Welcome to the great economic crusade of 2026.

It’s been a while since we’ve spoken.

What happened?

This sell-off began with a hawkish Fed on June 16. June also saw the dawn of open-weight models, followed by a collapse in token pricing. This proved there was demand for intelligence that lagged the frontier.

As I’ve discussed in this thread, Anthropic was a victim of its move to pull forward token demand.Anthropic increased context windows and token output for the launch of Opus and Claude Code in mid-April. This created a significant inflection in revenue, which they capitalized on by raising $65b by the end of May.

Then token pricing peaked. After stories of AI-sticker-shock, companies stopped token maxxing and sought out cheaper intelligence. In comes Kimi July 19 and the price war was officially on. After all, the natural forces of capitalism will always find a way to reduce prices of a commodity in shortage.

Iran had nothing to do with the top in this market. After all, oil bottomed weeks later.

That is a very powerful, exogenous judgement of the tech sector.

Bulls will fairly note that the rental prices for Blackwells and A100s remained stable throughout this period. Helpful, no doubt. Granted, there is a fair probability that token prices are a leading indicator for compute rental demand. But that remains to be seen.

It should be beyond clear to anyone in the market that the Fed bottomed stocks with Warsh’s press conference July 29. Stocks were heavy on the day of the conference, but once the news had been digested, semiconductors rallied 20% from the July 29 level. At what cost?

Welcome to Bellum in mercatus: Trump and Bessent’s great war on markets.

Great ambition

The market you are trading should be seen within the context of American wartime central planning. Within this framework, market pricing is only free insofar as it has government utility or escapes its control. Anyone skeptical is fighting against the Treasury and the Federal Reserve. The market will be controlled so as to maximize execution of the Trump administration’s imperatives.

What are the Trump administration’s imperatives?

• Stable bond yields.

  Exemplified by the liquidity facility unleashed Wednesday. This facility tops out at $16B per quarter,. This took momentary control of the long-end. Yet, the long end is digesting $2T per year of borrowing. Altogether, Scott Bessent really didn’t do much of anything permanent. People are calling it QE. It’s really more akin to miniature, symbolic yield curve control at a level too low to make a permanent difference. What he did is show the market he cared about the long-end and is willing to start intervening. A great irony considering Bessent’s work for Soros. The market promptly showed that what he did was nowhere near enough: a vote that the borrowing costs of the United States and its deficit are still materially underpriced given the inflation outlook.

  

  

• Control over the oil supply.

  No need to explain the why nor why it can’t be controlled. The administration has kept crude prices under stable as refined products fall into shortage. It has been almost 3 weeks since we were 1-2 days away from a deal.

The SPR:

The US consumes about 20m barrels of oil equivalent per day. Don’t forget, the government warned more than a quarter of the SPR’s inventory is inaccessible.

Refined products are the shortage to worry about. Ukraine is targeting refineries every week. Russia is tapped out. Hormuz is closed. If this becomes a problem, it will occur within refined products: see diesel and gasoline prices

• The AI buildout:

  the administration sees this strategically and opportunistically. It knows the future of war, cybersecurity infrastructure and the like will be powered by an arsenal of AI super-weapons. It also knows that building a ton of data centers will juice blue collar jobs and GDP, the opportunistic lens Trump understands. Trump is a real estate developer. He thinks that the US should pay less in interest on its debt when it’s doing well. This is the way someone trying to borrow money to run a real estate business operates. A hilarious comparison for a country whose deficit is widening. But within that context, you can imagine he just loves the idea of building a big warehouse that employs construction jobs and adds to his economic data.

• Wealth effect:

  jawboning the stock market is, in the short term, the cheapest way to create money in the economy. Companies can raise money at higher valuations. Executives get paid. Stock options cash.

• Resource nationalism:

  I will not waste time explaining Trump’s insane view of trade deficits. However, there’s obvious sense to wanting control of your own supply chain. Trump and his cabinet are right to ensure the US has access to the commodities it needs in the event of war with China or a fight over Taiwan. Don’t believe me, believe this Wall Street Journal reporting on July 30, buried within a story about Musk’s businesses.

Musk called the story fake news. Do you believe him?

Subscribe now

Utilizing a framework

There are some relatively straightforward implications to this kind of market. First of all, it’s to not be too greedy on the short side. Because of bonds’ role in building America, those who sell them short are called ‘bond vigilantes.’ I’d argue it’s only a matter of time until our love of stocks leads to calling short-sellers ‘stock vigilantes.’

For now, it is the golden day of the stock market. Trump is getting everyone’s kids into it with accounts of their own and sees the market as his scoreboard.

But this market is not healthy. It can go on for a long time. But the hallmark of a healthy market is broad participation and value creation. There is no clearer signal of an unhealthy market than the amount of off-balance sheet activity, borrowing, circular financing and - now in AVGO and NVDA - leverage and collateralization.

When the stock market ‘trades AI,’ what is it trading?

You, the stock market enjoyer of 2026, are trading the hardware purchases. You are trading the picks and shovels. The leading indicator of those hardware purchase, however, is lab ARR.

Tracking lab ARR is exhausting. The labs are very vague with their disclosures, and have the advantage of being selective with what they want to disclose at all. There was a period within the last two weeks that Anthropic’s revenue run rate missed estimates, then OpenAI’s Q2 disappointed, and then OpenAI rebounded in July and initial August numbers were strong. All citing leaks and meetings. It’s tiresome. It will be great to have exact numbers instead of annualizing weeks of revenue as ARR, and using that to determine the trajectory of the world economy. For now, all you can do is wonder how far up this S-Curve you are.

• How does the growth trajectory of the labs as a whole look to you? What does the trajectory of the lab imply for the publicly-listed companies most reliant on it?

• How does the demand for frontier intelligence look to you? What moats do the labs have if demand for the great frontier isn’t as economically viable as demand for the 85th percentile? 60th? What is the market pricing in?

• What is the trajectory of the labs? If the labs are now fighting to lower token prices, can the demand for tokens and new users offset the price cuts?

  

Is it possible for that S-curve to flatten now without a stock market panic? As far as equities go, in my opinion, this is the single most important thing in the stock market. It should be your myopic focus, as John Zito so perfectly put.

When exactly the S-curve flattens is the single most important issue in the stock market. If it flattens in 8 years at 9x its current size, every ounce of skepticism is wrong, the world will look very different, neoclouds will each have gotten a head-start at becoming their own version of AWS and power demand is going to soar. But there are many ways this can break. Your variables are:

• token price

  • Token prices can collapse due to compute efficiency or reduced demand for frontier intelligence. It can rise if demand outstrips generation at the pace of new compute being spent up less energy costs, etc.

