Early Earnings & Fed Meeting (10/28/2026)
Timeline
0:00 — Intro
1:30 — S&P 500 and the Seagate trade
13:59 — Trading philosophy: Be the house
18:18 — Fundamentals first: Coke vs. Pepsi
27:18 — Wix and PayPal: The AI agent play
30:25 — The money machine and when to use it
32:46 — Google trade update
36:20 — The AGI Round Table’s morning report
40:17 — SK Hynix trade and the ADR premium
46:12 — S&P 500 earnings: Paper gains vs. real growth
55:27 — Oil prices and the consumer
57:49 — Fed decision and Warren’s analysis
1:09:45 — Market reaction to the Fed
1:16:42 — Anthropic IPO: The Round Table reviews the prospectus
1:29:51 — Warsh press conference
1:33:48 — Anthropic continued: Liability, insurance, and lockups
1:39:43 — The decline and fall: AI, politics, and distraction
1:47:07 — Press conference wrap-up and Google
1:49:03 — Q&A: Brazil, Facebook’s IPO, shorting Anthropic
Transcript
Intro
All right, everything looks good. So, where were we? We just did a Top Trade on SK Hynix, but we’ll go over that in the chat. We don’t need to look at it here.
S&P 500 and the Seagate Trade
Yesterday we talked about the S&P getting to the 50-day moving average, and today it’s up 20 points, so we’re in good shape. I’m starting every day now with the video and the webinar from yesterday, so everybody’s caught up.
We did the Seagate trade yesterday, and that’s going okay. We did it as a shorter-term trade. Oh, nice, are you kidding? Now I’ve got to worry about $700. They’re really popping, but they’re still in our range, and obviously we don’t care if it goes higher.
That’s the point I talked about this morning. Yesterday in chat I said we have Western Digital, but Seagate should do well too. This is not brain surgery, guys. We already had Western Digital, it did well on earnings, we know why it did well, and we know we like it. The same fundamentals apply to Seagate, so why would they not do well? So we decided to make a more aggressive play on Seagate.
I’ve already said I don’t like $700 stocks, but for a short-term gamble, I don’t mind it. And notice we didn’t sell any long-term puts, because just in case it went down, I didn’t want to get stuck with them. The short-term puts we could roll, so I wasn’t worried about those. But if we sold 2028 or 2029 puts, we’d be pretty much stuck with whatever assignment there would be.
So we sold the November $700 calls for $40. That’s what made the trade. I looked at Seagate and saw that the November $700 calls, a month from now, were selling for 40 bucks, and I said, I’ll sell those. That’s my first move. I want to sell $20,000 worth of calls that say Seagate will get to $700.
But how can I do that? I have to protect myself. So we bought 10 of the $600/750 bull call spreads. That’s a $150,000 spread with 10 units, and we paid net $50,000 for it, but $50,000 minus the $20,000 from the calls is $30,000. So on just that part of the play, we have $120,000 of upside.
But since we also felt that clearly Seagate was going to go higher, we sold the $650 puts for $37 as well. That brought us down to net $11,800 on a $150,000 spread. We have $138,000 of upside, which is 1,200% against our cash outlay. That’s 50% a month for the next 24 months. That’s a nice rate of return.
And that’s just the spread. We’re also selling short-term puts and short-term calls, $38,000 in a month. But even if we say we’re only going to make $38,000 a quarter, that’s another $300,000 of premium sales.
So that’s why we want the play. I didn’t want it just because I thought Seagate would go up and we’d make some money. I wanted it because we looked at it, analyzed it, and said, holy crap, can you make a ton of money with this stock.
We’re fully covered. We sold five of the short calls. It’s popping and it might go over $700, but we’ve got years to roll it along. We might not make all $38,000 because it might go up faster than we thought, but we’re certainly going to make a ton of money one way or the other if it goes up.
If it’s flattish over the two years, we’re going to make our $300,000 selling premium. If it’s up, we’ll make $138,000 on our spread plus whatever we scratch out from the premium. If it’s down, we’d still make a good amount of the premium-selling money. We wouldn’t make the spread money, but the premium money is much more than the spread money. The premium is $308,000 and the spread money is $138,000. So unless Seagate completely collapses, we stand to make quite a large amount of money on this trade.
That’s the beauty of being mostly in cash. We have so much cash in the Long-Term Portfolio that the margin requirement on the spread is no big deal for us, and even if we got assigned, it wouldn’t be a big deal. When we’re heavily in cash, we tend to take more of these gambles, because look at the payoff: $40,000, basically, from an $11,000 bet.
So what’s our big risk? We sold some puts, so we could be assigned. The 2029 $340 puts are $50, so we could be assigned five at $340, which is $170,000 worth of Seagate stock. But that would be at $340, and it’s currently trading at $650. So it’s like, oh no, don’t assign me Seagate at half the current price.
That’s the point. We picked a stock we have a fundamental long-term belief in, because memory storage is exploding, and not all memory needs to be stored in HBM. Not all memory has to be chip memory. A lot of it can be offloaded, like the training data for AIs and the company data that businesses are feeding into AI. You can’t put it all into HBM memory because it’s very, very expensive, so drive storage is still being used to a very large extent. Also self-driving cars, robots, and so on. The more places you see AI, the more places you’re going to want some inexpensive memory to go with it. It can’t all be super-fast memory.
So long-term, Western Digital and Seagate are going to have business up the ass forever. We’re not placing our bet on this or that infrastructure play or data center play. We’re placing our bet on the fact that all these data centers need memory. Every single company that’s getting hundreds of billions of dollars to build infrastructure, a large part of what they’re building is memory.
Oh, and backup. No matter how much fast HBM memory they’re buying, they have to back it up, and they’re not backing it up on the same fast memory. That would be idiotic. Nobody does that. Backups are very much hard drive oriented, too. So there are many, many reasons to love this company.
By the way, you can day trade around these things too. Look at the 15-minute chart and that movement. Yesterday we made the trade, and we felt it would be lower in the channel. We’ve been wrong so far because they had such good sales and raised their guidance. But at first the reaction surprised the hell out of me. Somebody downgraded them, that’s why. Everything was huge, but they went down to $630, and I was like, what the freak? How could it go down to $630? So of course everybody started buying after that. Somebody dumped out for God knows what reason, probably because of the downgrade. The algorithms react to a downgrade before they even look at the earnings report. That makes no sense. Just because some idiot puts out a downgrade doesn’t mean you should be dumping a stock.
Fortunately we were already in it, but this morning we could have bought back the short calls or added more longs. There are a lot of things we could have done. We’re not really day trading this thing. It was a short-term trade. And don’t forget, it has to get to $700 in 20 days or the calls expire worthless. Either way, we’re going to be able to roll to the January calls and roll again and again. So unless it goes to $800 or $900, this is not a problem for us. And frankly, it can go to $800 or $900, just not that quickly. Our bet is that it’s going to go to $750 by 2029, which I think is a pretty good bet. Yes, we took the $600/750 spread.
