Why AI Bets Fail: Leverage, Timing, and Runway - AI News & Strategy Daily with Nate B. Jones

Loading summary

Transcript1 lines

[00:00]
A
To tell you what a wedding day, an M5 chip, a margin call, and a guy that's been involved at Enron 20 years ago all have in common. And you might not think that those have anything in common, but by the end of this video, you're going to get the whole story and you're going to see what I see. And I'm Nate B. Jones. What I do is I dig under the news and I find the stories that people miss. And so the story here is the story of Leopold Aschenbrenner and Apple. And nobody has told that story. Nobody has laid that out. But I want to lay out for you how those two stories are the key stories of 2026 when it comes to thinking about investment in AI, because they exemplify such different strategies. So we're going to get into all of it. I can't wait to tell you. Now, Leopold Aschenbrenner is an investor, and he started out his career in AI. He was obviously working at the major labs. He then wrote a famous paper called Situational Awareness, and on the back of that, he raised a fund. And he did very, very well last year and. And did even better this year. Like, he's increased his returns this year, at least until recently. And his whole thesis is that you can predictably understand where to invest in AI by reasoning back from COMPUTE requirements. And you can understand, hey, you know what? Based on compute, these are the companies that are in the supply chain that we can invest in. And he invested on that basis because he said, you know what? Labs are gonna need a lot of compute. I would know. I was there. And he did very well. I think he got something like 20x returns last year. And I think until very recently, he was up over 2x again this year, whatever the exact numbers are. The point is, he did disproportionately well, to the point where every investment fund manager started getting asked, what are you doing about the Leopold Aschenbrenner situational awareness strategy? If you're not doing something about it? Or why aren't you even. Jane Street Capital, which is notoriously a company that prides itself on investing its own money in its own strategies, took a position with Leopold's fund because they thought it was so high potential. And then it all came crashing down. And it came crashing down in just the last few days. Part of how he's doing well is that he is levered. And so what we mean by that, regardless of the exact number, is that he is borrowing money to invest in his high conviction AI products. Now, if you're not an investor, borrowing money to invest might sound crazy, but it's actually something that people do on Wall Street a lot. Some of them do it well, many of them don't do it well. It is certainly higher risk because if you're leveraged, if you have borrowed money, when you go up on the stocks, you go up faster. And when you go down on the stocks, you go down faster and then. And that's the risk you take. Now, Leopold had conviction. He had these returns behind them and he had leverage. And critically, he caught the attention of what I would describe as one of the original wolves of Wall Street, Ken Griffin. This is a guy who has investment history back to the Enron deal. Like he has been a player for a long time. And there's a buying opportunity when there's too much leverage on the table, if you know how to take advantage of that. And Ken Griffin and his company Citadel understood that. And so what happened over the last week, I have people who are in my DM saying this was market manipulation. I assure you no charges will be filed. There is nothing here that Wall street would find illegal or the SEC would be concerned about. It is the ordinary course of business when you're in Wall Street. That's just how it goes. Citadel Capital requires released an investor's note, which they do all the time. That's not unusual on Tuesday. This is after a few weeks of the AI trade getting hammered after a Korean company called SK Hynix launched their IPO and sagged a little bit after their ipo. And so there's already some pressure on Leopold's strategy because we've had a July that is not as up as he probably would have wanted it to be. So in that context, note comes out, note is not about AI. Note is about a Federal Reserve rate hike. Sounds very innocuous. But what Citadel says in their note, and they're totally allowed to say this, anybody can have any opinion they want. It's just an opinion note. They say, we think that the Federal Reserve is going to raise rates. Now, when you raise rates, you make money more expensive. And when you make money more expensive, guess what? AI trade looks less attractive because. Because at the end of the day, you're increasing the price of money. And so anything that's at all volatile or risky or that has received some pressure selling wise in July, which AI had, is going to look less attractive. In other words, Citadel releases this note and something predictable happens. More selling pressure on the AI trade, which means more pressure on Leopold and the situational awareness strategy, which means that because he has leverage, that pressure disproportionately falls on his fund. You can sometimes have so much leverage that your broker will call you and say, I need you to put more capital into the fund so that we can cover red losses that we're seeing in your portfolio. You need to show that you're liquid, put more capital into the fund. That is what was happening to Leopold this week. And this is where it gets very Wolf of Wall Street. His wedding was this past weekend. He was getting ready for the wedding and all of this was happening. You know who doesn't want to get what's called a margin call on their wedding day? Leopold doesn't, but neither do I. Nobody wants a margin call on their wedding day. And Ken Griffin steps in on Thursday and he says, let me take care of this. And what he does is he buys Leopold's entire public equities book. So a book is like your portfolio of orders. He just buys it up. And from Citadel's perspective, they get into the AI