• user growth + enterprise growth + coding product growth +agentic workflow growth = total token demand

  • Higher token demand from user growth and more complex products

• financing

The strategic implication of financing the AI buildout should not be lost on anyone. The AI trade rebounded when the Fed decided it wasn’t going to stop inflation itself. It was going to let the bond market do it. So the Fed sacrificed the near-term value of the dollar to incentivize financing deployment into productive assets. This kept front end rates lower and sent the long end higher. Good news for banks: this steepens the curve and increases their profit margin. Look at the moves in the yield curve the day before FOMC (yellow) and after (green).

This only works in so far as the long end stays under control. In that sense the Fed took a risk that oil prices would remain under control or the Iranian conflict would be wrapped up quickly. A bad bet thus far. That we are talking about financing at all has a few implications.

Bear in mind: The long end is the risk-free rate of duration and a traditional bellweather of asset prices and risk appetite.

The 2023 ChatGPT market rally was strong because hardware-demand growth was fueled by cash-rich hyperscalers. They could spend cash on hardware or utilize FCF. That is not the case today, well summarized in this infographic from the Journal:

I do not think it is a reach to say that hyperscaler balance sheet looking this way tells you that the easy days are over. Remember this?

Well, that’s the hyperscalers, what about compute production?

AVGO joined NVDA in taking the 2nd pill today.

To the AI complex’s credit, NVDA was initially in talks to guarantee $250b for the OpenAI datacenter. NVDA later committed to only $105b. For the rest, NVDA rightfully decided to just find money elsewhere.

Impressive. But why? Why keep talking about backstops and guaranteeing money for these labs? OpenAI has wanted a government backstop for almost a year. What does that tell you about their ability to get financing when counterparties see their books?

The other clear imperative is that if this buildout fails, it will take down an ever-larger part of the market with it. Which, of course, is supportive of the national imperative.

I believe this entire sector is so circular because value is not being created by the end user of all of this hardware, yet, if ever. The whole point is to backstop these labs and let them have time. But the issue becomes that and Felix have discussed quite a bit: securities are being sold. Liquidity is being taken by the complex, not provided. We are in a world where hyperscalers are selling debt instead of buying back stock. In Google’s case, they’re even selling stock

This leaves me with a bearish view intact. The easy part of the game is over. Unless the Fed is willing to abandon its principles entirely and make the dollar worthless, or the Iran war comes to a close, everything continues to get harder from here. Liquidity, financing, lab growth, token pricing, compute efficiency, and especially, public market scrutiny. Nevertheless, even with a ‘$500b Wall Street Facility,’ I don’t think things can sustain as they are for long until the labs can show that they can survive the threat of open source, token price deflation and financing costs. Important as we now know Anthropic will try and go public before the end of the month.

This is all bearish and speaks to my mid-long term view. But you really cannot say what happens tomorrow without risk of it being proven wrong immediately. You can, however, produce a framework.

I think through this market value acceleration with this overtly simple equation:

Consider this carefully as it is your roadmap for the market’s ability to accelerate or escape the administration’s Bellum in Mercatus control-initiative. Either the behavior of the labs’ end market, the labs themselves, the hardware prodcuers or the financing environment will show you where this change occurs.

Bond yields and oil prices are the greatest political challenge to the Bellum in mercatus plan. They work in synchrony and destroy the budget, Trump’s re-election chances cost of living, housing availability, and the business environment.

Trump has to decide between effective policy in Iran and the high interest rates. He is constantly weighing this against his desire for lower rates, a healthy market and a strong re-election chance. It is my opinion that the longer he tries to do both, the more it hurts him. An impossible quagmire now as munitions run low:

This is why Trump has tried to muscle data centers onto the voter base. A natural desire, but again, something about the American voter he does not understand.

I’ve long maintained that going public with an all-out push for data centers was a material political weakness for Trump.

Public opinion is a powerful thing. I believe it is public opinion that got the Fed to hike in 2022. Where is public opinion now?

But the war for data centers is a necessary fight in Trump’s Bellum in Mercatus. The administration believes it is a great way to grow the economy out of the national debt. Nevermind the obvious possibility that if this plan fails and there is a recession, then debt/GDP gets worse, the deficit is already blown out and government will have to battle a solvency question while trying to stimulate out of contraction.

In a stock market centered around two companies that are not yet creating value, something has to give. Oil prices, yields and elections are the only things that can cause this market to escape from the test environment that Trump, Bessent, tweets, TV apperances, Fed governor appointments and overhauls of the BLS have created.

The Anthropic IPO represents the first test of public willingness to underwrite the spending and business model with full information. It will represent a liquidity test as well. The path forward in the elections is a serious check on Trump’s power:

Although the market looks all the more under the Trump administration’s control every day, there are a few independent variables left. Both the AI labs and the policymakers trying to grow their way out of their debt, spending more than their balance sheet can accommodate and trying to preserve low rates, high equity prices and geopolitical control. That combination is an impossible triad until the war is over or the labs, hyperscalers, and neoclouds can achieve their ambitions and finance themselves.

A disclaimer about everything that I write, say and share:

Nothing you will ever read from me is investment advice.

I write about where I think the probabilities lie, and I hope my input in this debate has been of use to you. Invest at your own risk. I never sell stocks short, I never sell options short, and I would never recommend that you do either.

Nothing shared in my posts, comments, messages, or any other content on this platform constitutes investment, financial, legal, tax, or professional advice of any kind. All statements, opinions, analyses, strategies, or recommendations expressed are solely my personal views based on my own research, experience, and decisions. These are not recommendations to buy, sell, hold, or engage in any investment, trade, or financial action.

My content is for informational and entertainment purposes only. You should not replicate, copy, or act upon any information, ideas, or decisions I share. Any actions you take based on this content are entirely at your own risk. By reading, subscribing to, following, sharing, or otherwise engaging with my content, you explicitly acknowledge and accept full responsibility for all outcomes, including any financial losses, damages, or other consequences that may result. You agree to waive any claims against me related to your use of this information. Always consult with a qualified financial advisor, accountant, or other licensed professional before making any investment or financial decisions. Past performance is not indicative of future results. Investments involve substantial risk of loss.

Bear this in mind always, taking advice from anyone.

It’s been clear that the scale and returns of the AI buildout are hitting a critical mass. Not necessarily a stop sign, but a size that is going to make expansion difficult. We’ve discussed the amount of free cash flow required to continue to accelerate AI buildout.

