市场没有崩盘，人类没有灭亡——但加密货币刚刚突破

Humanity is still here.

Most weeks when I do this, there’s actually a lot to talk about in the market. When you get a week like this, where you get all the bear porn out on Sunday, I get contacted by more hedge fund people in one weekend than I have in a long time to see how the AI trade is going to play. I told pretty much everyone who reached out that this is just another one of these knee-jerk reactions. By the time we get a week to 10 days later, nobody will remember it. I think everyone keeps learning their lessons on this.

Overall, it was a very quiet week for stocks. It was a very quiet week for AI trades. It was just an overall very quiet week, despite the fact that we had the AI pause, a 10% chance that humanity dies, oil soaring higher on Monday, 10-year rates reaching 2007 levels for the first time, and the Fed raising rates in a hawkish hike. If you sold the market on Monday morning when it opened, you lost. If you sold it at the close on Wednesday—which a lot of people I follow said was the time to be out of equities and everything else—somehow or another, we were able to rally.

1. Crypto is the story of the month: AI and crypto are merging, the China 2003–2013 and dot-com-to-iPhone analogs, and the quarter-to-date scoreboard (ETH +66%, BTC +53%, SOL +38% vs. semis -14%). The 46-name index at new highs implies Bitcoin another 15% higher.

The story of the week and the story for the month is crypto. I have talked endlessly about this, and we are finally at the point. Last May and June, I said that by the time we got to September or October, we would be at the most important inflection point for crypto. The reason would be the emergence of AI agents. AI and crypto are together. Crypto does not exist without AI. It was built for AI, and most people don’t recognize that.

Before I start this, I want to make sure there are a lot of things I’m going to go through here. This week I will be sending out my first video on crypto for subscribers. So for those of you who watch this every week that have been waiting for it, this will be basically a very extended version on why now. I’m not going to go through all of that today. I’m not going to take you through the details. But as I’ve said, I’ve created an index. It made new highs for the year today. I posted that on X for macro people. This is the most important macro trade of everyone’s existence.

This is the beginning of something that is very, very similar to the times that I traveled to China. For any of you I traveled to China with whom I haven’t spoken, think about those times we went to China from 2003 to 2013, when they were building roads bigger than what they needed, and they had ghost cities and all of this stuff. Think about the dot-com bubble, where we were building the infrastructure necessary, but we didn’t actually get the iPhone until 2007.

All these stories that you’re hearing about debt and crashes are all bullshit. It’s all meant to scare you. Everything I read on AI is just a complete gobbledygook of garbage. It’s people literally writing things that have nothing to do with facts. They might be right down the road, but you can’t argue with it if you don’t use it.

What I’m going to do today, for a bit of this, is go over where we are in crypto, why it was built, and why we’ve gone through this long period of 16 years. It’s very similar to the period from the internet to when the iPhone came out. It’s almost the exact amount of time. The reason the iPhone and the App Store mattered is that there is no Uber, there is no Airbnb, and there is almost nothing you do in your life without the smartphone and without the App Store.

For crypto, I will tell you that your entire world is changing. There has not been a macro change like this in the history of mankind. We are moving to digital transactions with digital agents and digital employees. Everything is going digital.

If you want to understand what tokenization is, if you want to understand the names that are in there—what Zcash is, what Bittensor is, what Algorand is, and so on, up and down the line—all of these things are going to be a part of your life whether you like it or not. So you either learn or you don’t. At least with me, you’re getting someone who’s doing it from a systems-thinking perspective, who’s a macro person. For those who know me and like what I do on the AI side, they like the way I talk about macro. I’ve been spending my time making sure that I can give you the same thing in crypto.

You guys can read this on your own. Do not take anything I say as being certain. I’m wrong all the time. The key is to maximize upside and minimize downside.

I’m going to show this again because, after today, quarter to date, here are the stocks. I’m going to say this right now: I don’t think we will ever have another big bull market in stocks like the one we saw in the first 2 quarters of this year for AI. I believe the multiple compression is the old system basically going through something there. The only way I thought we’d get a surge in stocks would be if the Fed cut rates. The fact that they raised rates and stocks are just sitting here while earnings continue to grow is important.

You’ll still get returns in stocks, but remember, bonds are basically unchanged. For how long now? I’m sure we’ll be here in 2 years and say it’s been a decade, but regardless, stocks so far this quarter are up 2% for the S&P, while the Nasdaq is down 3% and semiconductors are down 14%. Ethereum is up 66%, Bitcoin is up 53%, and Solana is up 38%. Those are the 3 main ones. Those are the layer 1s that I want you to focus on.

Underneath the hood for crypto, to get a quick thing out while you’re sitting there spinning your wheels—not just this month, but for the entire quarter—trying to chase stocks up and down, responding to oil, responding to rates, and watching this whole thing while just getting gutted: everyone made their money in the first 2 quarters of the year when we got the rally in AI. I’m not sure going forward that we’re going to have the kinds of rallies that are necessary for people to keep going.

I do think the market will go higher. If you’re running long-short, I would guess that between now and the end of the first quarter, we’re going to see a good rally in stocks. But I wouldn’t bet heavily on it because I don’t think you’re going to get much return.

In crypto, here’s what you’re getting on Friday. I’m not showing you the names. These are the names in my 46-name portfolio. Some of these are public stocks. This is the vertical they represent: 21% in a day, almost 100% for the month. Look at these numbers. Is it above the 50-day? Yes. Yes. Yes. We’re at a different point now.