trade, which they can have long term conviction about at a great discount on the entry price. And because Citadel has a great reputation on the street, when news breaks during Thursday that Citadel has bought that public equities book, the fact that they did that makes them three or four billion dollars over the course of the day because there's so much confidence in the market, in their choices. Before we get to the Apple piece, let me ladder this up into a story for you. Leopold had high conviction on a particular thesis for AI. He did really, really well investing in that thesis. He got in over his head a little bit on leverage. And now that thesis is still intact, but it's effectively being separated out. And the public part of that thesis is being traded by Citadel. And the private part is something that Leopold is still managing himself. Apple is not an investment story. Apple is not a finance story. Apple is all about long term advantage acquired by investing in hardware and customer experiences. Apple has not a two year plan on AI, not a 10 year vision. Apple is thinking in terms of 20 years and 30 years down the road. When they think about AI and what they're doing is they are after a decade or more in a position where they are designing their chips, their chips are really, really good at local inference, by which I mean using a model locally, generating tokens locally, doing work with AI locally on your Apple machine. It turns out that when you invest in that, you get to be the default winner no matter what model people are using. And that is why startup after startup in the Valley uses Macs. That is why when you are trying to launch OpenClaw back in February, everyone says use a Mac Mini. That is why when we're looking at the future of Apple releases, we are thinking a lot less as AI people about the software and a lot more about what is the next chip going to look like? What is the M6 going to look like? We have M5 right now. How fast is it going to be? And that is why Apple appointed John Ternus to lead the company. Because guess what, his background is in, it's in chips. It's in chips. He has been at the forefront of the heart of Apple's technology ecosystem. And you might think, you know, Steve Jobs would want an ecosystem guy, a CX guy, someone who cares about the customer experience of holding an Apple product and making the iPhone really magical. But I would argue what Apple really needs is someone who understands that their chips play makes them a default winner in the AI race. No matter what lab wins, no matter whether open source wins, no matter whether a frontier lab wins, you still need the chips to effectively generate tokens on your machine. And that gives you so many options. And so all Apple has to do to win, and it's not a guaranteed win, but it's a really strong position. All they have to do is they have to keep memory prices and chip prices enough in check that they can make good Apple margins on their hardware play. And then all they have to do is sit back and continue to invest in that long term hardware position. And they have so many options to win no matter who's on the table. And this by the way, when you're playing smart like this, this gives you a win no matter what. Because if they want to get a model as they have, they go to Google, they cut a deal, they go to Anthropic, maybe they cut a deal there and they get access to a Frontier model and they have the cash on hand to do that and they can deliver the customer experience and it's all backed by the chips that they have. Now it may seem like I am setting up this narrative so Apple is the obvious winner. And if you compare the two, Leopold is the obvious loser because he had a very bad July. I don't think that's true. It is actually really, really hard to make returns in the market and to bank them. And Leopold has been able to do that. I don't think he's going to be out of the investment conversation for long. He is still managing a Fund with billions of dollars in private startup investments. He will be back in the conversation. This is a long race. I think one of the things I take away is, is that if you're thinking about AI investments, you really have to be thinking in the Same decade and 20 year time frames as Apple is. And I think that one of the things I notice is that if you're thinking in those timeframes, you probably are gonna take less leverage because there's room for a lot of volatility along the way to 10 and 20 year time horizons. There's gonna be days when the stock market is up, days when it's down. You wanna be in a position where you're capitalizing over the long term. You can be in a fantastic hardware position, but you have to intentionally capitalize on it to get where you want to go. I made a video a few months ago that a bunch of you watched about the idea that if Apple wanted to, they could get into a position to offer effective local compute to lots of businesses for AI. And so I think that one of the things that's really interesting right now is that you can have all of those advantages. You can be reaping a lot of passive margins from being in a good position on the ships business. And you still may be under monetized in AI because you're not fully realizing how impactful AI will be across the entire customer spectrum from enterprise customers to small business customers to end consumers like you and me. So it might seem like Apple's just the default winner in this story, but I think it's a little bit more complicated. And I'm sharing this because I'd love your perspective. Sound off in the comments, let me know who do you think has got the better strategy here? How would you play this? How do you think about AI as a story that you're invested in? And I've got news for you. You may not be a stock market investor, but because of AI's impact on the world, with our time, with our attention, with what we're doing with our tools, we are all invested in AI. And I hope that this has been a reminder to you that there are so many different ways to approach the future of AI and you have a lot of options on the table. What's yours?