In the past few notes, we focused on a few key topics:

• Token prices were falling and the market had Jevons Paradox wrong. Market expectations for Jevons Paradox did not account for the differential between efficiency-driven demand for intelligence and how much frontier-lab demand it would cannibalize. We reasoned Kimi and the efficiency-gaining process would not a) expand the compute TAM or b) be a net positive for the market given how much spending so much of the market expects from Anthropic and OpenAI.

• The Iran war was deteriorating and it was unlikely Trump would be able to get a workable solution that would keep the liquidity picture healthy. This, combined with the massive liquidity demands of AI capex and mega IPOs, could present a headwind in the supply/demand balance of equity prices.

• When the market hit a point at which bears would get eviscerated, it was surprising that it was still so heavy, perhaps indicating more selling was coming. This was especially true because of the unprecedented moves we saw in momentum stocks, implying a high risk of major overexposure that can unwind quickly. We noted it was likely that memory names could bounce quite aggressively yet NVDA still had quite a bit of downside in both the bull case and bear case. This came to fruition and was the short-lived bounce we discussed, which we wrote as a fair probability on the way to a collision course with 7000SPX

• That the net balance of the above made the setup asymmetric against the bull case for a market barely off of its all-time high, akin to hitting on 15 in blackjack, which has a 58% probability of busting.

This all assumed a competent, cooperative Fed and bond market. I was optimistic Kevin Warsh would bring a new day at the Fed: Sobriety would be a welcome fix and we could truly tackle inflation for the first time since the decade began. Thus far, I was wrong in cause and right in effect.

Warsh, as of now, has decided to let the bond market do the damage he did not want to do himself. He, and the administration, will see for once and for all that in an inflation crisis the Fed must either take control or the bond market will do the damage itself. Today, the selling in the long end told the Fed it was losing control, selling off and in effect, raising interest rates.

Look at the significance of where this is occurring:

The implications of this are ironic: by not taking control of the inflation picture, the Fed signed itself up for a tantrum instead. A dysfunctional and bearish bond market will continue to make rates more restrictive, slowing business activity in an economy driven by a cash-flow-negative business model that is going to need a lot of money.

Look at what’s happening to credit spreads. Credit spreads represent the risk premium that corporations pay to borrow vs. what it costs the government to borrow (given they are not ‘risk-free’ like government bonds). Operate under the basic assumption that corporations have to borrow at the prices set on the US yield curve + some premium. The premium makes borrowing more expensive. When things go risk-off, that premium exacerbates. Look at what’s happening to credit spreads.

Look at the inverse correlation between the Nasdaq 100 and credit spread blowouts.

We know AI capex is extremely cash flow negative. That kind of money can come from two places: it can be borrowed or stock can be sold.

The price to borrow went up today, and the equity prices reflected that reality.

I think Brent put it perfectly with this tweet. There is enough oil. There is not enough gasoline and diesel. Those are the consumer prices in CPI. That is your risk, plus many other pieces of embedded inflation caused by AI capex.

Look again at these prices in gasoline futures, diesel prices and unleaded prices.

I do not see why any of these things would improve into the fall. Iran knows it can continue to try and outlast Trump into midterms. It is also cognizant of the depletion of his already thin Strategic Petroleum Reserve. More than a quarter of the SPR’s inventory is unavailable for drawdown due to outages.

I’ve long read ’s work. His writing has always done a great job of monitoring the credit market. Forgive me if I’m paraphrasing incorrectly, but he has a good way of saying that you should always pay attention to credit spreads as it is the key caution flag for equities.

I say this to say it is the unequivocal time to be careful. Perhaps Trump capitulates to save the economy. Perhaps the US achieves its goals, Hormuz is open and the gasoline supply is fixed. Perhaps the labs really do ramp revenue like Semianalysis expects. But the conditions required for each of these things is becoming more onerous each day.

I’ve spent the few weeks talking about MSFT. They had terrific earnings today. Rational, cash flow yielding names are in vogue. Everyone knows how strong Apple has been. Perhaps the market can be held up by the great dispersion of healthy businesses in the economy. But you’re dealing with serious issues in the piping of the system now.

Look for example, at the constituents of QQQ, SPY, DIA or XLI. Check out how much of it is AI capex. I noted this months ago with CAT. Whether it’s the compute, the data center construction, the turbines or the financing, the entire market is set up for one trade. Look at how much each % Net figure is in NVDA alone!

As I’ve said many times this month, this is the time in markets where indexes can be up or down 3% for no good reason. It is a time where you can get really, really hurt being short-sighted. Be careful. There are red flags going up now that are beyond overvalued semiconductor companies. Consider, for example, how much customer risk NVDA has in its neocloud customers, or its megacap customers that investors are begging to cut capex.

We’ll see. But in my opinion, the downside risk to the economy has grown dramatically in the last few days. I say this to say: be careful. If the market soars in the next few days, I promise you there will be other great ways to make money in time. But it’s worth considering that the parabolic moves in these assets are going through a topping cycle and not a correction. That should be in the back of your head and influence your thinking on risk/reward.

Look at the size of this move. Think about how many buyers are trapped in those levels above. Think about how much overhead supply that creates in any attempt to bounce. Think about the market working off of the base assumption that there will be multiple $1T labs and nearly infinite demand for data centers. This is what is on the chessboard now.

Again, there is no way for me to say something now in the middle of vicious down-moves without high risk of being wrong in days or hours. But it’s an important moment to zoom out given what’s starting to see pressure. We all take our own risk. I’m imploring you to be careful.

A disclaimer about everything that I write, say and share:

Nothing you will ever read from me is investment advice.

I write about where I think the probabilities lie, and I hope my input in this debate has been of use to you. Invest at your own risk. I never sell stocks short, I never sell options short, and I would never recommend that you do either.

Nothing shared in my posts, comments, messages, or any other content on this platform constitutes investment, financial, legal, tax, or professional advice of any kind. All statements, opinions, analyses, strategies, or recommendations expressed are solely my personal views based on my own research, experience, and decisions. These are not recommendations to buy, sell, hold, or engage in any investment, trade, or financial action.

My content is for informational and entertainment purposes only. You should not replicate, copy, or act upon any information, ideas, or decisions I share. Any actions you take based on this content are entirely at your own risk. By reading, subscribing to, following, sharing, or otherwise engaging with my content, you explicitly acknowledge and accept full responsibility for all outcomes, including any financial losses, damages, or other consequences that may result. You agree to waive any claims against me related to your use of this information. Always consult with a qualified financial advisor, accountant, or other licensed professional before making any investment or financial decisions. Past performance is not indicative of future results. Investments involve substantial risk of loss.

Bear this in mind always, taking advice from anyone.