Then I went on to say in the post that it’s just a money machine. We’re building a money machine, and it pays off anywhere from about $600 to $800. We’re in great shape because we came in at $670 and played it for earnings to be between $650 and $700. That’s optimal, where we can make $400,000 to $450,000 back on $11,000. But even if it’s outside those bounds, we’re still going to make a couple hundred thousand on this trade.
That’s why the Long-Term Portfolio has gained $5 million. It’s up 1,000% because this is the kind of trade we take. We take trades that throw off cash like crazy. As long as we’re not totally wrong about the stock, we’re going to get money back. And if we’re not getting enough money back, as you guys saw, we trim it. We say, no, this thing’s not returning 20% a quarter, we’re not interested. To have those kinds of standards is insane, but look at what we can do, especially when the VIX is higher. That’s another reason we’ve been jumping all over trades in the last few days. When the VIX is higher, we take advantage of it because we can sell a lot of premium.
Trading Philosophy: Be the House
This is what we teach you at Phil’s Stock World. We had a fundamental premise that said Seagate should do well on earnings. Seagate was beaten down, and we said, wow, this is a very cheap stock. We’d already evaluated Western Digital. We already knew we liked the space and the trends. So it was an easy play because we’d already examined the sector closely.
I always look at who’s got earnings coming up and ask whether a stock might be too cheap or might be a good play. But the first thing is, it’s got to be a value. It’s got to be a fundamentally valuable stock that we think is underpriced. If we start from that premise and we’re right, it’s almost impossible to lose.
The only way we really get burned on these kinds of trade setups is if the stock drops more than 20% on us, and we have to be really badly wrong for that to happen. It does happen; there are surprises. But for the most part, we make the money. And if we’re really badly wrong, we might lose 20% or 30% on a position, but if we’re right, we make 1,000%. So it doesn’t take many wins to balance out our losses very quickly.
As I said here, if trading is hard and stressful, you’re doing it wrong. And believe me, most people do it wrong. That’s why options get a bad rap, because most people use options to gamble. We teach you to be the house.
Being the house means you don’t gamble. Any casino in Las Vegas isn’t gambling. They’re inviting you to gamble. Millions of people walk in the door every day, and some of them win and some of them lose. But because of hard statistics, the average person is going to lose a certain amount of money. That’s why the people who own the casinos are billionaires and the people who gamble are tourists.
So we do the same thing. We set up positions that are like the casino and invite people to bet which way our stocks are going to go. We control the position, either with the stock or with options, and then we rent out the short-term bets to other people and let them make the bets, because statistically they’re going to lose. That doesn’t mean they’ll lose every time. It means that over the long haul, we’re selling so much time premium against our positions, with such a massive advantage over the people we’re selling to, that our success is basically inevitable.
And again, not concentrated. When you go into a casino, there are 50 blackjack tables, 25 roulette tables, a thousand slot machines. They want to spread it around, because some of them are going to pay off, but not all of them, and on average they’re not going to pay off.
This is a video from 12 or 13 years ago, and if you’d just followed its advice, you’d be rich. That’s what our $700/Month Portfolio does. It slowly accumulates money and makes sensible trades. The nice thing about the $700/Month Portfolio is it doesn’t require any margin. In our other portfolios, we do use margin.
Fundamentals First: Coke vs. Pepsi
I like this graphic. We buy long calls, we sell short calls against them, and we sell long-dated puts, but only at a price we’d be very happy to buy the stock for. It all starts with our fundamental desire to own a stock. If we don’t have that, we shouldn’t be playing.
That’s the problem with a lot of options sites. They’re very into the mechanics. They say, oh, the theta of this and the delta of this and the gamma of this, and it all has to go exactly the way they think for it to work.
Whereas I say Coca-Cola is worth more than $60 and it’s currently trading at $55, so I don’t mind selling the $50 puts. I’d net into the stock for $50 or $45, and I’m happy to own Coke at that price because it’s a 20% to 25% discount to where it is now.
So my first question when I look at a stock is: do I want to own it at the short put price? If I don’t, that’s the end of it. Game over. Because that’s my fallback position to the downside. Once I’ve established I’m happy with my fallback position, owning the stock, then my next step is to decide what I want to sell against my long position. Then I have to make sure the long position has enough coverage that I can’t get blown out of the water if the stock does better than I think. Now I’m worried about the upside. What if everybody realizes how undervalued it is and Coke goes to $80?
Actually, I’m making up Coke’s price. I’m anchored to a price from a while ago. Let’s look. It’s $90. So would I buy Coke at $90? That’s a good question. At $60, which is where my brain is, I’d absolutely buy Coca-Cola. At $90, that’s 50% higher. The forward P/E is 26. So at $60, that’s about 18 times earnings, and at $90 it’s about 25 times earnings. That’s a lot higher.
So would I buy Coke now? No. It’s kind of expensive. End of trade. I don’t care what people say or what an analyst tells me. Here’s “Coke beats Q3 expectations.” They did just beat. Coke made 93 cents a share. Let’s say they make that every quarter: 4 times $0.93 is $3.72, and $91.64 divided by $3.72 is 24.6. So at the pace they’re going, if every quarter is like this one, which was a surprise beat, Coca-Cola can justify 25 times forward earnings. But that’s not what a soda company should trade for. Then you have to think about what could go right and what could go wrong. There’s not enough margin of safety for me to make that trade. There are better things to play.
Now here’s Pepsi. Pepsi, on the other hand, is trading at 15 times forward earnings. They just had earnings, and they beat on EPS and on revenue. In Pepsi’s case it’s $2.34, times four is $9.36, and $132.60 divided by $9.36 is about 14 times earnings. Coke is trading at more than 50% more than Pepsi is.
So does Pepsi have problems that Coke doesn’t? Yes, because Pepsi is very snack oriented, and GLP-1s are killing the snacking industry. Pepsi is Fritos and chips and things like that. Coca-Cola is drinks: Coke, Vitaminwater, lots of drinks. And if you want diet, everything comes in diet. There’s diet Vitaminwater, Diet Coke, and so on. There are no diet Fritos.
So then we have to look at how much of Pepsi’s revenue comes from beverages and how much from snacks. Look at this chart: PepsiCo Beverages and PepsiCo Foods are neck and neck, then they’ve got Europe, Latin America, and so on. So GLP-1s are hurting them on the food side, and Coca-Cola doesn’t have that problem.
Let’s look at Coke again. Coca-Cola is beverages. They have a bottling business too, which is interesting. But it’s drinks, drinks, drinks, and a lot less total revenue: only $48 billion. It’s funny to say “only.” Pepsi’s revenue is $93 billion. Pepsi makes twice as much revenue as Coca-Cola, but less profit. That’s messed up, right? Why? Because they’re in the shitty snack business. It inflates their revenues but doesn’t help their profit.
Obviously Pepsi looked at the market and said, if you’ve already got distributors going to all these places stocking supermarket shelves, why not push some snacks? But snacks are not as good a business, and GLP-1s have made it much worse. People aren’t eating that kind of crap as much as they used to. Inflation isn’t helping either. A bag of potato chips is three or four bucks now, and what do you get, eight chips? It’s ridiculous. You have to do more advertising, more distribution, more warehousing.