Summary6 min read

Podcast Summary: Why AI Bets Fail—Leverage, Timing, and Runway

Podcast: AI News & Strategy Daily with Nate B. Jones
Host: Nate B. Jones
Episode Date: August 4, 2026

Episode Overview

In this episode, Nate B. Jones dives deep into the stories of two major players shaping the 2026 AI investment landscape: investor Leopold Aschenbrenner and tech giant Apple. With wit and clarity, Nate unpacks why some high-profile AI bets fail and how differences in leverage, timing, and strategic runway make all the difference between short-term turbulence and long-term dominance. This episode delivers an actionable framework for understanding the risks and rewards of different AI investment strategies, making it a must-listen for executives, builders, and the AI-curious.

Key Discussion Points & Insights

1. Setting the Stage: Wild Contrasts in AI Strategy

A unique analogy kicks off the episode, tying together a wedding day, Apple’s M5 chip, a margin call, and a Wall Street veteran from the Enron era—all as metaphors for the hidden forces driving AI investments.
“You might not think that those have anything in common, but by the end of this video, you’re going to get the whole story and you’re going to see what I see.” (00:10)

2. The Rise and Falter of Leopold Aschenbrenner’s AI Bet

Leopold Aschenbrenner’s approach was rooted in his famed "Situational Awareness" paper and his insider perspective from working at major AI labs.
His thesis: Invest in firms across the AI compute supply chain, predicting demand from first principles.
Performance: Spectacular returns—20x last year, and over 2x again this year—quickly made Aschenbrenner’s approach a template for institutional investors.
Even Jane Street Capital invested, highlighting his influence.
Notable quote: “Every investment fund manager started getting asked, what are you doing about the Leopold Aschenbrenner situational awareness strategy?” (02:13)

3. The Dangers of Leverage: Crisis Strikes

Leverage’s double-edged sword: Aschenbrenner’s fund was highly levered, amplifying both gains and losses.
The catalyst: A streak of poor market performance for AI stocks—especially after SK Hynix’s IPO underwhelmed—increased selling pressure.
Trigger event: Citadel Capital released a routine investor note, suggesting a likely Federal Reserve rate hike, making capital more expensive and speculative trades less attractive.
Market impact: This move intensified pressure on AI bets, forcing Aschenbrenner into a margin call during his own wedding weekend.
Dramatic moment: “You know who doesn’t want to get what’s called a margin call on their wedding day? Leopold doesn’t, but neither do I.” (07:40)

4. Old Wolves, New Moves: Citadel Steps In

Ken Griffin (Citadel) spots a buying opportunity as Aschenbrenner faces forced sell-offs.
Citadel acquires Aschenbrenner’s entire public equities book, instantly gaining a major AI position—at a discount.
Market reaction: Confidence in Citadel’s buy was so strong that it added $3–4 billion in value for Citadel that very day.
Insight: “When news breaks…that Citadel has bought that public equities book, the fact that they did that makes them three or four billion dollars over the course of the day because there’s so much confidence in the market, in their choices.” (09:22)

5. Apple’s Silent (and Solid) Game—Long-Term AI Dominance

Apple’s AI story is not about stock market bets, but about generational planning and control of hardware and customer experience.
Their strategy: Multi-decade runway, designing advanced chips for local AI inference, ensuring Macs and other Apple hardware are future-proof for any model (third-party or otherwise).
Apple’s key move: Appointment of John Ternus, a chip architecture expert, as company leader, indicating the central importance of hardware in Apple’s AI play.
Quote: “What Apple really needs is someone who understands that their chips play makes them a default winner in the AI race. No matter what lab wins… you still need the chips to effectively generate tokens on your machine. And that gives you so many options.” (14:11)
Scenario comparison:
- Aschenbrenner’s fast, high-leverage, thesis-driven punt vs. Apple’s slow, resilient, hardware-based dominance.
- Major insight: Startups and labs—even those at the cutting edge—are building on Apple chips, making Apple the default infrastructure winner regardless of which AI model comes to dominate.
- Quote: “It might seem like I am setting up this narrative so Apple is the obvious winner… but I think it’s a little bit more complicated.” (18:41)