I think it’s going to be fun looking back on this moment in time. Stocks, and the president’s encouragement of them, have been undefeated for about 15 months in a row. Does it feel familiar to you?

With that in mind, a few days ago I suggested that it’s a good time to be cautious, unless you think the president can TACO and get Iran to play along. That was met with vitriol that I did not think was possible. Some took it quite personally; remarkable given I hardly could have said something less personal.

Vestiges of Q2’s momentum blowout, no doubt.

Markets slipped into a more precarious position this week. No one will argue we’re anywhere near as overheated as we were weeks ago. Where does that leave us now?

I will always attempt to waste as little words as possible:

It will be hard for POTUS to tweet his way out of this and upside is difficult given what assets have priced in. He will try like hell. But you can feel core tenants of the bull market starting to wane.

Look at the news in the last 18 hours. Which way do you think this is going? Is it clear at all? Sure doesn’t look like it to me.

So how do you play it? First, be willing to identify when it’s not clear.

Many of you will already be familiar with this framework:

1. Perception (Vibes): Think back to the ‘Google is an AI loser’ to ‘Google is going to win AI’ to ‘What the hell are they doing over there’ perception at Google. Or software. Or the arms race between Anthropic (this year’s winner) and OpenAI (last year’s winner).

2. Positioning: are asset managers over or underweight an asset? To what extent does that leave it vulnerable to a move in either direction? Who owns an asset and on what basis do they own it? What are they expecting? Do you think new buyers or sellers are coming and why?

3. Liquidity: How liquid is the asset? Can the float support the buying pressure it has? Think about a stock like ARM, which Softbank holds 86% of. Are there so little shares in the float that it can actually exacerbate a move to the upside? Or is dilution coming? In the case of SpaceX’s $75B IPO, is there just too much supply to outweigh the buying?

4. Gamma: How are asset managers and dealers positioned in the options market? To what extent is it notable or significant?

Assess where each of the above falls in the main components of our market: indicies, semis, oil, rates, etc.

I don’t think it seems clear. If nothing else, it seems now more than ever that we are in the middle or 3 binaries: rising or falling AI capex, more hawkish or dovish Fed and inflation picture, and more or less international flows of oil. Let’s think through each of these questions.

has said many times that it takes two to TACO. This is true. But this is not a two-front war anymore. Coalitions are set and there are now 3 frontlines: Hormuz, Bab-el-Mandeb and Ukraine.

Do I think that all of these conflicts can end quickly? Perhaps. But on what basis? The US and Israel first struck Iran in February and since then, we’ve transitioned to a more dangerous balance. Where we were once alternating between talks and threats, we have now only alternated between attacks and ceasefires. Iran knows it cannot outwardly extort the global economy (only the semiconductor companies can do that), and will play the game to its advantage:

But I cannot shake the feeling that the balance is deteriorating off of an extraordinarily-optimistic base. The war in Iran is not smaller, it’s larger. Iran is now controlled by its hardliners and the US seems unwilling to draw the war to a decisive close or accept Iranian terms.

This, of course, has direct implications on the inflation picture. The inflation picture and the Fed are not more constructive. They’re more hawkish. Until and unless Iran and Bab-el-Mandeb can be sorted, the Fed is unlikely to have any reason to cause the liquidity picture to impove.

Which brings us to AI, today’s main use of liquidity. The market was happy to support the AI-buildout when the liquidity picture was more constructive and the numbers weren’t as large as they are now. That has changed. The AI-buildout is now so large that its beginning to cause noticeable moves in hyperscaler credit default swaps.

I am not here to tell you that anything is going to go belly-up tomorrow or that this is World War 3. I am actually not here to tell you that I know where any of this is going. But I want to talk about what I think is becoming a death by a thousand cuts market. Especially because this consolidation/down-move was never caused by crude oil prices in the first place:

That sounds pretty dramatic

It’s not. Let’s talk about what that means. Any time you are short, the administration is working against you. Trump’s military does everything it can to only conduct strikes outside market hours. When the market is down, it smothers the media with Greer, Bessent, Hassett and Truth Social posts. The White House wants this war to hit American pockets as little as possible, and you can’t blame them for that. Plus, the president is always long the market, but this president is especially so.

So it bears repeating that if you are going to take a skeptical view of the stock market, it will be as painful a process as the powers at be can make it. They are going to push the dove headlines, they’re going to change things at the government agencies that calculate inflation numbers, they’re going to hide AI-debt off of the balance sheet, etc. Whatever it takes.

Again, let’s just look at a chart.

Why ignore the numbers? We know everyone is raising capex, we know this is a momentum story, and we know that the market lives between only a few binaries now: rising and falling AI capex, dovish and hawkish Fed, Hormuz open and closed. So if all of these things exist on the razor’s edge of a 50/50, how does your risk/reward look here? Are things continuing to go up? Down?

What is your hand? What is the dealer’s up card?

Any hand can win a game of blackjack and any trade can make money, especially on a long (or short) enough time horizon. Your job in markets is to find as many elevens as you can to double down and hit 21. When it feels like oil is falling into shortage, the fed and inflation picture are not conducive to risk-taking, and a war is in-between a rock and a hard place, it seems hard to imagine that the bulls have an eleven. The market has continued to punish AI spenders (GOOG) and sell good prints (INTC). AMZN reports this week and the market expects a whisper-number raise toward $210b in capex, another incremental raise. META reports and the market expects a $5b-15b raise. The market has not gone up as these expectations have.

While the market ex-AI has been resilient, the future seems unclear because of the following: even as Hormuz gets better or worse, it seems harder for the Fed and the inflation picture to let the AI trade get better.

I think that a lot of the market’s easy days are behind it for now. All of the above is uncertain. When certainty comes, price action denies it, and constructive progress in the main issues of this market has been fleeting. Markets do not like uncertainty, and much of the above is skewed to the downside.

Every time it seems the AI trade finds a new leg or inflation gets softer, the progress is fleeting. It has been this way a few months in a row now. We find questions, we panic, the market answers them, and then we find new questions. But stocks don’t make new highs, they don’t respond well to bigger capex numbers, and a lot of the components under the hood get worse. It’s not that anything is breaking, it’s that many small pieces are weakening one at a time: a death by a thousand cuts market.

This morning, a tanker hit a mine and exploded. Within hours, western news sources say we’re near talks to stop the fighting. Who do you believe? Trump, who has said that the US is near a deal with both Iran and Russia for 2 years? Or Iran, whose sole interest is to cause the US as much pain as possible and end the war on stronger footing?