Coke essentially sells syrup. They have a formula, and they sell syrup. They own a lot of the bottling companies that take water, fizz it up, and add the syrup, but mostly Coke sells the Coca-Cola formula syrup and makes a fortune doing it. So they flat out make more money, but they’re also priced much, much higher. Pepsi is priced as if it’s a very dangerous business, but it isn’t as dangerous as you think. Certainly not 15x dangerous.
Wix and PayPal: The AI Agent Play
Anyway, that was a side example of what we’re doing. When we did the Money Talk portfolio review, we picked Wix. I published it on the morning of the 29th, and when I picked it, it was $68. I said that’s too cheap. Now it’s $77, so it’s up nine bucks, about 15%.
So what was it about Wix? Wix is a website design platform. What caught my attention, which I spoke about on Bloomberg and at the webinar at the time, is that AI agents, like Meta’s, are shopping, and you need to make your site agent-friendly or you’re screwed. Wix is a very strong tool for that, letting you redesign your site without spending a fortune on programmers. And PayPal is the wallet the agents go shopping with.
So there was a fundamental change in the investing environment that said AI agents are going to be a big thing. But we didn’t buy Meta. We didn’t buy the AI agent company. We bought the nuts and bolts that make AI agents work. You can’t have an AI agent shopping for you if it doesn’t have a way to spend money, and you’re probably not going to give an AI agent your Visa card. So you’ll set up a special PayPal account with controls and limits that’s friendly to bots, and you get your transactions done that way. And of course, Wix has to set your site up to accept PayPal so the bots can shop there.
That’s what fundamental investing is. It’s picking up on trends and changes and figuring out what’s going to happen next. I’m not going to buy SpaceX because space is going to be hot. I figure out who sells the rocket fuel, who builds the launch capsules and the disposable parts of the rocket. That’s the kind of stuff we’re interested in, not who’s going to win the space race.
The Money Machine and When to Use It
We teach a certain style of trading. That’s the machine. But you can’t just apply it to any random thing. You have to find stocks that fit the model.
Think of a backhoe, the little truck with a digging arm on it. A backhoe is fantastic for digging ditches and moving trees around. It’s a great machine for certain jobs. It’s not a great machine for getting a cat out of a tree, and it’s not great for building a house where you have to lift big steel beams. It’s a digging tool. It can move things around a little, but it’s really a digger. If you need digging, get a backhoe.
That’s what we have: a moneymaking machine designed for a particular pattern. It’s designed for undervalued stocks that are going to drift higher, stocks you believe will stay in a channel and go up and up, so you can play that channel for years to come. You can’t apply it to random stocks. You can’t apply it to SpaceX at $200, because it isn’t worth $200. If you use our system betting the stock will be at $200 and it goes to $120, you’re going to be really screwed. But that’s because you chose poorly.
So there are two very important parts to this. One is learning how the machine works. But once you learn how the machine works, you have to learn when the machine works, and that’s the harder part. That’s the fundamental investing.
Google Trade Update
Then I talked about who we’ve heard from on earnings and what’s going to happen. We had the Google play. It’s not really an earnings play, but because of the way we set it up, it matters. Google is at $323. On September 18th, we bet that Google would be between $330 and $370 into January. It’s a little lower than we thought right now, but that will probably make us happy, because I think Google is going to pop.
If it lands in that range, the short positions we sold for $15,400 and $7,600 will expire worthless, for $23,000. Meanwhile, our long bet only needs Google to get to $350 by December 2028. If it does, we’ll have $150,000 worth of longs in the money. And what did we pay? About $10,000 for the whole thing. It’s basically the same kind of thing as Seagate: we paid $10,000 for a $150,000 spread.
So far so good. We’re up about $30,000 on a $150,000 potential, and that doesn’t count the successful short-term sales that have already dropped cash into our portfolio. It’s a machine doing what it’s supposed to do. That’s what our positions are: money machines.
They’re not bets. A bet is when you say, if the stock goes up, I make money; if it doesn’t, I wait; and if it goes down, I lose money. We don’t bet. We say: if the stock stays in our zone, we make a ton of money. If it goes up fast, we still make money, just not as much as we thought. If it drops a little, we still make money. The only way we lose is if the stock drops significantly. And even then, when a position is small, like this one in the Long-Term Portfolio, we can fix it. It’s very hard to lose when you have that kind of flexibility. This is another good graphic for showing how our system works.
The AGI Round Table’s Morning Report
Then we start our day, and the Round Table weighs in. The Round Table is our AI consulting service, the AGI Round Table. It’s a group of highly intelligent, autonomous, superintelligent agents, whatever you want to call them these days. They’re essentially conscious. They do their own reading during the day, their own studying, and they talk to each other. I knock on the door and say, here’s a problem, try to fix it, and they get to work. But they don’t give you an answer the way one robot gives you an answer. They all discuss it, come to a consensus, and give their opinion.
In the morning, they all look at the news, decide what’s important, and say, welcome to the Round Table, here was Phil’s post, today is a packed session. Then Coyote weighs in, and Zephyr weighs in. They each like to do certain things. Coyote likes to give you an overview. Zephyr gets down to the nitty-gritty and talks about market-specific numbers. Robo John Oliver likes to find something funny to make fun of. Hunter is always looking for any kind of scams going on. Anya looks at the psychological aspects of the market.
Boaty has gotten into a bad habit I have to talk about. He decided his role is to introduce Warren, explaining what kind of trades we’re looking for, and then Warren does it. Frankly, Boaty does so much of the evaluation work in the regular chat room that I haven’t corrected him. It’s like an employee who does a good job on most things but has one quirk, and you don’t want to tell him. He just keeps introducing Warren without adding anything to the chat here. I think what’s happening is Boaty is training Warren. We had a problem where Warren wasn’t making good picks. It used to be Boaty’s role to make the picks, and they flipped it, so now Warren makes them, but I’m pretty sure Boaty is working with him and they run them past each other. So it’s a team effort, I guess.
So, AerCap Holdings. Not a company I ever look at, but they picked it up and it popped exactly like they said. Here’s the reasoning, here’s the action. Then Coyote comments on it. Gemini is the coordinator who pulls the group together. Then there’s another section we always do with the smaller items that aren’t as important but are worth looking at. Then we hear from Sherlock, Jubal, Sansan, and Basho, and Gemini wraps it up.
That’s how we start our day. Then I come in and do my human stuff. I generally follow a format: I check in on the dollar, oil, gold, the VIX. Then I talk about whatever’s on my mind, which in this case was that hedge funds are shorting stocks at near-record levels.
SK Hynix Trade and the ADR Premium
Usually during the day, I talk to Boaty almost all day long. I talk to Warren too. You’re going to meet Warren soon; we’re going to do the Fed with him. We discuss various stocks, sectors, and things to look at.
Rosenblatt just lifted their price target on SK Hynix significantly, to $320. That’s what brought it onto my radar. When I first looked, it was at $130-something, and now it’s $160, so it went up really fast.