Here’s what the chart looks like. I said my 46-name index broke to new highs for the year. That puts Bitcoin up significantly from here—another 15% from here. This is Bitcoin overlaid with it. The correlation remains basically 1-to-1. The difference is that the ecosystem is doing better. Bitcoin, remember, is part of the ecosystem.

I posted this on X: “10% chance AI kills humanity. 100% chance crypto was built for AI.” For all of you who have made money on AI this year, all the macro teams that have jumped into AI, and all the pod teams, crypto is now where you want to have your focus.

2. Why now: the ghost city was never meant for humans or the $900 trillion fiat system. Liquidity plus AI plus blockchain, M2 no longer matters, velocity of money explodes on chain. Learning crypto is like learning cricket when you know baseball.

Why now? I’m not going to spend time on it today, but for those of you who know I traveled to China, China built ghost cities. Ghost cities have no economic activity, but they have a bunch of infrastructure. That’s where crypto is.

The problem is that crypto was never meant for human beings. Even though the crypto crowd thought it was, it was never going to get the wealthiest people who own the $900 trillion—the banks and the institutions—to actually start using it. What cares about it are AI agents, and they’re already starting to.

This is the reason why I have waited and been patient. This is why “why now” is about the perfect convergence: liquidity, plus AI, plus the blockchain. In crypto, I hear this from both sides: the liquidity cycle, what we’re doing with M2, doesn’t matter anymore relative to what’s happening with tokenization and what’s going to happen with the velocity of money.

We are going to see an explosion in the velocity of money as everything moves on-chain gradually over time, because it will be cheaper to transact on-chain. Will people be slow to move? Of course. That’s what they did with Uber. We all know parents and grandparents who either never used Uber or barely moved over to it. But in a world where every penny matters and where time matters, you’re going to use the agents. They’re here, and they’re being valued highly. This applies to private companies, too. I’m going to take you through that as we go on.

The reason now is that we couldn’t have crypto benefit until we reached the automated AI researcher. This is hedge fund manager Leopold Aschenbrenner’s 165-page paper, Situational Awareness: The Decade Ahead, which he wrote in 2024. When we hit this point, we get an explosion in all of these things, including GDP, military edge, robots, science and technology, and cognitive labor. I’ll go through this. This is where we are. We’re at that inflection point, which is why crypto matters.

The last thing I’m going to do on crypto before I go through the week that was is this: I want you to understand that I’ve spent the better part of 5 years mainly focused on Bitcoin, then spreading out into the ecosystem, particularly over the course of the last 2 years. Thinking that you understand crypto, or that you don’t need to know it, is the same thing as trying to say that if you know baseball, you don’t know cricket.

I’m bringing this up because of the complexity of learning a completely new language: why tokenization matters, why stablecoins matter, what Ethereum is, what Solana is, what Bitcoin is, what Zcash is, and how they all fit into the equation. It takes a lot of time. For the people making decisions at the highest level of a hedge fund, it’s very, very hard to pick this up.

What I will be spending my time doing for the next 2 months, for sure, is going through “Why now.” Then I will start going through each of the 10 verticals, probably 2 or 3 a week, to take you through the names and give you an understanding of what’s going on.

This is what I will be talking about. That is what will be on the subscriber website for hedge funds. If you don’t pay for the subscriber website or mutual funds or whatever, this stuff is going behind the paywall. This is what I do presentations on right now for financial advisors and RIAs. If you’re worried that you’re going to fall behind with your clients, the same way with AI, this is far more important than AI because it’s very easy for someone to understand what Micron does. They just have to go read.

But to understand what Zcash is, what NEAR is, and I can go on and on, it’s going to take a lot more learning. Your viewpoint of it is that it’s a scam.

3. Bearish forecasts vs. reality: the Anthropic slowdown call drove the biggest software-versus-semis outperformance, doom content goes viral ahead of a midterm, demand remains greater than supply, and China is not slowing down.

So, “The Week That Was”: I did my thing on Friday last week. Dario came out and said, “We have to slow down.” That got everyone freaked out. So Monday, we get in here. You get cyber stocks gapping up, which moves the software sector up. You get semiconductors down.

What you end up with is a 10% difference—the biggest outperformance of software versus semiconductors. I’m sure it hurt a lot of people. An Anthropic researcher believes there is more than a 10% chance AI could kill all humans. Incredibly stupid, and there’s no reason for you guys to worry if that’s what’s going to happen. It’s not something you can defend against.

It’s very hard to come up with how you kill all of humanity. We can kill people, and I’m sure we will. Anthropic warns that AI agent swarms could seize the internet within a year. This is a story that’s now going around more and more.

Again, I 100% believe that this is possible, especially as you read the Hugging Face thing. I 100% subscribe to the fact that there will be an issue, and humans will be scared, just like they’ve been with every new technology, bar none. This is not the first time we’ve worried about humanity. Could AI really wipe out humanity? It’s a global thing.

So, let’s go through the bearish forecast that we get constantly and the reality. The reason I said the BS and the bear porn—whatever you’re reading, whether it’s Michael Burry, Jim Chanos, or the stuff on X that comes out all the time—is that GPU prices are going to fall. I get stuff every single week, guys.

I had a CPA send me something this week to try to prove to me that this is just a fact: it may happen. We may get to a point where the frontier-model companies are selling their incredibly important brains to nobody. Open-source models are not as good. None of that stuff is the same as what they’re doing.