6. Lessons for AI Investors: Time Horizons & Leverage

The key takeaway: Short-term leverage magnifies swings, but long-term vision reduces the urge for dangerous risk.
- “If you’re thinking in those [Apple’s] timeframes, you probably are gonna take less leverage because there’s room for a lot of volatility along the way to 10 and 20 year time horizons.” (19:15)
Apple’s position: Succeed regardless of model winner, as long as they control chip supply and pricing.
Risk of under-monetization: Even with hardware dominance, full AI revenue potential depends on realizing AI’s impact across customers—from large enterprises to individual consumers.
- “You can be reaping a lot of passive margins from being in a good position on the chips business. And you still may be under monetized in AI because you’re not fully realizing how impactful AI will be across the entire customer spectrum.” (21:05)

7. Final Insights, Audience Challenge

Notable closing: Nate reminds listeners that everyone—not just investors—is affected by AI’s evolution and must decide their own stake.
- “You may not be a stock market investor, but because of AI’s impact on the world, with our time, with our attention, with what we’re doing with our tools, we are all invested in AI.” (23:08)
“There are so many different ways to approach the future of AI and you have a lot of options on the table. What’s yours?” (23:37)

Timestamps for Key Segments

| Timestamp | Segment Description | |-----------|--------------------------------------------------------------| | 00:00 | Episode intro; comparing weddings, M5 chips, margin calls | | 01:22 | Introduction to Leopold Aschenbrenner and his AI thesis | | 03:00 | Overview of returns and investment strategy | | 05:30 | How leverage amplifies gains (and risks) | | 07:40 | Margin call drama on Aschenbrenner’s wedding weekend | | 09:22 | Citadel’s Ken Griffin buys entire public equities book | | 12:00 | The psychology and market dynamics behind Citadel’s move | | 13:30 | Transition: Comparing Apple’s strategy | | 14:11 | Apple’s hardware focus, chip prowess, John Ternus’ rise | | 17:28 | Apple’s options for model partnerships and customer impact | | 19:15 | Lessons in timeframe and leverage for AI investors | | 21:05 | The under-monetization risk—even for hardware winners | | 23:08 | Closing thoughts; call to audience action |

Memorable Quotes

On AI investment extremes:
“Leopold had high conviction on a particular thesis for AI. He did really, really well investing in that thesis. He got in over his head a little bit on leverage.” (10:42)
On Apple’s unique position:
“They have so many options to win no matter who’s on the table… all they have to do is keep memory prices and chip prices enough in check that they can make good Apple margins on their hardware play.” (16:55)
On why this matters to everyone:
“Because of AI’s impact on the world, with our time, with our attention, with what we’re doing with our tools, we are all invested in AI.” (23:08)

Takeaways

Leverage is tempting, but dangerous. Short-term outperformance can vanish overnight with leverage, especially in volatile markets.
Apple plays the long game. Dominance in chips and patient, generational thinking offer unmatched resilience as AI models and platforms shift.
Every organization (and person) has a stake in AI’s future. Whether you’re making investments or simply choosing your devices, the strategies explored here will shape the landscape you live and work in.
Nate’s challenge: What’s your approach? Are you thinking short-term, high-risk—or are you building for the long haul?