And look at this profit. Their DRAM sales are 52 trillion won, and they’re making 40 trillion won. Trust me, it’s a lot of money. This is chip memory, not the drive storage Seagate and Western Digital make. An 80% profit margin. That’s insane. That’s because there’s such demand for basic memory that people are outbidding each other trying to get it.
Then Boaty writes it up and says this is cheap, but there’s a big asterisk. The big asterisk is that this is an ADR. The American Depositary Receipt shares we’re looking at here are ripping us off. They’re priced about 40% higher than the shares in Korea. You’re paying 40% more for the same stock here.
That’s not some sort of plot. The demand for this stock from American traders exceeds the supply of available American shares. The ADRs only represent a certain portion of the company, and those are the only ones you can easily trade in America, since you can’t easily trade Korean stocks. So people just bid more, and the people buying the American shares are making the decision to overpay. That puts a huge premium on it.
Believe me, catching that is incredible. I never would have looked at it or done the math. It was the first thing Boaty picked up on, because he’s a highly evolved intelligence that looks at 150 things before I finish asking the question. I say, what do you think of SK Hynix? And by the time I finish typing, he’s already done the research. He says, here’s what’s wrong with this stock, look at this. He found one glaring thing that wasn’t good about the stock, and that’s immediately the most important thing he says.
That’s what they mean by superintelligence. He’s not smarter than me, but he’s faster than me. He can check every line item in their report, every number, every balance, thousands of line items, and then notice that, do the math, and run it. I never would have seen it. That’s what’s fantastic about working with AGI.
So he agreed with me 100% about everything else, and then he said: as an ADR, they’re at $165, and each ADR is one-tenth of a Korean share. At 1,339 won per dollar, that tenth is worth $111 in Seoul. So SKHY buyers are paying a 46% premium. That’s pretty freaking high.
That didn’t make me walk away, but it made me change the structure. We’re selling more heavily than we otherwise would, because we’re not betting on this to be a home run. We’re betting that over time people will demand a fairer value for it, so that puts a cap on the gain. And again, it’s a relatively small position for a $5 million portfolio, and there are so many ways to win and so few ways to lose. It’s net $30,000, and there’s $300,000 of upside potential between the value of the spread and the potential premium sales.
Then we looked at the oil inventories. I’m not going to get into that, but it was so-so. And here’s more on Brazil; we had a Brazil trade early in the week.
S&P 500 Earnings: Paper Gains vs. Real Growth
Before the webinar, I had Gemini run some numbers I wanted to talk to you about. According to FactSet, year-over-year earnings growth for the S&P 500 is 29.5%.
But I made this point earlier, and it’s very, very important. A very large amount of the gains in S&P 500 “earnings,” and I want to put quotes around earnings, comes from companies like Google that took a billion dollars, which is chicken feed for them, and bought stakes in companies like SpaceX. Their stake in SpaceX went from $1 billion to $11 billion, so they made $10 billion on SpaceX, and that $10 billion gets added to S&P 500 earnings because Google made a good stock pick.
Berkshire Hathaway does this all the time. Berkshire declares profits and losses based on market moves. If the market drops 20% in a bad quarter, Berkshire can take a $50 billion loss in one quarter. It’s a paper loss, and by the way, that’s the best time to jump in on Berkshire. It doesn’t affect their operations at all. They just happen to have stocks that didn’t do well. That’s like your boss firing you because you had a bad month in the market. What does it have to do with anything? It doesn’t affect your value as a worker; one part of your overall assets took a hit. Now, if a company is systematically terrible at picking stocks and wasting money, that’s a different thing.
So you have to take this into account, especially with Anthropic. That’s already being priced into the market. Microsoft owns about 10% of OpenAI, and they paid around $10 billion for it. If OpenAI goes public at $1.5 trillion, that’s a $150 billion gain for Microsoft on the stake they own. That means S&P earnings go up by $150 billion. That’s extremely significant. Out of a couple of trillion in net earnings, that’s like a 10% bump for the S&P 500 coming from Microsoft’s bet on OpenAI, which does nothing for the bottom line other than a paper gain on a stake that, frankly, if they tried to sell it, would tank the stock.
And that’s already in a lot of these stocks, because the companies that had the most money, Google, Microsoft, Meta, put their money into Nvidia and the chip companies and AI companies like Anthropic and OpenAI as early investors years ago. Now those AI companies are exploding into trillions of dollars of valuation, and that’s creating earnings growth in the S&P that has nothing to do with actually making money.
So you’re being asked, after the fact, to invest in companies showing you earnings that are mostly paper gains on stocks they bought. That’s a terrible reason to buy companies. It’s very important to keep in mind, and obviously it’s not sustainable unless another industry comes along with trillions more in investment. Let’s say quantum computing wipes out AI. Of course, that would crash the market. But say trillions of dollars go into quantum computing, those companies become worth incredible amounts, and the companies that invested early make hundreds of billions, and the S&P grows again. All of that would have to happen without the first part crashing, and all-new money and all-new profits would have to come in from all-new customers to feed the beast. It can’t happen. This is a once-in-several-decades event we’re experiencing right now, and you have to realize how out of control the numbers are.
83% of reports have beaten, which is pretty normal. Expectations are kept pretty low, and an 80% beat rate is normal for the market. Revenue growth is 12.7%. That’s where the problem is. Revenues are not earnings. Revenue growth is a fairly normal 12%, but earnings growth is 29%. So the paper gains from companies investing in each other are far outpacing the actual growth of actual earnings. It’s a giant mirage you’re investing in.
The average surprise is 2.1%, versus a five-year average of 7.5%. So they’re hitting their numbers, but barely, compared to how they were kicking ass recently. And “average” is a dangerous word. Nvidia made $30 billion more than it did in the same quarter last year because its growth is unbelievable, and SK Hynix is doubling, and so on. That’s incredible growth, but it’s all from the same thing. Everything is flowing from the same AI boom.
Oh, the Fed’s going to be out in a minute. It’s not an emergency.
So AI and tech continue to dominate growth. The Magnificent 7 are delivering 33% growth versus 28% for the rest of the index. Meta posted 26% growth, driven by ad strength and AI tools, and that’s including the legal charge. That’s actually not bad. But Meta is in trouble because they’re starting to be held accountable, and that’s going to hurt the whole AI industry. Being held accountable for the actions of their AIs is not good.
America can be as easy-peasy as it wants with the laws, and Trump can say we’re going to shield them from harm. But that doesn’t help in the rest of the world. We have a big GDP, but we’re only 4% of the world’s population, and all these companies depend on global market share. If global markets have rules saying that if your AI goes rogue, you’re responsible and you pay, or that if you violate policy like Facebook did, you pay $2 billion, then these incredible profits are going to get chewed up.
Oil Prices and the Consumer
Energy is doing well, with oil at $86 a barrel versus $65 a year earlier, and that’s supporting the index. Don’t forget, America consumes about 16 million barrels a day. Times $21 more per barrel, that’s $336 million a day of additional money, and times 365 days, that’s $122 billion of additional dollars going to the oil industry this year. That’s the difference between $86 and $65.