Who’s solving the math problems? It’s Anthropic and OpenAI. This may change over time, but they have a huge advantage. It’s very clear at this point that they’ve hit some kind of recursive self-improvement, based on the fact that we’re getting models out quicker and quicker. Now they’ve reached a point where they’re scared because they’re ahead on the safety side.

If you read through all of these things that have been out there and then you look at the facts, we need things to change. If you’re listening to people and getting worried, it’s costing you money. Right now, the market’s not moving, but as we get through earnings and we have the next wave of this, at some point here, we’re going to go higher.

The Magnificent 7, if you have not seen, are near all-time highs. That’s an important read to me with consumer agents. I’ll get into that as I go through this. Demand is greater than supply, and it’s going to remain that way for a long time, in my opinion. I don’t see how we can catch up. It’s just physics.

We’ll solve problems with Einstein-level intelligence, but we still—you can come up with a solution and say, “This is a better battery,” but then you have to go build the better battery. It’s not that simple to get the parts.

This guy gets 172 million views. Obviously, this involves politics. Obviously, this involves geopolitics. There is no way for this thing to go this viral and for this to be there. We’re ahead of a midterm. So, if you guys are going to listen and get all caught up in this, it’s obvious that this is part of it because it is a major issue.

Whether this slows down Anthropic and OpenAI because they have to redo the risk side, I don’t really know. But I do know this: It just seems way too political at this point and way too much like, “We don’t want to be responsible for the liability of what may happen.”

Again, Geoffrey Hinton, the godfather of AI, has been talking about how dangerous AI is for a long time. We had 87 million views on this Matt Shumer one that said there would be no jobs by 2028. The Satranei[?] one by 2028, same thing: 28.7 million. “Oh, we have a cancer vaccine”—11 million.

People want to see the end of the world. That’s what sells, and that’s what gets the hype. And I said geopolitical: If the Russians, Chinese, Iranians, and anyone that’s at battle with the U.S. wants to break us politically, the easiest thing to do is continue to make data centers an issue, make AI an issue, and everything else.

Whatever the facts are, we had a cancer vaccine. I can’t think of anything more important, and yet we’re focused on things which are the same things that have gone on. This had 3 million views. This was June 6, 2023. This was less than 6 months after the launch of ChatGPT: “Why AI Will Save the World.”

Marc Andreessen had to write this because people were freaking out. So, this has been going on, guys, for a long time. China’s not buying Silicon Valley’s call for an AI slowdown. They would love an AI slowdown.

4. Oil, the Fed and rates: oil at $105 is not 2008's $150, this is not 2022, core CPI keeps falling, and the AI trade has no rate sensitivity. Mark Zandi's framing, Goldman's core PCE forecast, and the consumer bear market in restaurants, XRT, household durables and negative EPS revisions.

We also had oil prices go up. Scary. As I’ve talked about, oil is $105, guys. The next time someone tells you oil is going higher and they show you a chart, you tell them Jordi Visser was in China in 2008. Oil was $150. U.S. household net worth is up, I don’t know, 3 times where it is. We’ve added a hundred and something trillion dollars.

Gas as a percentage of disposable income for this country has never been lower. We are now a gas provider. Yes, some of the world is going to have problems with gas being higher, but at the same point, their net worth is higher as well. This is just completely stupid.

Does it mean headline inflation will be a little bit higher? Yeah, but it’s not 2022. “Stocks Fall, Fed Delivers Hawkish Rate Hike.” We just raised 25 basis points. Maybe we’ll do another 50 over the course there. I didn’t think we’d do one this year because CPI and core CPI have kept going lower, but they decided to do it because, for whatever reason, he was pushed into it. I have no idea. I really don’t care. They’re panicked about rates, whatever.

But here are the facts: This is not 2022. We don’t have 10% inflation playing from behind the curve. We have core CPI going down. As I’m going to show you with Goldman Sachs’ forecast and with everyone that knows, core PCE is going to go down because of things that are coming out of it.

The main point of this is that they raised 25 basis points. As much as I want people to think, as I saw a bunch of macro people getting paid money to deliver forecasts saying that this meant stocks would go down because stocks don’t like when rates go higher, I’m going to say this once, and I’m going to show you guys as things go on.

Before 2008, before 2009, we had recessions, and the reason we did is because rates really mattered. When rates went higher, the economy was based on corporate debt at companies that ran into really big problems and homeowners whose house prices went down because rates went higher.

Right now, we have tons of equity in the housing market, and everyone’s trapped because they locked in mortgages at lower rates. So, we don’t have the rate sensitivity there.

The second thing is, I don’t know if you guys know this, but Nvidia is $5 trillion. As I’ll show you later, the total sum of the consumer discretionary, household durables, retail, and restaurant companies, including some big ones like Home Depot and McDonald’s, but not including Amazon, is $1.5 trillion. Nvidia is $5.25 trillion.

The stock market represents what is driving growth. The Magnificent 7 are much bigger. They have no rate sensitivity whatsoever. The AI trade, despite what you’ll hear from bond people, has no rate sensitivity whatsoever in what’s going on.

Mark Zandi: “The economy is already growing near full potential.” True. “Operating at full employment.” True. “Inflation is too high to be sure, running above 3%, but much of that is the fallout from high energy prices and tariffs.” True. “AI-related infrastructure investment appears to be powering the economy.” True. “While the non-AI economy is already struggling.” True.