Have we reached peak-Trump-power with midterms 5 months away? Have we reached peak Iran war with two global oil routes nearly shut? Have we reached peak American war-buy-in with munitions storage low? Have we reached peak-AI-spending expectations with AI capex-debt estimates hitting $7T in 3 years, second only to the American housing market?

Again, it’s not that any one thing is snapping. It’s that many core components of the bull market are starting to get a bit shaky. Because of this, the best way I can put it is that I think there are a lot of cases in which the market is hitting on 15 in blackjack. Hitting on 15 has a 58% chance of busting.

The clock of western oil and munitions supplies is ticking and AI has hit something of a critical mass, with obstacles making it hard to continue to scale. Sometimes the dealer busts, and as I said in my last post, markets can quickly resolve higher on good news. It is human nature to find a positive solution in any conflict, and Trump sure is trying. But given the developments of the last week, chances of a sustainable resolution to the upside do not seem greater than the alternative, for now. The path is unclear.

A disclaimer about everything that I write, say and share:

Nothing you will ever read from me is investment advice.

I write about where I think the probabilities lie, and I hope my input in this debate has been of use to you. Invest at your own risk. I never sell stocks short, I never sell options short, and I would never recommend that you do either.

Nothing shared in my posts, comments, messages, or any other content on this platform constitutes investment, financial, legal, tax, or professional advice of any kind. All statements, opinions, analyses, strategies, or recommendations expressed are solely my personal views based on my own research, experience, and decisions. These are not recommendations to buy, sell, hold, or engage in any investment, trade, or financial action.

My content is for informational and entertainment purposes only. You should not replicate, copy, or act upon any information, ideas, or decisions I share. Any actions you take based on this content are entirely at your own risk. By reading, subscribing to, following, sharing, or otherwise engaging with my content, you explicitly acknowledge and accept full responsibility for all outcomes, including any financial losses, damages, or other consequences that may result. You agree to waive any claims against me related to your use of this information. Always consult with a qualified financial advisor, accountant, or other licensed professional before making any investment or financial decisions. Past performance is not indicative of future results. Investments involve substantial risk of loss.

Bear this in mind always, taking advice from anyone.

I spent about 20 minutes coalescing the biggest datacenter accelerator buyers’ actual/forward estimated FCF. It is interesting to look at the totals by year (last row) and company (last column), As you can see, it’s not just that the capital intensity of the buildout is uneven, but it will enter its apex in the next 18 months. Per Bloomberg, the only companies above that will produce FCF at all are GOOGL ($14b!) MSFT (42.4b) and META ($1.3b)! Without those companies, the 2026 FCF burn from the list above is about $124.7b this year.

That FCF burn will require the following :

• the debt markets will have to remain open

• the public must continue to endorse OpenAI and Anthropic’s public spending path: as much as possible for as much intelligence as possible.

• component costs have to remain stable. Many are getting worse, like memory

• No company that has seen the share price punished for this build-out cuts or slows capex

It is critically important you look at the column on the right and consider the implications of CRWV burning $85b, APLD burning $21b, IREN burning $21b, NBIS burning $65b and ORCL burning $43b.

This becomes an important framework to understand as questions around the frontier lab model continue. As I noted a few weeks ago, META is now selling its excess compute, and so is XAI.

It gets pretty crazy when you look under the surface. There’s private credit and off-balance sheet work at META, for example. Or take a look at Google.

The last year saw things get pretty circular in Google’s relationship with Anthropic. Last night’s Google cloud growth included revenue from Anthropic, which was market share that they won after investing $40b in Anthropic. Anthropic’s valuation, which Google supported, then helped Google post a $99b gain on equity securities in the quarter (together with SPCX). This is a large part of why their FCF was negative last night. It’s one thing to create an infrastructure plan that enables growth, but there are quite a few examples of the labs forcing Mag7 to buy growth outright, such as Microsoft’s early relationship with OpenAI.

As this buildout starts burning more and more cash, consider the following:

• the risk of the Federal Reserve’s disposition becoming restrictive

• the amount of liquidity coming to market (hyperscale borrowing to spend, mega IPOs, bond issuance, etc)

• the chance that any compute is now excess (META/SpaceX) or capex is cut

• the presumption that OpenAI/Anthropic’s spending plan will be supported as Chinese open-source models force a rethink

• the risk coming from the now expanding war in Iran

My intention is to write to you all when I feel like I have something really valuable to say. I feel that I owe it to you to check in at least once a week, and especially when I think I can discern something significant. Today, day 2 in the life of a substacker, turns out to be one of those days.

Why?

I spent most of the period between April and June sitting on my hands. Then, I decided to change my views in early June. Why? I had assumed, like many, that Warsh would be a Trump acolyte. He would do the bidding of the president who nominated him and lower interest rates. This may have seemed like a naive assumption because that’s not how Trump’s other nominee, Jay Powell, behaved. But it was reasonable to assume that Trump had learned his lesson this time and been deliberate in vetting his pick, especially because he had said his nominee would lower rates countless times.

Turns out, that was wrong.

I could not have been more wrong. You may have heard the phrase ‘Strong opinions, loosely held.’ I am definitely one of those people, and this was one of those moments. You open your ears, you listen, and you see if your convictions stand the test. Mine did not. In fact, it was time to put the car in reverse.

This FOMC marked the perfect re-test of the top and all-time-high failure in the Nasdaq.

Look at the crosshaired date. Look how differently the Nasdaq 100 trades before and after that date.

Look at the uptrend from April to June. Notice how many of those candles in the uptrend are white: meaning the index closed higher than it opened. During the trading session, buying power won out against selling power, pushing the index higher.

Look at the downtrend since June 22. Notice how many of those candles are blue: meaning the index closed lower than it opened. During the trading session, selling power won out against buying power, pushing the index lower.

So what?

So look at where we are today. I like to use SMH or SOXX because it tracks semiconductors, the undisputed king of our bull market so far.

