sitting with, and what would need to happen in the next 90 days for you to change your mind about it?