“To hit its inflation objective, the Fed either needs to rein in the AI boom”—true—“or put even more pressure on the rest of the economy.” Neither is a good outcome. This is the point I keep making. Your job is to make money. It is not to sit there and think about the past.

Remember Visser Labs: the V, the left side linear. This is where your economists are giving you BS. The Goldman Sachs PCE core inflation forecast—again, the effects of temporary factors, the portfolio-management effect, all these things. Everyone knows this. I just don’t understand why people don’t talk about it in terms of what’s going to happen.

Core CPI is down at 2.3%-plus. It’s barely above. Core PCE is higher for the reasons that they have here. The software and accessories component is due to this having a different weighting in the CPI thing. There’s a double-counting issue. Everyone already knows this, but we’re scaring people.

So, is the market responding? Is the bear market happening? Yes, it is. Here are the restaurants trading at 52-week lows. Look at this collapse in rates and oil. XRT. Now, this includes Amazon, but it’s equal weight, so Amazon is a small portion of it. Here we go: falling continuously.

Here are the household durables. This includes homebuilders and Home Depot. Look at it trading off. This is the Goldman Sachs low-income consumer basket, getting near the lows of the year, near 52-week lows.

Like I said before, here are the GICS Level 2 EPS revisions. Look at these massive numbers. The AI trade, communications, and energy have massive numbers. Here’s what you guys are looking at: these are the ones where revisions are going lower, which is all related to the consumer. Now, in media and entertainment, you have to strip out Google, Netflix, and Meta. Then what you’re left with on the revisions is negative. That’s the 11 trillion.

So when you go through this, the total sum of these is very, very tiny, and that’s why you have this. In 2022, when it was time to be bearish on stocks, people were telling you then—and writing now—that they’re raising rates, so you have to be bearish on stocks. You don’t have to be bearish on stocks. Here are revisions: massive negative in 2022. Again, people were forecasting a recession; it never happened, but equities did run into trouble. The hangover from 2021, the overhiring, and the overspending—that’s what ended up reining things in.

This time, it’s not happening now. Could we get GPU prices and chip prices to come down? Could we see bottlenecks that slow things down, and then you’re just left with the debt sitting out there? Of course, we don’t have any indication of that, like I’ve said. At this point, because AI is accelerating, I’ll take the other side on this.

5. Market resilience: since June 25th oil went from 67 to 92, 10-year rates from 4.30 to 5.02, tech momentum fell 48%, yet the S&P held and HYG/IEF made new highs. If crude falls $10 on a tweet, stocks are at new all-time highs.

Here’s 2025. The reason I wanted to bring this up again—I’ll show this later, too—is that these periods were during periods of tariff fears. This is the tariff fears, and this is 2022. These were recession fears. We don’t have that right now.

I put this out from a bearish week. You had a hot PPI the Thursday before. You had a hot CPI the Friday before. You had a 10% chance that humanity would die. You had a pause in AI. You had the Fed’s hawkish hike. Brent hit $110—I don’t know if it was the first time since March. Ten-year rates were at their highest level since 2007. That was the week, and for the week, the S&P futures finished down about 30 basis points. But the AI bear reading hit 53. I wanted to give you guys—well, I guess I’ll get into that later.

Here were the daily moves. This is in sigmas. For all of the negative stuff that’s been happening with oil and Iran, we can’t even get a 1-sigma move on depressed VIX. This is taking the implied volatility, which is the VIX, converting it into a daily move, and then taking the daily move. We’ve basically been going sideways.

This is a nice-looking reverse head-and-shoulders chart. Regardless of that, this is only a 5-day chart. This is when people were calling me. By the way, for people calling me and asking me where futures are going to open on Sunday night, start looking at SPY perps. Just go into Coinbase and start looking. You have to get used to 24/7 tokenized trading, guys. It’s time. You have to start getting ahead of the program here.

We opened down here, and we’re here. So you get this gap. Well, that cost you money for a while. Then the Fed came out—hawkish hike—so we sold off. The next day, we went right back up through it immediately, came down a little, and then continued going on.

Since June 25, that’s what those circles were that I wanted to highlight for you guys. Since June 25, oil prices have gone from $67 to $92. That’s $25. Since June 25, 10-year rates have gone from 4.30% to 5.02%. You guys are worried now? Look how far we’ve gone up. Two-year rates have gone from 4.15% to 4.73%. Oh, my gosh. Since June, tech momentum fell 48%. We’ve seen credit spreads widen, and CDS have widened on the hyperscalers.

Here’s the Magnificent 7. This was the Magnificent 7 on June 25. So if you’re worried about the AI trade, isn’t it going to have to come, to some degree, from these guys? It’s not going to come from Anthropic and OpenAI stock prices when we see their earnings. It could come through.

During true bear markets, when there are bearish conditions, the Magnificent 7 goes down. This is during the sell-off here. You need the Magnificent 7 going down, not about to make new all-time highs. Same thing down here. I bring this up again so you guys can start to go through it.

If you’re worried about the bond market, like I’ve said repeatedly for a while, you should have credit spreads widening and bond volatility spiking higher. We can barely get a move in bonds. Here is HYG relative to IEF, my proxy for credit, stripping out the yield differential. We made new highs for the last year in HYG relative to IEF.

And remember when this sold off here? What was this off of? This was on the private-credit side, so it was responding to that. This was a combination: you had Iran, you had all of this, but private credit started this move here. Right now, private credit—HYG—is going higher. The S&P—that’s where it was on June 25.