You may think that because of some of my more recent content that all I look at charts. That’s not true. But they’re the most important thing to watch when it has become clear that a speculative mania has reached a point of madness. It became clear to me that that was likely when the third quarter ended and momentum stocks dominated any set of modern precedent:

So sure, the momentum checks out. Do valuations? Yes. Every equity is an investment in an asset based on its appreciation and its future cash yield. And the reality is people haven’t payed this much for the Nasdaq cash machine in a long time. We’ve nearly doubled the Price/FCF of the Nasdaq. So by any stretch of the imagination, this can be where things are vulnerable. Look at the Nasdaq’s performance after each spike in the green line:

Okay so what are you getting at

Having assumed in June that the following things in my last post could work against the market:

1. Fed pivot to hawkish (I think this thread says it really well)

2. Memory earnings gouging the rest of the AI trade, even causing real inflation

3. A ton of liquidity coming to market (think about it: if buybacks are good because they buy the stocks you are buying, then liquidity coming to market is just as destructive as buybacks are constructive)

4. The Iran ceasefire looked fragile if active whatsoever

5. Some compute mega-buyers have started selling it

6. The cost of tokens is going down because of a price war, hurting the frontier labs, who are particularly vulnerable because they are unprofitable and spending an obscene amount of money

7. Chinese labs found out how to make frontier-level intelligence again, almost certainly for a lot cheaper

I started working under the assumption that the AI stocks were finally going to respond to a new reality. This was fruitful. That brings us to today:

As a put-buyer, I will be the first one to tell you that when a momentum stock looks like the one above, this is part of the cycle where you expect bears to get eviscerated. It becomes way too easy for a stock like this to bounce, quickly and at extreme scale. This not happening is, in my opinion, extremely noticeable.

In my post yesterday, I specified that I did not think it was smart to enter any short AI trades today because bears were in dangerous territory. That said, the market spent the whole day losing all of its gains. This is exactly the kind of thing bears would want to see.

There is a great phrase about things like this: “if my aunt had male parts, she’d be my uncle.” In the same spirit, if my stocks had net buyers, they’d be going up.

While I have spent the better part of the last month buying and owning puts in the AI trade, I still think it is too late to enter any new directionally-short trades. You are asking for it. I’m not saying not to do anything or to get long. I’m just saying that even as someone who owns puts, implied volatility and bid/ask spreads have become too expensive to push it buying puts. Owning them is a different story. While I came in today expecting to be relatively convinced that the market would consolidate, find a bottom, and ease its way into hyperscaler earnings, that was not the case. This is encouraging for my hypothesis that because MTUM rose 43% and SMH rose 77% in one quarter, if we are to correct, it is not likely that it is a brief and small correction.

Moreover, when Trump wants to pump the market, he does it every single day. The president has been noticeably silent about the stock market and continues to escalate in Iran,.

Where is the value?

So that leaves us at our conclusion. I spent a bit of this post helping you all understand my thinking. I know you all come here from my twitter and have seen me come at the AI trade for a lot of its logical fallacies for a long time. Forgive me if this felt repetitive. What I hope to make clear now is this: even as someone who thought that puts in the AI names and the regional banks were cheap last month, I expected this trade to get much harder today and Friday. That the grand majority of tech’s Friday/Monday gains were erased is a major red flag; a signal to me that there is a higher than usual probability that this is likely to be a correction at the scale of the momentum run up we saw in the first place.

All that to say, I do think there is one part of the trade that is particularly interesting. NVDA is cheap, a term that is often times a red flag for cyclical stocks. Many have argued NVDA isn’t cyclical anymore. Perhaps, but not relevant.

During last week’s selloffs, NVDA became the ‘safe’ trade. It was optically cheap and coming into the 200DMA, which is usually good for a lot of support. Those two things coincided and the stock bounced 10%.

NVDA has been pretty stable lately. In my opinion, it’s likely serving as a funding long against many higher beta tech shorts. I think that gets into a riskier place now:

• if the AI trade continues to break, NVDA has downside risk

• if the AI trade begins to work again, it will involve AVGO, AMD, ARM, QCOM, Trainium, and every other DC-wannabe taking market share.

• If the hyperscalers raise capex, it’s fairly likely a decent bit of that is because of higher memory prices, or NVDA competitors

• If hyperscalers lower capex, it reduces NVDA’s TAM.

NVDA has market share to lose in the bull case and TAM to lose in the bear case. It will likely have to continue to use its balance sheet to support neoclouds (like it has CRWV/NBIS/IREN) and growing parts of the AI stack (GLW/NOK etc). I think NVDA getting a higher weighting from any active manager will be a difficult task unless the situation changes in a way that is extremely obvious. If the good news does come, I would still think it nderperforms on the upside, whereas the downside risk is material. There’s also always that risk that buying pressure loses out if it does break through the 200DMA just a bit below where it is now.

I would heavily advise against taking on any new trades this late into a sell-off, especially buying options contracts and shorting stock. But with the repricings we’d already seen, I figured it was safe to assume this sell-off was nearer to its end than its beginning. Today’s action makes me think we may have one more swoop below and that there is a good chance the indicies make another major move, perhaps challenging SPX 7000.

Having started this Substack deep into this sell-off, it’s difficult to say much of anything feels advisable now. But if my stocks had net buyers, they’d be going up. They are not going up, and because they are not going up at levels at which any traditional correction would pause, that’s notable to me. Do not lose sight of the fact that the SPX is barely off of its all time high (!!!) and the risk to the Strait of Hormuz is quite significant. In a binary, I think it’s much more likely that stocks continue to go lower than higher, and I think NVDA is one of the most mispriced assets given its risk balance.

Good luck. Be safe: don’t short-sell stock, don’t short sell options, and don’t take any of this as investment advice. I hope my thinking is of use to you.

A disclaimer about everything that I write, say and share:

Nothing you will ever read from me is investment advice.

I write about where I think the probabilities lie, and I hope my input in this debate has been of use to you. Invest at your own risk. I never sell stocks short, I never sell options short, and I would never recommend that you do either.

Nothing shared in my posts, comments, messages, or any other content on this platform constitutes investment, financial, legal, tax, or professional advice of any kind. All statements, opinions, analyses, strategies, or recommendations expressed are solely my personal views based on my own research, experience, and decisions. These are not recommendations to buy, sell, hold, or engage in any investment, trade, or financial action.

My content is for informational and entertainment purposes only. You should not replicate, copy, or act upon any information, ideas, or decisions I share. Any actions you take based on this content are entirely at your own risk. By reading, subscribing to, following, sharing, or otherwise engaging with my content, you explicitly acknowledge and accept full responsibility for all outcomes, including any financial losses, damages, or other consequences that may result. You agree to waive any claims against me related to your use of this information. Always consult with a qualified financial advisor, accountant, or other licensed professional before making any investment or financial decisions. Past performance is not indicative of future results. Investments involve substantial risk of loss.

Bear this in mind always, taking advice from anyone.

The market has spent the last three weeks digesting a few different variables since semiconductors topped on June 22:

1. A Fed pivot into a new chairman and new regime. The Fed is now no longer outwardly accommodative. If anything, Warsh’s outlook is quite hawkish, and he seems dead-set on doing what’s necessary to get inflation under 2% YoY for the first time in about 5 years. This is a massive difference; Powell’s regime enabled and outright encouraged inflation for the entirety of his post-Covid tenure.

