THE LATEST ANALYSIS OF THE U.S. ‘BOY OF THE VERSION’: 9 BILLION DOLLARS TO MAKE AN EMPTY NVIDIA, BULLETS TO POWER AND MEMORY PLATES

Original video title: Leopold Aschenbrenner says "No More Stocks!"
Source: Limitless Podcast
Original language: Deep tide TechFlow

Summary of highlights

Leopold Aschenbrenner, who is considered one of the world’s most radical AI investors, uses about $9 billion in nominal warehouse space in the open market to empty NVIDIA, ASML and Oracle while shifting funds to electricity, memory, data centre networks, and deeper AI infrastructure and model assets such as Anthropic。

According to the two facilitators, this does not mean that the AI bubble has broken down, but it is more like a rotational signal that infrastructure transactions have shifted from "chip priority" to "energy, network, room building priority" especially after NVIDIA has just completed $25 billion bond financing and the Anthropic valuation has been pushed up。

Summary of outstanding views

Core transaction logic for Leopold

– "The most classic" for shovel sales in AI: "The deal is too crowded, and Leopold's recent change of position is sending the signal."

• “HIS JUDGMENT IS NOT THAT THE AI INFRASTRUCTURE HAS PEAKED, BUT THAT CERTAIN LEVELS IN INFRASTRUCTURE STACKS, ESPECIALLY SEMICONDUCTORS AND TRADITIONAL HOTSPOTS, HAVE BECOME OVERCROWDED.”

• “If the question turns, where will the money go next? There are two answers. The first is the flow to the next real infrastructure bottleneck, namely, electricity, memory and data centre networks. The second answer is the mysterious investment that was exposed only a few weeks ago."

• “He has been betting on very infrastructure-oriented things, both for optical and electricity-related companies.”

• "If he is careful about NVIDIA, then the money goes to the power, the memory, and at the same time he wants to go directly to the mine," instead of continuing to buy only shovels," and Anthropic is his favorite mine

NVIDIA FINANCIAL RELEASE SIGNAL

– "THE QUESTION IS NOT WHETHER NVIDIA WILL CONTINUE TO MAKE MONEY, BUT RATHER WHY IT NEEDS TO BORROW $25 BILLION FROM OUTSIDE A COMPANY WITH A VERY HIGH PROFITABILITY AND AN ALREADY LARGE AMOUNT OF CASH ON THE BOOKS."

• “IF A COMPANY BUYS SHARES IN THE SAME MONTH WITH HUGE AMOUNTS OF MONEY AND THEN BORROWS MONEY, IT OBVIOUSLY DOES NOT BORROW FOR LACK OF MONEY. AND THE MORE RATIONAL EXPLANATION IS THAT IT'S CHEAP MONEY, AND AI IS MAKING A SLIGHT SHIFT IN THE WAY IT IS FINANCED

NEXT WAVE OF AI INFRASTRUCTURE DIVIDENDS

• “THE REAL BOTTLENECK IS NOT JUST THE GPU, BUT POWER, MEMORY, THE DATA CENTRE NETWORK AND THE ABILITY TO ACTUALLY BUILD UP THESE THINGS.”

• “You can't build data centres fast enough to expand enough memory chip production, and you can't expand power grids, power lines and related infrastructure immediately.” There are not so many men on the ground, and approval, supervision and procedures are stopping you."

• “Whoever has the capacity to build a data centre will make money”

Light modules, copper and fibre optics

• "WHEN THE GPU IS GROWING IN SIZE, THE COPPER LINE WILL BECOME HOTTER, ENERGY WILL BECOME MORE EXPENSIVE AND INEFFICIENT, AND FIBRE-OPTICS WILL BE THE NEXT COURSE OF UPGRADING IN THIS CASE."

• "The short-range transmission of how high the bandwidth is, copper is almost the only material you really want to use. Only when it begins not to apply, such as too far away or too much heat, will it reach fibre optics, so that the market is now in strong demand for copper and fibre optic combinations.”

• “The copper futures have recently gone very strongly, essentially because everyone needs it, which is the most critical base material in short-range high bandwidth, and fibre-optics is the next step”

• “Current copper remains the most critical material in the short-range high bandwidth transmission, but once the distance is lengthened and the heat is too high, it must be turned to fibre optics”

• "The next step is to finance infrastructure companies that don't sound sexy."