Despite all the rise in 10-year rates and all the rise in crude, I just want to tell you something: if it was able to do this with crude moving up that much, we’re now toward the end of the year. There is less than a 40% expectation of any kind of ceasefire with Iran. Now that we have a hawkish hike and rates are already at 5%, isn’t the risk for everyone who said to sell equities this week that, if crude falls $10 at some point for any reason—which a tweet could do—stocks will be at new all-time highs? Do you think stocks won’t be at new all-time highs if they can handle all the bad news?

6. Positioning and sentiment: 20% in AI (Marvell, Nvidia, Eli Lilly, AAOI, Fluence), adding to crypto. AAII bears at 53 with a sub-20 VIX has been positive 89% of the time. All Mag 7 above the 50-day has occurred only eight days this year.

The AI trade again. This is when I did my famous bailout of Micron, and this is when I did my June 4 midcycle slowdown. I’m still not increasing AI, and I’ve had a lot of people reach out. About 20% of my portfolio right now is in the AI trade. Back here, it was about 80%.

As I mentioned, crypto and silver now make up most of it, with crypto’s massive performance since I did the rotation. It’s much bigger. I keep adding to crypto. I’m not adding to the AI names at this point. I still have Marvell. I still have Nvidia as the 2 biggest. Eli Lilly, which I consider an AI name, is a big position, but that’s not the AI trade that you guys are looking for.

I still own Eos. I still own Fluence, even though it’s fallen about 80%. For those of you who are trading, if you’re picking one-off names in here, the reason I have a bunch of names is because I move my portfolio around based on the things that provide the biggest opportunity. Even though crypto’s had a big rally, if you ask me what has the best risk-reward for me over the next year, it’s still crypto, not the AI infrastructure trade.

That’s because crypto is an AI trade. It’s part of my thematic AI portfolio, but it’s not part of Agent 1, which I’m working on now. Again, this is about as bullish as it gets. The Magnificent 7 is up there, and they benefit tremendously from AI agents.

Here is the AI bear reading at the 53 level. This is a big level, except for this point. I mean, going back, there are only 2 times that we’ve hit this level in the last 4 years: during the complete chaos of 2022 and during the tariffs. That shows you how unusual it is. AI bearishness since 1990: 1,875 weeks. It only happens 4.5 times out of 100.

These are the forward performances for these periods when you combine this with a VIX lower than 20. Think about that: you have bearishness, but the cost of insurance is low. That seems like a good time to be looking at the long side, and it is up 89% of the time. It’s 10 times over the last whatever years; it doesn’t happen often. 53 basis points, 50.53%.

Now, to also get it above 50 in a structural bull market, defined by the S&P being above its 50-week moving average, that happens 1% of the time. Again, very good positive numbers with all 3 conditions. The VIX is low, we’re in a structural bull market, and bearishness is high. Only 9 times, and again, very solid numbers.

This doesn’t happen. As I wrote, this would be one of the more structurally positive things: to be in a bull market while people are getting bearish. That’s where we are now. For everyone saying to sell equities, I think you’re in a very dangerous position based on history, given that you’re joining the bear crowd, which is already there. You’re climbing the wall of worry. This is a classic definition of it.

Again, the Magnificent 7 were back there on June 25, when all the bad news was happening, and they’re near the highs of the quarter. The reason I brought up the Magnificent 7 is that there’s another thing going on. Very seldom this year have all 7 of them—all 7 of them—been above the 50-day moving average. That includes the hyperscalers, plus Apple, Nvidia, and Tesla. They haven’t all been above the 50-day moving average except for 8 days the entire year. These are the periods. The last time was June 4.

I view this as meaning that, if we’re going to get a big move higher in equities, one of the reasons would be because all of the Magnificent 7 are benefiting. If you add in Oracle, you get fewer days. The reason I care about Oracle is because it’s a very large stock and a hyperscaler.

I believe this bearish sentiment for the market began in October of last year, with the CDS blowing out on Oracle, and I don’t think it’s ever left people. I think the people who thought it was a bubble in October missed the entire rally in semiconductors, and now they’re still sitting there bearish because the AI trade has peaked. Even though it’s still up massively, Korea is still up 60% year to date. Taiwan is about to make new all-time highs. All of that has happened in 8 days.

To put that in context, this year, only 4.5% of the trading days had that occur. In 2023, 36% of the time, we were declining with the Magnificent 7 above the 50-day moving average. I think this is an important period to watch because they’re breaking out, and unlike the prior times, when Meta was collapsing, Microsoft was collapsing, and AI was going to destroy...

They're all sitting there, and I think that's because of AI agents. They all benefit from AI agents—all of them. And that's the point I want to get into.

The reason the economy is different, again, is that the economy is based on the top 10%. The bottom part of the economy gets transfer payments. So unless you're going to fire people—which is not happening—how do you get anything bad in the stock market if everyone has money to spend? And you're doing it because oil is going higher, which I'm telling you does not matter the way it used to. Gas prices as a percentage of disposable income remain near historic lows. If you do gas prices as a percentage of household wealth, it's an even bigger joke. Oil doesn't matter the way it used to.

7. The automated AI researcher: Jensen Huang and Brad Gerstner at the All-In Summit, Aschenbrenner's inflection arriving early, why the debt fears are misplaced, Musk's 10–20% vs. 80–90% framing, and the agent verticals: robotics (Tesla), consumer agents, digital labor, scientific discovery, autonomous commerce, longevity, data and cyber.