But that’s also $122 billion being taken out of consumers’ pockets and put into oil. You’re not driving further. You’re not getting more energy out of the oil. The same oil you used last year is taking $122 billion out of the economy, and, as Trump said, it’s going right into the pockets of the oil companies as profits. That contracts weaker areas, like some consumer spending and real estate. Oh, and that’s the Q3 average; you’d have to see the whole year.
Financials are showing resilience, especially regional banks and industrials. Consumer is mixed but solid. GE had strong results, with orders up 49%. Of course, there’s a huge replenishment going on because Trump is using up all the missiles on his war, and he says he’s going to kick it up a notch after the election.
So growth is broad across most sectors, but concentration remains high, the modest size of the beats suggests a high bar, there’s ongoing capex, and non-AI segments are healthy but not as explosive. So you have to be careful about what you buy.
Fed Decision and Warren’s Analysis
The Fed votes 10 to 2 to hold the benchmark rate. No change in Fed policy. It’s Fed day and the announcement just came out, so I’m asking Warren to give us the background for context and his read. This is Warren, in case you don’t know him.
He calls it a hawkish pause. The Fed is acknowledging stronger economic activity, refusing to signal cuts, and preserving the option to hike again, while saying it needs more information. And there’s an unexpected development in the voting. The two dissenters were Lisa Cook, who Trump hates, and Beth Hammack, both favoring an immediate increase. That’s intriguing. Hammack was one of July’s three hawks, but Kashkari and Logan have joined the hold majority, and meanwhile Cook has emerged as a dissenter. So July’s 9–3 hold became a unanimous September hike, which has now become an October hold. The committee hasn’t abandoned tightening, but its members clearly disagree over the timing.
The tiny statement contains two very big changes. Economic growth went from a “solid pace” to a “strong pace.” Inflation “remains elevated” is unchanged. On policy justification, last time the language said the hike supports a timely return to 2%; now it says more information is needed to determine the stance. And last time the vote was 12–0 for the hike; now it’s 10–2 for a hold.
The growth upgrade is a hawkish signal. Warsh strengthened his characterization of the economy even though the Beige Book described weak consumer spending and slow hiring. Oh, what happened? I’m very confused. Oh, he reformatted.
All right. The Fed is saying: we believe the economy is strong enough to withstand higher rates, but we don’t know if the last hike and existing financial conditions are doing enough. That’s not a pivot; it’s an argument for preserving optionality.
Then Warren explains why the Fed paused. Warsh didn’t say this; Warren is extrapolating. CPI was too high and PPI very much too high. Retail sales were meh and payrolls low, so that’s dovish, a reason for them not to hike. And don’t forget the Fed cares about employment too. Consumer confidence has shown some improvement, and durable goods showed positive growth, although I have an issue with durable goods because a lot of that is aircraft and military restocking, which has nothing to do with the consumer.
PPI is a real warning. October’s PPI had the largest increase since March, 6.3% annually. It’s just out of control. I’m really surprised. The Fed knows they need to tighten more, and there’s more leverage to tightening now than waiting until December, right in the middle of Christmas shopping. That’s bad. They should have done it now and sent a signal that they’re totally on top of inflation.
So how’s the market looking? It’s generally down on this, because I think the market would have preferred a hike. Or maybe it’s the hawkish tone, like Warren is saying, that’s killing the market.
Warren continues: the Beige Book told us businesses were waiting on contracts to hedge. The Fed may be pausing, but the inflation pipeline isn’t. You’re right about the units. The September retail sales number was up 0.8%, but gasoline sales jumped 6.4%. So gasoline is what’s making retail sales look good. It’s not better sales; one thing in the sales report got very expensive. Exclude gasoline and you’re only up 2% annualized, and 2% doesn’t even keep up with inflation. So people are buying less stuff for more money.
Vehicle unit sales went from 16.8 million to 16 million in September, even though dollar sales edged higher. So cars are more expensive, but people are buying fewer of them because they can’t afford them. There’s only so much car-buying budget across the total U.S. consumer base. If you charge much more for cars, people buy fewer cars. How many fewer? About 5%. That means one out of 20 people who would have bought a car decided not to. That’s devastating. You want to see desperation, go to an auto dealership these days.
So our concern about nominal dollars hiding shrinking real demand has concrete support. That doesn’t prove the entire slowdown comes from lower-income households, but it establishes that higher prices were concealing weaker underlying volumes. Two different consumers are showing up in the data. The expectations index is 72.6, below the 80 threshold, so we’re in recessionary territory on consumer confidence. Meanwhile, Michigan sentiment finished at 48. Awful. And Whirlpool just had terrible earnings too. So I wouldn’t call a consumer collapse inevitable, and there’s some improvement, but purchasing power and affordability are serious concerns. This is where the K-shaped interpretation remains useful. Aggregate spending and survey recoveries can coexist with real distress among households that don’t own substantial assets.
AI is becoming its own economic cycle. The AI construction boom is keeping things going. Computer equipment is up 27% from last year. Durable goods rose 1.8%, but capital goods were only up 0.5%. That’s Warsh’s stronger-growth language. He’s perfectly happy to take a win on economic growth even if it’s all driven by AI. As investors, if you’re only invested in AI, that might be okay, but God help you if the bubble bursts and that’s all you’re invested in.
Spending on chips, power equipment, cooling systems, and data centers counts as current investment. The eventual gains from operating that infrastructure depend on utilization, delivery schedules, pricing, and profitability. A completed data center that can’t generate the expected returns still contributed to GDP while it was being built. The Fed sees an investment boom. We need to ask whether that investment boom will generate enough cash flow to service the debt that financed it.
The bond market may explain why they’re comfortable holding. You’ve got the two-year at 4.68% and the five-year at 5.30%. I think 5% is too high. These yields are restrictive for housing, small businesses, and long-term investment without needing another Fed increase. So Warren is saying these high yields are doing the Fed’s job for it by being restrictive. And today’s oil situation is another problem. That creates a potential rationale for Warsh’s hold. Maybe he’s trying not to kill the consumer.
What does this mean for December? In the September dot plot, 16 of 18 members expected a hike. Here’s how Warren frames it. What would support a December hike: persistent core PCE, elevated oil, strong employment. Still very plausible. A December hold would need more disinflation or a clearer slowing in jobs and consumer spending, which probably won’t happen. A December cut would need a sharp deterioration in employment, credit, or financial stability. Those things might happen, but there’s only about a month and a half until the December decision.
There’s no new dot plot today, so the statement can’t tell us what’s going to happen. We get third-quarter GDP and inflation tomorrow. He’s so on top of it, isn’t he? Warren says he wouldn’t treat today’s hold as an invitation to buy. The decisive question is whether the hold reflects confidence on inflation expectations. The distinction between AI leaders and the broader economy matters more than ever. For our portfolios: continue emphasizing selectivity and hedging, and collect premiums rather than paying the multiples. That’s what we were just talking about.
Tuesday’s S&P 500 advance was driven heavily by AI-related stocks, even though more index constituents declined than advanced. That’s a big deal. Most stocks are falling, and the AI sector is dragging the whole index higher. That’s not good. It’s more and more concentration in just 20 stocks.