Why is energy the best bet

• "I'VE ALWAYS BEEN VERY INTERESTED IN ENERGY BECAUSE EVEN IF THE DEMAND FOR AI SLOWS DOWN, ENERGY ITSELF IS STILL A GLOBAL DEMAND, AND THIS DEMAND WILL ONLY INCREASE."

• “The single trend that will continue to rise in any situation is our demand for energy, electricity and power, companies that I would most like to do in the long term.”

"What I'd like to talk to is the companies that Jensen pitched, at the same time, with Leopold's logic. So the closest I can now is Marvell

• “The best long-term position is not necessarily the hottest chip company, but rather those macro-levels that do not bypass the electricity infrastructure of the past.”

AI portfolio for Leopold

Josh Kale:Leopold Aschenbrenner, a 24-year-old young man who specializes in AI, is now almost the best AI investor in the world. There are rumours that his nominal warehouse size has exceeded $20 billion. When we looked at Ejaaz's post a month ago, the size of the fund was only $13.7 billion, which is almost double every quarter。

THIS TIME WE HAVE A FEW VERY IMPORTANT NEW CHANGES IN HIS RECENT INVESTMENT MOVEMENT. WE TALKED ABOUT HIS PORTFOLIO IN THE LAST ISSUE, AND THE MOST SURPRISING THING WAS THAT HE WAS DOING A COMPANY THAT ALMOST EVERYONE KNEW, I.E. THE HIGHEST GLOBAL MARKET VALUE AND HIGHEST HEAT IN NVIDIA. A LOT OF PEOPLE DON'T UNDERSTAND WHY HE'S MAKING AN EMPTY LAYOUT FOR A COMPANY LIKE THIS FOR OVER $9 BILLION。

NOW WE HAVE A NEW CLUE THAT MIGHT EXPLAIN THIS. NVIDIA IS FINANCING, AND IT IS FINANCING THROUGH DEBT. IT'S NOT REASONABLE ON THE SURFACEWHY WOULD A COMPANY LIKE NVIDIA, WITH A HUGE VOLUME AND A VERY HIGH PROFIT MARGIN, TAKE ANOTHER $25 BILLION IN CASH JUST FINISHEDToday we'd like to combine Leopold's portfolio and talk about why he's making so much money, what he's looking at next, and what the financing of NVIDIA really means。

Ejaaz Ahamadeen:Let me give you a background. Leopold Aschenbrenner, a former researcher at OpenAI, raised a fund about a year and a half to two years ago, initially on a small scale, which I remember was about $200 million, but from his last 13F, the fund’s open hold was worth $13.7 billion。

The market was naturally curious to know which slots he had put in, what the core investment logic was, and where the next big deal would fall。

To understand this, just a month ago, Leopold was really optimistic about the entire AI track, especially about the logic of "sell shovels" -- that is, NVIDIA, such as GPU, and upstream hardware suppliers。

But about a month ago, the market found that he did not look that much at the semiconductor line。HE STILL SEES THE REAL BOTTLENECKS OF MEMORY, ELECTRICITY, AND PERHAPS NEW CLOUD MANUFACTURERS, BUT HE PREFERS THE WORLD'S MOST VALUABLE COMPANY, NVIDIA。More specifically, he was able to crush some $9 billion in falling positions in companies such as NVIDIA, ASML and Oracle, which are considered to be core beneficiaries of AI infrastructure。

LOGIC FOR EMPTY NVIDIA

Ejaaz Ahamadeen:AS SOON AS THIS THING GOT OUT, A LOT OF PEOPLE STARTED TO WORRY ABOUT WHETHER THE AI BUBBLE WAS BREAKING. AFTER ALL, ON THE SURFACE, NVIDIA'S GPU IS STILL SELLING, AND DEMAND IS NOT CLEARLY WEAK

AND THEN WE CAME UP WITH A FEW NEW CLUES, THE MOST IMPORTANT OF WHICH IS THAT NVIDIA JUST TOOK $25 BILLION FROM OUTSIDE THROUGH BOND FINANCING. THIS MEANS THAT IT IS NOT SIMPLY USING THE MONEY ON ITS OWN ACCOUNT, BUT IS ADDING ADDITIONAL LEVERAGE. SO THE QUESTION IS:Why do you have to borrow $25 billion from the world's most profitable, most profitable, most cash-flowing company