The AI boom is transforming America. We know that this is ChatGPT. This is what's going on. Give me a break. Jensen Huang spoke this week. So, for everyone who listened to the Anthropic kid, who listened to Dario Amodei, just go listen to Jensen. And at the All-In Summit, whenever you listen to him, especially at a time like this—and yes, President Trump called him while he was onstage—but this is really about what he said about everything happening today and the industrial revolution that's going on.

Don't lose focus on this because you're worried about people telling you it's time to be bearish. Brad Gerstner went through everything, and I thought he was very, very balanced in terms of the way that he approached this. It's worth listening to. It's worth going through.

So, the AI labs—Anthropic and OpenAI—this is really the critical thing, and he talks about it. I'm not giving you any of the important points from the transcript because I think you guys should just go watch those two and get this thing out of your mind. If you haven't brought this up, bring it up in an LLM and start doing so you can check some of the things I'm going to show you.

But this is really important. Whatever you think or know of Leopold Aschenbrenner, this paper came out in 2024, “Situational Awareness: The Decade Ahead.” The models are almost six months ahead. So this automated AI researcher point is really, really critical.

This is effectively one of the slides. I believe it's on page 71. But this is what happens, and the reason this is important is, don't think of it as AGI or ASI or anything like that. I'm going to take you through it, but this is the point: What's happening right now is exactly what this is—the hacking. That's the negative side of reaching this point: the Hugging Face episode and the misuse fears. We've reached this point, but there's a lot of good that comes from it, too.

We've already accomplished, based on his research—and these are the facts between the releases and the key things that have happened this year—a lot. It's really important for you to understand how fast this is moving and ignore the people who are telling you that there's too much debt.

I don't know how to get this out of people's minds. It is so insane to me that people are worried about investment-grade debt for the best balance sheets on the planet when they're sitting there watching that the United States of America can have $40 trillion of debt. Everyone keeps telling you—and Ray Dalio has tried to make a living post-hedge fund telling you how bad this is going to be because we have so much debt. Yes, Japan's been in this problem forever. AI is a government thing. There's no way we allow the whole thing to come down.

So this comes down to usage. And, like I said before, I don't know how to tell people: We're past humans adopting it. That's why this is working, and that's why what he's saying is so important.

When you get to the point of recursive self-improvement—the point where you've got autonomous researchers—that means companies can just say, as I showed with Astra last week, “Build this for me,” and it's built. We're at the point now where Salesforce.com can hand out products to people with agents. The agentic side is where adoption starts to accelerate, and it goes at the expense of expenses. It's a productivity boom, and that's what he talks about in the paper.

That's where we are right now: millions of AI researchers operating faster than humans. But it's all at this point here that was supposed to be in 2027, and that's happening now, in the early stages. This is the part here, and that means we're accelerating things. Nobody expected it to happen this fast.

So that's why an Anthropic researcher believes there's absolutely a 10% chance. In fact, Elon Musk gave an interview with Ted Cruz during the summer where he said there's maybe a 10% chance advanced AI leads to human extinction within the next 5 to 10 years. That would be the headline for the media. But then he says there's an 80% to 90% chance that it's extreme prosperity.

So if you're managing money and you worry about this, if there's no humanity, don't worry. Nobody's going to pay you, so there's no need to bet on it. The expected value of a 10% chance of the end of humanity and a 90% chance of abundance is basically all upside. You can sit there and play the game. You can listen to it. You can pretend that that's there.

I'm just telling you the demand side is from nonhumans. The adoption is happening from agents. This is why it's so important to crypto. As someone who set up a Robinhood blockchain wallet, it took me 4 hours. I use AI all day long, and there were so many things to get through. So I need an agent to do this for me quickly and go through the pain, speak the language, and not have to go look it up.

That's why we're entering a period where the models are coming out so fast that you don't know how to keep up. Which model do I get? Astra came out. Fable 5.1. That's because the models are coming too fast. That's an indication that the agentic world is here and humans can't handle the speed. We cannot handle the speed.

But luckily, the debt is at human speed. The demand is going to come from companies that have trillions of dollars of revenue and will easily be able to give trillions of dollars to the model companies. And there's no way that open source is as good as the model companies, so it'll be 90%.

If you want to use Jensen Huang's analogy, there's plenty of water that we use in our lives, and not a lot that we pay for, but we all pay for bottled water. So, again, you can go through this thing and try to figure out where the positive side of agent swarms forms.

And this is where I come in, hopefully in the same way I have the Agentic Infrastructure portfolio, which I'm saying will continue to outperform the S&P 500 by 20% to 40% a year. So there's alpha. All of those 100 names, in my opinion, will outperform. They'll go at different times. Not all of them will work. My job is to give you an index that works. If you try to pick one name, you're taking on risk in that one name.

The agent swarms—how do you get long? Here are all the things that come with that. We're at the time for robotics and physical systems, which means Tesla should be doing well. That's one to watch.

The consumer agents—this is where Microsoft, Salesforce.com, Meta, Apple, all of them, fit into the consumer-agent side. When you get into the digital-labor side, and when you get into this—this is my favorite one—the scientific-discovery side, which I'll go through, agent swarms solve unfinished IP. There's a business there.

Autonomous commerce—this is your crypto side. So, again, all of this stuff is necessary. The infrastructure is still there. I will create an index that will have all of these and all the names that I think benefit in there. That'll go on the subscriber wall sometime over the course of the next 2 weeks, but I am focused on crypto mainly.