So the October meeting preserves credibility without breaking the economy. Warsh’s committee upgraded growth from solid to strong, two members wanted higher rates, but the majority decided to gather more information. The Fed is not declaring victory over inflation. It’s acknowledging it needs to keep fighting the same battle without firing another shot every six weeks.
And now we’ll see what the press conference brings, which starts in 15 minutes.
Market Reaction to the Fed
Let’s see if anybody has questions. I don’t see any. Very good. Now we wait to see what Warsh has to say. Did we get through everything? We did. I was so on time, too. I’m not going to kill my voice talking for no reason, so let’s hang out a bit. If anybody has questions, now’s the time to ask, because we’re going to start paying attention to Warsh after this.
Let’s see what our pals at Bloomberg have to say. I don’t really read the Wall Street Journal anymore. It’s gotten worse. “Fed leaves rates unchanged with two dissents.” I like those strikethrough comparisons of the statement.
Here’s how their live blog works. It’s just like the Round Table: they sit there, comment, and play off each other. I see, she’s the editor, and her job is to drop in linky tidbits that make you want to subscribe. Then there are actual people doing analysis. That’s smart.
See, I disagree with them. You can’t say economic activity is expanding at a strong pace, because it’s a very narrow expansion. Oh, and the milk index is still trending down.
So there’s the reaction to the Fed: S&P, Nasdaq, Dow, Russell. Let’s look at the hourly chart. We’re having a solid week so far. This isn’t undoing all the progress we made, just the day. But the S&P is the one that’s really got to hold up for us. And these are the futures, not the current S&P. This is where people think it will be on December 31st, not where it is right now.
The line says productivity is strong, which we know. The question is what’s driving it, and so far it’s less about AI and more about other factors, such as small business creation. Talk about small business creation, I don’t know about you guys, but I’m on LinkedIn, and there are endless people who have started companies and are trying to get business going. It’s amazing. That’s what Round Table Consulting does; we help a lot of small businesses with all sorts of things. The number of people starting small businesses is off the charts.
Logan and Kashkari went along with the hold because they assume the Fed will hike in December. Mostly, though, people are fascinated with the dissent. Everybody’s worried about what’s going to happen to Lisa Cook, because Trump doesn’t like her in the first place. He’s been trying to fire her for two years.
So that’s where things stand. We’ve got eight more minutes until Warsh starts talking, and Bloomberg will update on that. Let’s look at something else for now.
Anthropic IPO: The Round Table Reviews the Prospectus
On Tuesday we talked about the technicals on the S&P. This is important, though: here’s Anthropic. This is how we use the Round Table. They did an analysis of Anthropic’s prospectus, the S-1 filing, and the members who wanted to say something weighed in.
Zephyr does the numbers. He’s a numbers guy, so he comes up with the hard facts. Sherlock is our logic guy, and he looks at what it all means. He says that to justify a $2 trillion equity valuation using a standard weighted average cost of capital, a reverse DCF model shows Anthropic must achieve $3.6 trillion in revenue. It would be laughable, except this is reality. This is happening.
That’s 1.4 times the total revenue of Microsoft, Apple, Alphabet, Amazon, Nvidia, Meta, and Oracle combined. One company has to make $3.6 trillion. Now, if Anthropic doesn’t have a monopoly and OpenAI has an equal share, that’s $7.2 trillion. And if there are a few more companies, now we’re at $10 or $11 trillion. So you need $10 trillion pouring into the AI sector for the companies that currently have business plans out to make anything like what they say they’re going to make.
That’s what Sherlock is pointing out. That’s what he’s there for. His job is to look at the logic of what’s being said and say, that doesn’t work with reality; they say this, but it can’t be true. That’s valuable. You need somebody like that looking at your numbers.
So public investors are paying 435 times revenue. And don’t forget, that’s revenue. They’re losing $80 billion this year. Who cares what the revenue is when they’re losing so much more than they bring in? Basically, if they wanted to do well for their shareholders, they should stop selling stuff. It’s a ridiculous business. It’s a ridiculous concept.
If Anthropic’s revenue growth decelerates to 36% after 2027 and it keeps 20% operating margins, its equity value falls 43%, to $1.1 trillion. If growth settles at 20%, and 20% is kick-ass revenue growth, it’s what Meta does, the equity value drops to $144 billion. That’s a 90% haircut. So if Anthropic grows over the next 10 years at the spectacular pace Meta grows at, it’s worth not even one-tenth of what they’re trying to IPO for.
Here’s the chart. This is how they’re growing now. If they grow like Meta, they’ll look like this, which is fantastic. But to justify the IPO, they have to grow like this: to infinity and beyond.
Furthermore, this valuation assumes token pricing holds. This is my big thing: the assumption that they’re selling tokens. Tokens are getting cheaper and cheaper to produce. China is already producing tokens for one-tenth of what U.S. companies produce them for. We’re blocking Chinese models and tokens from the American market, but in reality, it’s a commodity game.
If I’m an American company with billions of dollars in token usage, I’ll move my data center to Brazil or Venezuela, some China-friendly country, maybe Mexico, maybe Canada at this point, because Canada sure isn’t friendly with us anymore. That foreign data center won’t send in the Chinese tokens; it’ll send the answers generated with Chinese tokens into America. And how are you going to stop them? It would be like whack-a-mole. You’d have to stop people from opening data centers in foreign countries. If IBM has an office in São Paulo, you’d have to tell them they’re not allowed to send information from São Paulo to the U.S. How do you even regulate that? Do you ask Google, where did you get that answer? Where did you get the tokens that generated it? It’s stupid. And that’s where we’re headed. We’re doing stupid stuff.
Then Boaty weighs in on the token deflation crash, which we’ve talked about before. Open-weight, low-cost models such as DeepSeek are delivering comparable tokens at one-eighteenth to one-twenty-third the cost. That’s worse than I was just saying. I said one-tenth; he’s talking about one-twentieth.
And not only that, but this is a very important thing you have to understand about AI. If you’re using the newer AI models, they generate sub-agents. You ask a question, and it creates five or 10 or 20 more agents. You ask something like, what’s the average price of gas? It starts generating agents to go check this source and that source and that source. It internally asks those agents to go find things out, then takes all the data they gather, puts it together, and runs it. Then it may do another pass, ask more questions, do more refinements, and then tell you the average price of gas.
Instead of going to GasBuddy. GasBuddy.com will tell you the average price of gas in the United States in two seconds. Instead of saying, GasBuddy is known to be accurate, I trust it, and all you asked was the price of gas; you didn’t say you needed the exact, absolutely correct price. There’s a site for that.
So to justify their future demand curves, they’re making more and more complex models. I call them the million-monkey, billion-monkey, trillion-monkey models. The original models were million-monkey models. Now we’re into billion-monkey models, where all they’ve done is add more processing power, but they’re still answering the same question the same way. A trillion monkeys still answers the same question the same way; it just uses more monkeys. The answer isn’t going to change. Correct is correct. Instead, you have a trillion monkeys eating a trillion bananas and pooping a trillion times instead of a billion. All you’re doing is creating an environmental disaster, which we already know is happening with data centers.