Josh KaleAnd at first, they only planned to absorb $20 billion, and eventually to expand to $25 billion, more than three times the subscription. When we talked about this portfolio in the previous period, we also said, "Don't worry about bubbles, because these companies, despite their huge capital expenditures, have a sufficiently high income to theoretically support expansion on their own balance sheets。

However, for the first time since 2021, NVIDIA has apparently financed from outside rather than directly using cash on its own account. I remember it was about 12 billion dollars in cash. Put them all together, there's a strange tension: On the one hand, Leopold is doing nothing, on the other hand, NVIDIA is paying debts when it looks like there's no limit to cash and profits. So what happened

NVIDIA BOND FINANCING DISMANTLING

Josh Kale: Ejaaz, can you help us open the deal? Because it is not financing in the general sense, it is the issuance of bonds. In the end, now NVIDIA's balance sheet is over $25 billion, and interest rates should look low。

Ejaaz Ahamadeen:I PUT OUT TWO EXPLANATIONS. NVIDIA HAD ABOUT $13.7 BILLION IN CASH ON ITS BOOKS, THAT IS, IT COULD HAVE SPENT ITS OWN MONEY DIRECTLY. THEN WHY DO YOU HAVE TO FINANCE IT FROM OUTSIDE? THE SIMPLEST ANALOGY IS BUYING A HOUSE. MANY PEOPLE, EVEN IF THEY HAD THE FULL AMOUNT, WOULD STILL OPT FOR LOANS BECAUSE THEY HAD THE CAPITAL TO DO SOMETHING ELSE, AND THE COST OF BORROWING WOULD BE MORE ECONOMICAL IF IT WAS LOW ENOUGH。

THE INTEREST RATE ENVIRONMENT HAS NOT BEEN FRIENDLY IN THESE YEARS, BUT IF YOU ARE NVIDIA, ONE OF THE MOST VALUABLE AND POPULAR COMPANIES IN THE WORLD, THEN YOU CAN BORROW MONEY ON FAIRLY GOOD TERMS. THIS TIME, THE MATURITY OF $25 BILLION IN BOND FINANCING RANGED FROM 2 TO 30 YEARS, ALMOST AT A VERY LOW COST, AND INTEREST RATES WERE ALREADY CLOSE TO US TREASURY DEBT RETURN。

And it's probably four times oversubscribed, in other words, there's $85 billion on the market that wants to break into $25 billion, and NVIDIA can almost pick investors. If only official statements are made, NVIDIA explains that this is mainly a financial arrangement used to repay and refinance part of the existing debt. Google did something similar a few weeks ago, once this February. So, of course, you can accept this explanation as financial optimization。

But it's hard to ignore the other side:Over the past month and a half, NVIDIA, Amazon, Google and a few other super-large cloud manufacturers have been financing almost entirely externally。Some pay debts, some sell stocks. Leopold's opinion may not be completely unreasonable. Is this the signal that the foam starts to loose and the card house starts to shake? But if you look at the financial structure alone, it is not clear that it is dangerous。

Josh Kale:SO DO I. NINE BILLION DOLLARS TO EMPTY NVIDIA, WHICH IS REALLY A VERY BIG SLOT. BUT WHEN WE DID THE RESEARCH, WE SAW ANOTHER THING: ON MAY 18, THE NVIDIA BOARD JUST AUTHORIZED AN ADDITIONAL $80 BILLION TO BE BOUGHT BACK, ADDING 25 CENTS PER SHARE, A DIRECT INCREASE OF 25 TIMES。

IT IS CLEAR THAT IF A COMPANY BUYS SHARES IN THE SAME MONTH WITH A LARGE AMOUNT OF MONEY, INCREASES DIVIDENDS AND BORROWS MONEY, IT DOES NOT BORROW IT FOR LACK OF MONEY. AND THE MORE LOGICAL EXPLANATION IS THAT IT'S CHEAP MONEY, AND AI IS MAKING A SLIGHT SHIFT IN THE WAY IT IS FINANCED。EVERYONE WANTED TO BE INVOLVED IN THESE CAPITAL OPERATIONS, AND NVIDIA REALIZED THAT IT WAS EVEN CHEAPER TO PAY FOR ITS OWN DEBT THAN OTHER FORMS OF FINANCING, SO IT JUST WENT ALONG WITH IT. AT LEAST NOW, NVIDIA ITSELF IS STILL DOING WELL。