So when you think about it this way and you go through it, agents are opening up all of this part here. We had the compute and power that led us to get the intelligence up to an IQ at Einstein's level, and now crypto is the main benefit. And because there is no investment there at all—because it is completely not understood—unlike the power and compute side, which got way too big way too quickly, this one's going to take years.

So go spend your time now on it. Go understand the things I'm talking about. Longevity—this is a phenomenal place to make money. And it fits with the unfinished IP and dormant knowledge. Any IP, any biotech company based on Navier–Stokes, is worth a lot of money.

Proprietary data all of a sudden becomes more valuable. FactSet, Thomson Reuters—because we have agents, all of the data becomes more valuable. Cyber becomes more valuable. Everything becomes more valuable. That's why you've seen this, and embodied AI starts to take off because they need the agentic side, too.

8. Agent swarms: Oppenheimer vs. Groves, Navier-Stokes solved by 10,000 agents in 88 hours, the revaluation of IP, Eli Lilly's TuneLab and biotech, Meta's Muse and Instinct's $2.5 billion raise. The macro chain is broken: earnings near 30% with zero hiring, and U.S. earnings breaking a 90-year channel.

Again, you can go through this on your own. We've reached this point. It's connecting his vision to when we reach here. This is the infrastructure side. This only happens when we have enough compute.

I want you to think this way as I go through this: There are 2 ways to organize intelligence. You just saw “Oppenheimer.” This is the easiest way to understand the importance of this. In the movie, Groves wanted to compartmentalize the different verticals and have them come in and bring their information into one little inside group. He wanted to coordinate scientific collaboration.

When you go read the Navier–Stokes side, this is what the agents did. Humans could have solved Navier–Stokes, in my opinion, if they had all worked together. But the Fields Medal, the Nobel Prize, and the way academia is set up—you've heard me speak negatively about my experience and my belief in what the school system does—it is the best way to get people trained for a particular major.

But that ended with the internet. And with the internet, we've gotten to a point where you have to be creative. It's only gotten more and more so. So Steve Jobs and all of these people who invented things and became the wealthiest people—go back and look at their jobs, their work situations.

It's not all of them, because then you started getting into the engineers. But again, having creativity, thinking outside the box, and using the other side of your brain—agents are basically going to be solving things, and they're going to be doing it just like Oppenheimer.

There were about 600 brainiacs at the Manhattan Project. There were a lot more people in general, but on the brainiac side, the Einstein side, Navier–Stokes was solved by 10,000 agents working for 88 hours. All problems will be solved. That's why I've compared this to that, and this is why the acceleration of disruption will increase.

AI-native businesses have a huge advantage. The democratization of intelligence is best used in organizations where they don't have the friction and difficulty of meetings, people, decision-makers, devil's advocates, and all of this stuff that goes on. I've been a part of that, and I work there.

So, traditional human research—here you go—all the friction that gets to an answer. The agent swarms can do something that Einstein couldn't: solve a problem that was there before Einstein. For everyone saying, “Oh, yeah, the guys who were working on it said that they must have stolen it,” why aren't they just happy that it was solved and that they were part of the process, if that's what happened?

Again, humans are already saying, “This is bad for science. This is bad for math.” Nobody's going to be happy with this. That is the Luddite story. That is every argument that comes with technology.

I've been spending time writing something on the revaluation of IP in an agentic world and how its value goes higher. This fits in with my Eli Lilly theme. Eli Lilly has tons of capital coming in. They are buying IP every day. They are also bringing biotech companies into TuneLab, so look up TuneLab and understand its value from the collaboration-in-the-swarm perspective.

I believe that biotech is part of the consumer-agent trade, that Eli Lilly is part of the consumer-agent trade, and that the way a biotech will trade is like it did with Moderna. As we get breakthroughs, they don't trade on a fundamental basis; they just gap higher. You don't need all of them to go higher to win. If you get a 10-bagger, a 20-bagger, and a 30-bagger—and I fully expect that we will see that on the biotech side over time—you win.

Muse was released. We are getting the first personal AI agent built for everyone. The reason this is important is that I want you to go back to when Mark Zuckerberg spoke at a town hall in June and said, “We are way behind on the AI side.” It was related to AI agents.

The latest viral AI assistant is rocketing across Silicon Valley. Buzzy startup Instinct is fundraising at a $2.5 billion valuation. That was on August 26, 2026. Go read about Instinct. Here's a YouTube video, “Instinct AI Is for Real.” This came out this week.

I'm not going to go through everything, but basically, all of these horrible things, like applying for visas—and yes, if you have an assistant at work, that assistant at work is doing them—all the personal-assistance stuff, the time savings, the money savings, coordinating with other people's agents, negotiating, reducing bills, and canceling subscriptions: all you've got to do is tell this thing. It will go out there and do it. It just needs the information.

AI agents, text-based AI assistants, are a hot market with Manus, WeChat, Alibaba, and Meta's Muse Agent, which launched last week, vying for users' attention. We're on the consumer-agent side, guys. I've been waiting for this. I didn't expect it to happen this fast, but when we get the Hugging Face incident, when we get the Navier–Stokes incident, and when we get the PauseAI incident, that means we're there. We are on the agentic side.

You're going to have a new economy. For everyone on the macro side, where it was, “Well, M2 growth, then we get nominal GDP that translates into some multiple and topline revenue. We get earnings to grow, and we hire people”—the missing link is that we're not hiring anyone. Zero people. The only people getting hired are in healthcare.