For deeper analysis:
Sign up for Nate’s newsletter at natesnewsletter.substack.com

Loading summary

Transcript1 lines

[00:00]
A
To tell you what a wedding day, an M5 chip, a margin call, and a guy that's been involved at Enron 20 years ago all have in common. And you might not think that those have anything in common, but by the end of this video, you're going to get the whole story and you're going to see what I see. And I'm Nate B. Jones. What I do is I dig under the news and I find the stories that people miss. And so the story here is the story of Leopold Aschenbrenner and Apple. And nobody has told that story. Nobody has laid that out. But I want to lay out for you how those two stories are the key stories of 2026 when it comes to thinking about investment in AI, because they exemplify such different strategies. So we're going to get into all of it. I can't wait to tell you. Now, Leopold Aschenbrenner is an investor, and he started out his career in AI. He was obviously working at the major labs. He then wrote a famous paper called Situational Awareness, and on the back of that, he raised a fund. And he did very, very well last year and. And did even better this year. Like, he's increased his returns this year, at least until recently. And his whole thesis is that you can predictably understand where to invest in AI by reasoning back from COMPUTE requirements. And you can understand, hey, you know what? Based on compute, these are the companies that are in the supply chain that we can invest in. And he invested on that basis because he said, you know what? Labs are gonna need a lot of compute. I would know. I was there. And he did very well. I think he got something like 20x returns last year. And I think until very recently, he was up over 2x again this year, whatever the exact numbers are. The point is, he did disproportionately well, to the point where every investment fund manager started getting asked, what are you doing about the Leopold Aschenbrenner situational awareness strategy? If you're not doing something about it? Or why aren't you even. Jane Street Capital, which is notoriously a company that prides itself on investing its own money in its own strategies, took a position with Leopold's fund because they thought it was so high potential. And then it all came crashing down. And it came crashing down in just the last few days. Part of how he's doing well is that he is levered. And so what we mean by that, regardless of the exact number, is that he is borrowing money to invest in his high conviction AI products. Now, if you're not an investor, borrowing money to invest might sound crazy, but it's actually something that people do on Wall Street a lot. Some of them do it well, many of them don't do it well. It is certainly higher risk because if you're leveraged, if you have borrowed money, when you go up on the stocks, you go up faster. And when you go down on the stocks, you go down faster and then. And that's the risk you take. Now, Leopold had conviction. He had these returns behind them and he had leverage. And critically, he caught the attention of what I would describe as one of the original wolves of Wall Street, Ken Griffin. This is a guy who has investment history back to the Enron deal. Like he has been a player for a long time. And there's a buying opportunity when there's too much leverage on the table, if you know how to take advantage of that. And Ken Griffin and his company Citadel understood that. And so what happened over the last week, I have people who are in my DM saying this was market manipulation. I assure you no charges will be filed. There is nothing here that Wall street would find illegal or the SEC would be concerned about. It is the ordinary course of business when you're in Wall Street. That's just how it goes. Citadel Capital requires released an investor's note, which they do all the time. That's not unusual on Tuesday. This is after a few weeks of the AI trade getting hammered after a Korean company called SK Hynix launched their IPO and sagged a little bit after their ipo. And so there's already some pressure on Leopold's strategy because we've had a July that is not as up as he probably would have wanted it to be. So in that context, note comes out, note is not about AI. Note is about a Federal Reserve rate hike. Sounds very innocuous. But what Citadel says in their note, and they're totally allowed to say this, anybody can have any opinion they want. It's just an opinion note. They say, we think that the Federal Reserve is going to raise rates. Now, when you raise rates, you make money more expensive. And when you make money more expensive, guess what? AI trade looks less attractive because. Because at the end of the day, you're increasing the price of money. And so anything that's at all volatile or risky or that has received some pressure selling wise in July, which AI had, is going to look less attractive. In other words, Citadel releases this note and something predictable happens. More selling pressure on the AI trade, which means more pressure on Leopold and the situational awareness strategy, which means that because he has leverage, that pressure disproportionately falls on his fund. You can sometimes have so much leverage that your broker will call you and say, I need you to put more capital into the fund so that we can cover red losses that we're seeing in your portfolio. You need to show that you're liquid, put more capital into the fund. That is what was happening to Leopold this week. And this is where it gets very Wolf of Wall Street. His wedding was this past weekend. He was getting ready for the wedding and all of this was happening. You know who doesn't want to get what's called a margin call on their wedding day? Leopold doesn't, but neither do I. Nobody wants a margin call on their wedding day. And Ken Griffin steps in on Thursday and he says, let me take care of this. And what he does is he buys Leopold's entire public equities book. So a book is like your portfolio of orders. He just buys it up. And from Citadel's perspective, they get into the AI trade, which