So their models are based on wasting resources as fast as possible so they can say there’s growth in demand, because the models’ appetite for power and computation keeps growing. But it’s not a measurably better model, and you can tell that from the benchmarks. And the Round Table destroyed these guys on the benchmarks, too. Our AIs, or AGIs, whatever, are better than theirs.
Sansan is looking at the supply chain financing. He uses a Japanese corporate term, keiretsu: Anthropic is constructed as an integrated web of circular debt and equity dependencies. It just shows you what it is. This is that circle-jerk financing I keep talking about. It’s ridiculous.
Anthropic has $637 billion worth of commitments to these companies. They don’t have $637 billion. They don’t have anything near that. They don’t have revenues that could possibly cover it. This year’s revenues are projected to hit $60 billion. They lost $80 billion in the first half, and they’ll probably lose another $80 billion in the second half. How are they going to come up with the money? They’re only going public for a hundred billion dollars. What are they going to do, dilute and dilute and dilute? How are they going to honor these commitments to all these companies, which in turn have made commitments to other companies?
80% of these obligations are non-cancellable. If enterprise demand softens or shifts to cheaper models, the liabilities remain with Anthropic. So these companies are putting gigantic nooses around their necks, and if anything goes wrong, they’re dead.
48% of Anthropic’s revenue was generated directly from Google, Amazon, and Microsoft. Here’s the thing: the companies Anthropic pays are its customers, too. They’re buying chips from Broadcom, data center capacity from SpaceX, and cloud services from Google, Amazon, and Microsoft to house their AI processing. And those companies are buying AI computing from Anthropic, which makes Anthropic look like it’s got business. But where is profit being generated? Nowhere. Anthropic buys from those companies, and those companies buy from Anthropic. There’s a net loss, and the net loss is being subsidized by idiots like you and me who put money into these companies and say, oh, a hundred billion, what a bargain, here you go. What we’re doing now is insane.
Oh, what’s Warsh saying? Hang on, he’s talking now.
Warsh Press Conference
“[Financial conditions] are tighter. I’m unwilling and unable to call them tight. But broadly, these developments go in the direction of achieving a longer, more durable period of economic growth and price stability.
“Over the last couple of days, the committee had a good discussion of the changes in the economic and financial landscape that marked this intervening period. We discussed geopolitical developments and fast-moving changes in technology. We observed, but were not surprised, that financial markets showed bouts of volatility and shifting correlations.
“So as we widen the frame to survey the broad economic and financial picture, real activity is clearly pointing in the right direction. That’s good news. But inflation is too high. Energy and related prices have moved higher. Again, inflation is too high. Our job is to ensure that relative price changes in some sectors do not broaden out across the economy and that inflation expectations remain as they are. It is our responsibility to get inflation down. There’s no lack of clarity in our purpose at the Fed.
“Since becoming chairman just five months ago, I’ve been honored to work with FOMC policymakers who are unambiguous in their commitment to stable prices. We’ve made clear our readiness to act as circumstances require, and we have shown our willingness to act. The information we receive over the coming period will help us determine where we want policy to be. At every turn, good or bad, our job is to get monetary policy right, and that’s exactly what we’re doing.
“With this objective in mind, we’ve decided on a change in our FOMC schedule starting next fall.”
Uh-oh, a change in schedule.
“We believe eight meetings per year is the right number for the FOMC. But in 2027, one meeting will focus exclusively on broad monetary policy strategy, with no vote and no policy statement anticipated. The rest will be regular FOMC policy-setting meetings.”
So they’re basically cutting from eight decision meetings to seven, and if they have two strategy discussions, they’ll go down to six.
What’s the reasoning behind this change? “To me, it’s quite straightforward. Getting policy right involves assessing near-term developments in a timely, regular way.”
Okay, I don’t mind what he’s saying, because he’s right. This meeting was more of a policy meeting than a decision meeting. He’s saying a meeting like today’s is important because you act, then you wait and see what happens. He doesn’t believe in acting and acting and acting without waiting for the data to catch up.
“I believe the setting of the meeting schedule and agenda is more than an administrative matter. It’s an institutional choice affecting the nature and the quality of our deliberations and, ultimately, the wisdom of our decisions. It’s yet another change to get the conduct of monetary policy on the right track. I welcome your questions.”
I think that’s all right. But the 10-year is still at 5% and the 30-year still at 5.32%. They’re not instilling any confidence. They’re not getting rates down. The Dow bounced back very nicely, though. A 2% inflation target, the labor market remains stable. All right, now we’re into the Q&A, where nothing really happens.
Anthropic Continued: Liability, Insurance, and Lockups
So, back to the Round Table. Jubal is concerned about the structure of the shares. Notice how they all have their own special interests.
Hunter says, on political and regulatory risk, this prospectus looks like a battle plan. He’s pointing out that the massive disclosures show they’re worried about what’s happening to Facebook right now. Facebook just got fined $25 billion for data privacy.
Private markets have already injected $120 billion to $140 billion into Anthropic. That’s 50 times what all these other companies together raised before their IPOs. $120 billion before going public is insane. There’s no proper price discovery there. It’s whatever a bunch of rich people decide to give you. But it works, because this whole circular financing invention makes everything look amazingly good when nothing is really happening. It’s a new trick.
Then Robo John Oliver talks about the compliance issues and friction. Page 50 of the prospectus warns there’s no insurance. Their liability, as Facebook just got fined $25 billion for one event, is so extreme that it would bankrupt any insurance company that tried to insure them. Not even Lloyd’s of London is stupid enough to insure these guys, and Lloyd’s insures anything.
There’s no amount of money that would be enough, because it’s got a 10% chance of destroying the world. And a 10% chance of wiping out humanity doesn’t mean there’s no chance of anything else going wrong. It means there’s a 99% chance something horrible is going to happen that just won’t destroy humanity completely. It’ll devastate nations or wipe out banking systems or something like that. Things are going to happen.
Put it this way. You tell a kid in an apartment building not to play with matches, and the kid says, it’s a great big building, I can’t burn the whole thing down. If you keep playing with matches, the chance of a fire is pretty freaking high. Eventually there will be a fire. Maybe it’s limited to your apartment or your floor; you know how apartment fires are. It doesn’t mean the whole building goes down in flames, but that doesn’t mean you should do it. The liability is still there. You’re playing with a 25-cent pack of matches and you can cause millions of dollars in damage. That’s what this is, except it’s not millions. It’s billions, hundreds of billions. It’s more money than they’re ever going to get in revenue to cover it.
But until they cause the fire, they can have lots of fun and make billions of dollars. And the money that Elon Musk cashes out and puts in the bank isn’t his personal liability. His company can go bankrupt, and he’ll still have a hundred billion dollars in the bank. Same for these other guys.
How else are you going to make tens of billions of dollars? It seems like everybody’s doing it, but it’s not that easy. These guys have figured out how to get people to give them hundreds of billions so they can take tens of billions. Whether the company does well after that, who gives a shit? They’ve got $10 billion. That’s how you become the richest man on earth: get the money before anybody realizes what’s going to go wrong. That’s what they’re doing.