   

2. Memory earnings: Memory is now a component the AI stack working against the rest of the industry. Memory companies were once plagued by cyclicality. The industry is now plunged into deep shortage, and the companies are using their newfound pricing power to strike long-term agreements and control their own destiny for once. This pricing power is so dramatic that the cost of memory relative to the rest of hyperscaler capex has increased from ~8% in 2023 to about 36% of the budget.

1. The hyperscalers are no longer being rewarded for their AI spending. They are being judged for it - in a very fair way. Google has been rewarded for its AI spending, courting Bekrshire Hathaway as a new investor. But this flipped: Google was the first to sit down in a game of musical chairs, raising a shocking $84.75b in dilutive financing and private placements. Meanwhile, Microsoft lost as much as 27% of its value YTD, Meta 20%, and Oracle 35%: their models are not seen as competitive and their spending not seen as worthwhile

1. Liquidity: I will let the numbers speak for themselves:

   1. SpaceX raised $85.7b in its IPO, made its way into the Nasdaq, and spent $60b on Cursor

   2. SK Hynix raised $26.5b in its ADR offering

   3. Anthropic is looking to go public for around $1 trillion

   4. OpenAI wants to go public but actually failed to hit the $1 trillion valuation it is seeking

      Consider the implications of this liquidity waterfall as the fed becomes more hawkish.

2. Iran: I will not waste your time. I will assume you too see the fragility of the deal with Iran, which seems to have now collapsed. Western nations have petroleum reserves but they do not have reserves of refined products. I worry that because Trump has been so outwardly proud of the stock market, Iran is going to target it going into his most vulnerable time: the midterms.

   

   This is compounded by Ukraine’s continued attacks on Russian oil infrastructure. Russia is said to be importing gasoline from India, a shocking development that speaks to the fragility of the situation.

   

   The tail risk to crude and crude products are real.

3. Some compute buyers have become compute sellers and a token war has begun. Meta and XAI are now both leasing compute to other parties: Meta leasing to Anthropic, XAI to Google and Anthropic. AI bulls have quickly noted how easily the compute was purchased by the AI labs. A good omen, no doubt.

   But this misses the point: We have transitioned from every hyperscaler amassing as much compute as they can to 1) those who are not doing a good job being punished for it - which at some point will force management to either moderate this behavior or stop it and 2) those who have too much compute selling it instead of hoarding it

   This matters because there is a whole industry based on the idea that there is scarcity to compute and it should be bought up as quick as possible. Neoclouds can only survive if their depreciation and cost of capital is negated by the demand for compute and what they can charge. They have collateralized their semiconductors (a depreciating asset) many times. Should the demand for this kind of debt run out, what happens to the total demand picture? What kind of deleveraging and rerating occurs if the demand can’t be met?

   Most importantly, it’s not just that some compute is being resold, Mark Zuckerberg is taking on the labs in a token price war:

   

   Why? Because he can. He can use META’s EPS as a weapon against the unprofitable labs. He will happily offer a 2nd tier model, gain users, and learn from the process and threaten competition.

   So where does that leave OpenAI? OpenAI has over $1T in compute spending commitments, yet it doesn’t seem to be better than Anthropic at anything. Its main backers (Oracle and Softbank) are trading under serious stress:

   

4. Kimi: Surely a familiar moment for anyone who traded Deepseek last week:

   

A Chinese lab has, again, figured out how to achieve comparable intelligence with a much smaller budget. This is a great thing for anyone who is not long the equity of an AI lab or anyone not trying to sell semiconductors.

Jevon’s paradox indicates that efficiency gains in token generation will be met with a rise in the use of the resource as it becomes more attainable. AI bulls rejoiced when this seems to have been reaffirmed yet again as Kimi said that usage gains had left them short compute:

This misses the point.

There are 2 major datacenter chip companies and 3 new entrants: Nvidia, Broadcom, and their new threats AMD, Arm, and Qualcomm. There’s also the Amazon Trainium project. They are guiding for some aggressive growth. Look at those numbers in NVDA and AVGO!

That needs financing markets to support either debt/equity financing for the AI labs to buy this or the labs to become profitable. That was looking pretty likely until Zuck cut token prices, as we said above, and many firms created token use limits, such as Uber and Microsoft.

There are a lot of pieces of the stack showing stress in their willingness to allow unprofitable enterprises to borrow money or raise equity to buy semiconductors. Debt markets can tighten if Iran isn’t a quick fix. Nvidia can only do so many circular deals with Coreweave. How much market share can Anthropic take from OpenAI until it starts to lose investor patience?

Those questions could only answered in a calming way until the Chinese labs showed that the cost of intelligence may be much cheaper than we thought. So what does this Kimi tweet mean?

I do not think it means that the compute market has grown. It is incredibly unlikely that the marginally cheaper cost of tokens at Kimi has created new marginal compute spend that didn’t exist last month. It is far more likely that Kimi took market share from OpenAI, Anthropic and Gemini. To compete, incumbent labs need to lower price or create demand with a better product. If neither can be done then the net future value curve of compute spend has moved marginally more flat. This has major implications in a world valuing every bottleneck as a major company and every data center stock doubling its sales off of a massive number in the next 2 years.

So what? Are semiconductor stocks overpriced?

Simply put, semiconductor stocks have put in a massive head and shoulders top (below) after a breathless rally. The sector ETF was up over 70% in one quarter. As I noted below, this is extraordinary over-exposure to one industry, making positive performance and beating expectations an extremely difficult task.

As an enormous skeptic of this entire industry, I will be the first person to tell you that this market has already gone through a fast, drastic repricing, and it is a smart time to use very light exposure. The future is in flux, looking likely to the downside, but bear market rallies can be fast and drastic.

I think that the market has done an extremely effective job repricing the space according to its idiosyncratic company risk. It’s no secret that the KOSPI has led a lot of this charge and it certainly led their correction. Memory and anything down a lot is an extremely dangerous trade in either direction: megacaps report earnings this week and hyperscaler guidance will set the tone for the week (that said, be mindful of how much of it is compensating for memory price increases).

I do not think that anything is in place to create a conducive environment for risk until something changes. Hot take, no. But the point is that I don’t think this Kimi statement changes much of anything, the Iran situation has gotten worse, the Fed has gotten more hawkish, and the AI trade is getting difficult. If you have AI trades on, you may reach the apex opportunity to take them off soon, and the risk/reward becomes difficult as the megacaps report earnings. You are entering a binary environment and without any edge, a 50/50 game is nothing but luck.