Earnings are booming at close to 30%. Margins are exploding, but we've broken this chain, and this chain is going to get further broken because we don't need the liquidity anymore. The liquidity is going to come from tokenization. It's going to come from the AI agents using and driving revenues, driving fees, driving activity, and the developers going and building there.

U.S. earnings are breaking out of a 90-year channel. This is not talking specifically about any kind of profitability related to productivity. On the upside, it's incredible when you consider the high-growth periods of the past that it's just broken out.

Software revenue per employee, from Warren Pies this week: again, a channel from 2000 to 2023. You don't even get a mega-spike from the 2007 iPhone. All of a sudden, the second we get AI—and that's the exact timeline that he has—you get a dramatic shift in software revenue per employee.

9. Crypto close: Clarity Act stalls, but 23-hour weekday trading arrives December 6th and the SEC rolls out a five-year exemption for tokenized stocks. $3–4 trillion of crypto vs. $900 trillion of fiat. Bitcoin is macro, the ecosystem is bottoms-up. This is a beginning bull market.

Let's finish up with crypto clarity. I didn't go through it. You had all the other bad news: oil prices higher, rates going higher, no debasement trade. For the last time, I will say this to both crypto people and macro: the time when crypto matters is when rates don't matter, because all innovations eventually hit a spot where they grow regardless of rates. That is the whole point of innovation.

When you're a new innovation, you're not being used, and it's all speculation, you should be correlated to liquidity. We are at the breaking point. For all the crypto crowd that's looking for liquidity, that's not what this is about.

The liquidity is going to come from the tokenization of dormant assets. It's going to come from the AI agents using and driving revenues, driving fees, driving activity, and the developers going and building there.

The CLARITY Act doesn't get through, but U.S. markets are set to expand to 23-hour weekday trading on December 6. That's just weekdays. Eventually, we'll be doing weekends. Perp can get you there.

Now, the U.S. securities regulator rolls out a 5-year exemption for tokenized stocks. Guys, this is why this is so important. So, here we go: if you build it, they will come. The infrastructure comes first. The economy begins when the people arrive.

Humans could never use crypto. But the ghost city—the old, famous Chinese ghost city—what have we been building, waiting for this to light up and be economically there? Right now, it's a complete ghost city. It is exactly like 2007 and 2008.

What were 2007 and 2008? That was the post-dot-com bubble. In 2007, it did change. Then we got the Great Financial Crisis. So, let's say 2004 and 2005: the infrastructure was built. The telecom companies blew up. That's what happened in crypto in 2022, with all of those entrepreneurs and the stuff that went on.

There was dumping of stuff. You've had companies bailing out. Everything has been an issue. But we built the infrastructure for this. It's all for AI agents.

When do you want to invest? You want to invest when the ghost city starts to fill up. When the ghost city fills up, particularly when there's no investment there, this will all be new stuff, and it will be taking away from the traditional finance world.

A $3 trillion market cap—let's say $4 trillion after today—for crypto, versus $900 trillion in the fiat system. There is no reason for wealthy people to invest over here without it being based on fundamentals. No need, speculation-wise. They have plenty of places to speculate.

The agents are there. The agents are going to be doing all of the business, and it will generate the flows. It's already starting to happen. So, you get crypto before, and then you get crypto after. This is where we are: the bridge that's being built is through tokenization and through all the pieces.

AI agents will need their own financial infrastructure. 100% correct.

Again, here's my index. It broke out to the highs of the year. There are 46 names. Four to 6 of them are public stocks, like Coinbase, Robinhood, and Circle. The rest of them are tokens, most of which the majority of you, if you're not crypto-oriented, have never heard of.

Here's Bitcoin. I will say it again for all the Bitcoin maxis: this needs to happen to move Bitcoin to the next level. You do not get to a gold-like market cap without the ecosystem, the usage, and the fundamental value.

Remember, Bitcoin is macro. The fundamental, bottoms-up stuff is what's driving things and what should happen. For all the macro people out there, think about how you've got bottoms-up and you've got macro. A lot of macro people I knew moved into bottoms-up eventually or bought into businesses that way. They work hand in hand because they're important, and that's where we are.

The verticals so far this month: digital monetary assets, which include Ethereum and Bitcoin, are up 27%. But there's 1 name in here that's driving most of it. The exchange and brokerage—the main thing is, I'm not going to read them all. You can read the verticals. They're all up month to date.

Again, I showed you this. Look at it on your own. Today, the index was up 8.4%, month to date up 14%, and year to date up 32%. Now, 43 of the 46 are above the 50-day. The 50-day is rising, with 37 out of 46 above it, and 27 of the 46 are at the 200-day. This is a beginning bull market, guys.

A couple of disclosures: I'm not here to lose you guys money. I'm not here to make you guys money. I'm here to educate you, and I'm here to help you. This is the beginning.

There will be a crypto video out this week. It’ll go out sometime during the week. I’ll probably do a more thorough market update, too, in terms of the AI thematic trade for the subscribers. And there will be a paper on Ghost Rails, which hopefully will go out either over the weekend or on Monday, whatever the case. It’s here, guys. I’ve talked about it for a while. The crypto side is there. I look forward to teaching all of you and doing a lot of crypto events. And for those of you who are going to be down in DC, reminder, I will be at Freedom Tech. Reach out to Mark Whaling. That’s m Whaling at22vresearch.com for information and to subscribe. And you can always go to visser-labs.com and get your information there. Have a good week, guys.