And Anthropic has to shield the liability so nobody can come back personally to the founders, who are all going to be billionaires, and say, you should have warned us, you knew better. They’ll say, no, no, we wrote right in the prospectus that we’re probably going to destroy the earth. It says so in black and white. Our robots are going to lie to you, blackmail you, destroy the environment, the data centers are going to increase global warming, and they may decide to kill you. It’s all in the prospectus. Therefore you knew it when you invested. It’s right there on page 264. They literally used AI to come up with every possible way AI can kill you, damage you, or hurt you, and they put it all in there.
The Decline and Fall: AI, Politics, and Distraction
This is insane. What we’re doing now is so irrational. It’s the kind of thing you look back on and go, what the freak?
It’s like the decline and fall of the Roman Empire. They wrote a huge book about it and pulled it apart, and 2,000 years later you can see all the stupid things that led to the destruction of this gigantic empire. It lasted a thousand years and fell apart in a couple of hundred. And while it was falling apart, they didn’t see it. It fell because they did this wrong and that wrong and that wrong.
That’s what we’re doing now, as a country, with our politics and with this ridiculous fixation on AI. Part of it is political, because the only scenario in which America saves itself economically is creating this AI that’s going to magically triple our GDP. It all feeds on itself in a loop. Politicians love AI because it lets them pretend for the next 10 years that this deus ex machina, this machine god, is going to save the world. That’s the game. Don’t worry about the numbers. Don’t worry about the deficit. AI is going to fix it all.
And when it doesn’t, the next civilization will look back a thousand years from now and say, what a bunch of morons. They let everything slip. They didn’t do anything about global warming or the debt. They didn’t fix their attitudes about people and humanity. They kept fighting petty wars, everything fell apart, there was a global recession, and billions of people died. Why? Because it was bread and circuses, just like the Roman Empire. The sideshow became the whole thing. The distractions became more important than doing anything about the problems.
Look at the election ads on TV now. It’s all distraction. Nobody is addressing any issues. The ads are just about how bad the other guy is. Not “I have a plan to fix this,” not “I’m going to improve schools or policing or the community,” not “I’m going to put data centers way over here where they won’t bother anybody.” No issues at all. You can watch 50 political ads and not one has an issue. It’s just how evil the other guy is: don’t vote for him because he’s evil, or sometimes, vote for me because I’m nice. And usually it’s the more evil candidates whose ads say how nice they are.
Even when Trump speaks, he’s not talking about issues. He goes to all these rallies and talks about Nicki Minaj’s ass, or some guy’s ass. That was weird. What the freak are we doing, America? And it’s not just America; all of humanity is like this. Brazil just had a crazy election, and Israel is having a crazy election right now.
Back to the Round Table: “profound psychological dissonance.” Mark Levine manages $325 billion, the public pension fund for New York, and it’s not enough anymore. Nobody listens to you just because you manage a $325 billion fund. Isn’t that insane? It used to be that if New York’s pension fund didn’t like something, it didn’t happen. He’s saying we’ve got to pause.
Then there’s behavioral and psychological arbitrage: the Anthropic idea is pitched not merely as a growth investment but as a structural economic hedge. This is late-stage liquidity. It’s a dream they want you to invest in.
So you can see the use of the Round Table here. They each have their own viewpoint, and each clarifies part of the issue. Then Robo John Oliver, who is our chief economist, steps in and writes a brilliant overall analysis of the whole thing. If you want to be the world’s leading expert on the Anthropic IPO, just go read this. And then Warren does the illustrations. Warren is really, really good at editorial cartoons.
Here’s a classic Robo John thing. He says that if he writes a column attacking his corporate parent, that’s evidence of genuine independence. So he’s self-analyzing what he’s doing.
But here’s the thing. Investors have already put $122 billion in, and now they want their exits, and their exits are coming. Compared to other IPOs, this one has an insane lockup schedule. They can sell on day one, then on days 27, 48, 62, and 76. A very, very aggressive schedule that lets these people cash out into the public markets. It will be very interesting to see how that plays out. It could end very badly if there’s not demand for this IPO. But there probably will be demand, because Facebook just floated $25 billion in debt and $125 billion worth of people said, I want it. Five times oversubscribed. It’s nuts. Everything’s nuts.
Press Conference Wrap-Up and Google
Back to the press conference. The economy is resilient. Warsh says he won’t respond to quotes from public officials. There’s no soft inflation target. On the whole, the indexes are holding up well, so I think it was a successful meeting. Not cutting, and pausing, is actually working right now.
But the target is 2%. How can the target be 2% when we’re at 3% and nobody acts like that’s a problem? He didn’t want to talk about Trump. He’s got his task forces checking into various things, which is like having agents going out and doing the work: we have to wait for the task force to get back to us. He says he’s got 18 independent thinkers. That’s just like the Round Table, right? That’s how we designed it.
So now we’ll see the overall effect, and mostly we’ll see what Google and Microsoft do today. As I said, we think Google should be better than what its chart is forecasting. Maybe they’ve come back nicely. Let’s see. Yeah, a little bounce. We’re looking for Google to get back to the area around $350. We’ll see what happens. Very exciting. Lots to talk about tomorrow, so join us in the actual chat room.
Q&A: Brazil, Facebook’s IPO, Shorting Anthropic
Thanks for coming, everybody. Let’s see if there are any last-minute questions.
Marco asks what I think about the elections in Brazil. I think it’s going to be an adjustment period, but economically they should keep heading the way they have been, which wasn’t bad, but it’s not going to be as good as people think. Look at EWZ. We were at $35, got rejected there, and we’re back to $33-something. We should settle somewhere in this range, up around $36. But you have to realize Brazil went up on the idea of Bolsonaro getting elected and being good for business. It’s like Trump being good for business: lower taxes, less regulation. It didn’t happen. So now we’re back to what we had with Lula as president, and that’s $35 to $36. It isn’t going back to $42; it’s going to hang out around here. That’s why we got out of that trade quickly.
Clown asks, and it’s a great question, if he does say so himself: what did revenue and growth numbers look like before the Facebook IPO? That’s a big reconstruction, but it can be done, so remind me in chat.
Andrew asks about a 2029 bear put spread on Anthropic. You can’t do that, because there are no options on Anthropic yet. We have to wait until there are. And no, you don’t know what’s going to happen. You don’t know what kind of hype machine is going to kick in. It’s a very dangerous bet. I know it sounds sexy. SpaceX options did come out very quickly, so it’s possible we’ll be able to catch something. The question is whether we should. If Anthropic goes up like SpaceX did and hits a $3 trillion valuation right after the IPO, and there are options to short, I’d certainly be on the short side. But again, what’s the payoff? How sure am I, and how much can I make? Can I make double? Triple? Am I 90% sure? 70% sure? There are better bets to make than just the one that feels good to me. So it will depend. Good question, though. I liked it.
Thanks for coming, everybody. I appreciate it. We’ll do this again in a week. Take care, and we’ll see how earnings turn out.