they can have long term conviction about at a great discount on the entry price. And because Citadel has a great reputation on the street, when news breaks during Thursday that Citadel has bought that public equities book, the fact that they did that makes them three or four billion dollars over the course of the day because there's so much confidence in the market, in their choices. Before we get to the Apple piece, let me ladder this up into a story for you. Leopold had high conviction on a particular thesis for AI. He did really, really well investing in that thesis. He got in over his head a little bit on leverage. And now that thesis is still intact, but it's effectively being separated out. And the public part of that thesis is being traded by Citadel. And the private part is something that Leopold is still managing himself. Apple is not an investment story. Apple is not a finance story. Apple is all about long term advantage acquired by investing in hardware and customer experiences. Apple has not a two year plan on AI, not a 10 year vision. Apple is thinking in terms of 20 years and 30 years down the road. When they think about AI and what they're doing is they are after a decade or more in a position where they are designing their chips, their chips are really, really good at local inference, by which I mean using a model locally, generating tokens locally, doing work with AI locally on your Apple machine. It turns out that when you invest in that, you get to be the default winner no matter what model people are using. And that is why startup after startup in the Valley uses Macs. That is why when you are trying to launch OpenClaw back in February, everyone says use a Mac Mini. That is why when we're looking at the future of Apple releases, we are thinking a lot less as AI people about the software and a lot more about what is the next chip going to look like? What is the M6 going to look like? We have M5 right now. How fast is it going to be? And that is why Apple appointed John Ternus to lead the company. Because guess what, his background is in, it's in chips. It's in chips. He has been at the forefront of the heart of Apple's technology ecosystem. And you might think, you know, Steve Jobs would want an ecosystem guy, a CX guy, someone who cares about the customer experience of holding an Apple product and making the iPhone really magical. But I would argue what Apple really needs is someone who understands that their chips play makes them a default winner in the AI race. No matter what lab wins, no matter whether open source wins, no matter whether a frontier lab wins, you still need the chips to effectively generate tokens on your machine. And that gives you so many options. And so all Apple has to do to win, and it's not a guaranteed win, but it's a really strong position. All they have to do is they have to keep memory prices and chip prices enough in check that they can make good Apple margins on their hardware play. And then all they have to do is sit back and continue to invest in that long term hardware position. And they have so many options to win no matter who's on the table. And this by the way, when you're playing smart like this, this gives you a win no matter what. Because if they want to get a model as they have, they go to Google, they cut a deal, they go to Anthropic, maybe they cut a deal there and they get access to a Frontier model and they have the cash on hand to do that and they can deliver the customer experience and it's all backed by the chips that they have. Now it may seem like I am setting up this narrative so Apple is the obvious winner. And if you compare the two, Leopold is the obvious loser because he had a very bad July. I don't think that's true. It is actually really, really hard to make returns in the market and to bank them. And Leopold has been able to do that. I don't think he's going to be out of the investment conversation for long. He is still managing a Fund with billions of dollars in private startup investments. He will be back in the conversation. This is a long race. I think one of the things I take away is, is that if you're thinking about AI investments, you really have to be thinking in the Same decade and 20 year time frames as Apple is. And I think that one of the things I notice is that if you're thinking in those timeframes, you probably are gonna take less leverage because there's room for a lot of volatility along the way to 10 and 20 year time horizons. There's gonna be days when the stock market is up, days when it's down. You wanna be in a position where you're capitalizing over the long term. You can be in a fantastic hardware position, but you have to intentionally capitalize on it to get where you want to go. I made a video a few months ago that a bunch of you watched about the idea that if Apple wanted to, they could get into a position to offer effective local compute to lots of businesses for AI. And so I think that one of the things that's really interesting right now is that you can have all of those advantages. You can be reaping a lot of passive margins from being in a good position on the ships business. And you still may be under monetized in AI because you're not fully realizing how impactful AI will be across the entire customer spectrum from enterprise customers to small business customers to end consumers like you and me. So it might seem like Apple's just the default winner in this story, but I think it's a little bit more complicated. And I'm sharing this because I'd love your perspective. Sound off in the comments, let me know who do you think has got the better strategy here? How would you play this? How do you think about AI as a story that you're invested in? And I've got news for you. You may not be a stock market investor, but because of AI's impact on the world, with our time, with our attention, with what we're doing with our tools, we are all invested in AI. And I hope that this has been a reminder to you that there are so many different ways to approach the future of AI and you have a lot of options on the table. What's yours?