Some will point to Gavin Baker’s commentary this week. I have a ton of respect for Gavin Baker and I’ve listened to every one of his podcasts when they’ve come out. But as my friend told me, Gavin has a long term view, he is not taking a 30% position in SOXL (lol).

My earnest opinion is that if anything, the GPU majors are at the most downside risk this week given that stocks like AMD and NVDA have barely corrected in any noticeable way. But options premiums will be expensive and I never recommend shorting stock. Shorts are unlimted risk. I always structure my trades with long options positions (buying puts, buying calls, buying put spreads or call spreads, etc.) I think there is a decent probability that the GPU leaders start to correct more meaningfully as the memory stocks stabilize, but the risk/reward to entering any/all of those trades is not attractive.

The only asymmetric bet I saw this week was in oil, which I think drastically underpriced risk to the upside. The first derivative of this is higher crude prices, the second is oil stocks, the third is the inflation outlook and the fourth is regional banks and AI capital markets. Given this, I think that regional banks are at risk of repricing lower. A reacceleration in inflation expectations means net interest margin compression, credit deterioration and multiple contraction. Any sort of deleveraging, reacceleration in inflation expectations because of Iran or a refined product shortage, or cracks in the AI market will have major implications in tightness of credit and the banks’ ability to earn off of a steep yield curve. The trade is not that ‘hard’ because it’s difficult for regional banks to go much higher very aggressively in this environment and the stocks are at highs not seen in 4 years.

This is not investment advice. I think this is where probabilities lie and I hope my input in this debate has been of use to you. Invest at your own risk. I never sell stock short and I never sell options short and I would never recommend you do either.

Nothing shared in my posts, comments, messages, or any other content on this platform constitutes investment, financial, legal, tax, or professional advice of any kind. All statements, opinions, analyses, strategies, or recommendations expressed are solely my personal views based on my own research, experience, and decisions.These are not recommendations to buy, sell, hold, or engage in any investment, trade, or financial action. My content is for informational and entertainment purposes only.You should not replicate, copy, or act upon any information, ideas, or decisions I share. Any actions you take based on this content are entirely at your own risk. By reading, subscribing to, following, sharing, or otherwise engaging with my content, you explicitly acknowledge and accept full responsibility for all outcomes, including any financial losses, damages, or other consequences that may result. You agree to waive any claims against me related to your use of this information.Always consult with a qualified financial advisor, accountant, or other licensed professional before making any investment or financial decisions. Past performance is not indicative of future results. Investments involve substantial risk of loss.

Subscribe now

A manifesto

What does it mean to think independently?

I started a twitter account a few years ago to anonymously share what I thought were under-served opinions. Back then, my biggest concern was the Trump administration’s brazen embrace of cryptocurrency and the massive inflation of the bubble of Microstrategy:

It was my view then that the market had mistaken Saylor’s overleveraged, convertible-bond/stock/debt fueled buying power for mass-adoption of bitcoin as a store of wealth. Therefore:

1. That was likely to hit a critical mass

2. When that flywheel stopped working for bitcoin, it would work against it

3. MSTR’s use of debt instruments actually made it riskier than bitcoin

4. MSTR’s use of debt instruments and dilution meant it would probably underperform bitcoin and trade below its NAV

5. MSTR’s position in the market was so great and so unsustainable that it was a threat to bitcoin’s bull-run itself

   

The timing on these ideas has turned out to be fortuitous. MSTR has lost most of its value, the debt began to work against it, and the stock traded below the value of the bitcoin it held, taking bitcoin 50% lower in the process.

Then I found myself bothered with the claims about Robinhood. At a simple level, it had gotten to be so big that its valuation was pricing in impossible growth and overlooking the free perks they were giving away.

It bothers me when markets exaggerate and overhype their claims. Innocent people are sucked into the hype and they lose real money doing so. The grandeur makes longs/insiders rich via stock-options and long positions.

So I started having fun with it. It wasn’t without some disagreement.

But the timing there worked quite well, too. It turns out those trends and guides were not sustainable and long-investors were working themselves into a dangerous situation. HOOD lost over half of its value within a few quarters.

But it’s not always bad: this framework of independent thinking allows you to find dramatic mispricings on the upside too:

Fortunately, this framework proved correct: Trump took special interest in Intel and was a positive catalyst for the company’s change: the stock continued to 7x.

It’s not been a perfect run - I’ve often heard that Citadel prides itself in its best traders being right about 52% of the time. But the important revelation has been that independent thinking can help protect investors, find unique opportunities, and make sense of the world around us when the noise is so loud. It is clear that there is a powerful network of equity research, political jawboning, and narrative control that tries to drive markets and society at large. Going along with it will yield results totally commensurate with the rest of the population. Thinking outside of these constraints - whether in monetary policy, the future of AI, geopolitics or markets - opens the door to differentiation.

Therefore

This kind of communication can be difficult on Twitter. I try to limit my posts to the character limit so that my thoughts are easily digestible. It’s been a positive experience: I’ve spoken with a number of great people and I truly enjoy the discourse with all of you. I also hope that, perhaps, my thoughts have made your lives a bit better; and I think that a longer, more insightful, and more detailed forum can do even more of that.

So what’s different?

Twitter is nuts. It is a chaotic and combative environment. You have to love it for that, but it makes detail impossible. I can only share so much on Twitter. I’ll continue to share some, but not like I used to. I will not waste your time on Substack. I try not to on Twitter and I won’t here. Your time and money are valuable, and I respect that.

By subscribing, you will get access to more detailed thoughts, analysis, and implications on the developments of the world around us. I come from a special situations background: great opportunities do not come up every day, but when you do find them, they’re usually worth your while.

I have priced this membership at $19.99/month because each month has 20 trading days in it. For a dollar per day, you’ll get all the insight I have

Disclaimer

Nothing shared in my posts, comments, messages, or any other content on this platform constitutes investment, financial, legal, tax, or professional advice of any kind. All statements, opinions, analyses, strategies, or recommendations expressed are solely my personal views based on my own research, experience, and decisions.These are not recommendations to buy, sell, hold, or engage in any investment, trade, or financial action. My content is for informational and entertainment purposes only.You should not replicate, copy, or act upon any information, ideas, or decisions I share. Any actions you take based on this content are entirely at your own risk. By reading, subscribing to, following, sharing, or otherwise engaging with my content, you explicitly acknowledge and accept full responsibility for all outcomes, including any financial losses, damages, or other consequences that may result. You agree to waive any claims against me related to your use of this information.Always consult with a qualified financial advisor, accountant, or other licensed professional before making any investment or financial decisions. Past performance is not indicative of future results. Investments involve substantial risk of loss.