Why'd he reposition it

Josh Kale: This is another question. Leopold, what are you thinking? Why does his judgment change? The stock-price chart you just showed shows that NVIDIA's not really doing very well lately, but it's not that bad. It's also the world's largest company, nearing the market value of $5 trillion, falling by 7 per cent a month, not much in the context of other AI stock booms。

Ejaaz Ahamadeen:I DON'T THINK NVIDIA WILL DISAPPEAR. IT'S GPU, INCLUDING THE CPU PRODUCT LINE JUST A FEW WEEKS AGO, AND I THINK IT'S GOING TO BE VERY GOOD. AI ' S PRODUCT DEMAND IS NOW EXPONENTIAL, AND THE CORE MACHINE SUPPLIER THAT CAN TRULY ABSORB THIS DEMAND IS STILL PREDOMINANTLY NVIDIA。

But I do thinkThe most classic "sell shovel" deal in AI has become too crowded, and Leopold's recent change of position is sending the signal。Look at his recent 13F, he knows that he's seen the line of a clearly emptied semiconductor, like NVIDIA, ASML, Oracle, and several other infrastructure companies。

At the same time, however, he re-engineered the direction of memory, electricity and new clouds. That means..HIS JUDGEMENT IS NOT THAT THE AI INFRASTRUCTURE HAS PEAKED, BUT THAT CERTAIN LEVELS IN INFRASTRUCTURE STACKS, ESPECIALLY SEMICONDUCTORS AND TRADITIONAL HOTSPOTS, HAVE BECOME OVERCROWDED。

If the question becomes where the money goes next, there are two answers. The first is the next real infrastructure bottleneck, electricity, memory, data centre networks; the second is the mysterious investment that was exposed only a few weeks ago。

Anthropic space unexpectedly exposed

Josh Kale:This was the most surprising thing, and I didn't know until yesterday, after listening to you, that my first reaction was impossible。Did Leopold's fund, "Situational Awareness", have 20% of its equity in AnthropicThere are now rumours that the company holds about one fifth of the space of the Leopold Foundation, as the Wall Street Journal and several other media outlets have said, and that has been confirmed by people very close to trading。

This became one of his combinations that was completely unforeseeable by the market。

Because 13F can only disclose open market holdings and not private equity, and Anthropic is exactly a large, unlisted shareholding. It was also for that reason that one began to understand why the outside world would carry his portfolio valuation to $20 billion。

If 20% of the fund was Anthropic, and he was probably in early 2025, that year's return was like seven years on Anthropic. This change would have led to a very significant revision of our understanding of his entire portfolio。

Ejaaz Ahamadeen:Yeah. The first time he went through private channels or funds to invest in Anthropic was in March 2025, when Anthropic valued about $60 billion. It has now been set at $965 billion according to the latest valuation round。

It's almost 15 times the increase. According to the algorithm that we're showing on our show today, his most recent 13F disclosure of the liquidity portfolio was $13.7 billion, and if added to the Anthropic space in the Wall Street Journal, it would probably add another $7 billion, the size of the entire fund would be $20 billion。

How far is this going to go? Bill Ackman, the top investor in the market for 30 or 40 years, is about $20 billion in the size of Pershing Capital. Leopold entered the game only a year and a half, and he was only 24 years old, with little real investment experience。

But he made some amazing judgments, and it's crazy that he wrote it all in advance. When he launched the fund a year and a half ago, he published an AI long-page 65-page version, “Situational Awareness”, which provides almost a complete picture of the whole logic, including how funds can rotate from semiconductors and parts of infrastructure to other bottlenecks. The market is now moving along this line, which is really amazing。

Next wave of infrastructure

Ejaaz Ahamadeen:So tell me where the next money goes。If he was careful about NVIDIA, he would go to the power, the memory, and at the same time he wanted to go directly to the mine itself, instead of continuing to buy only shovels, which was his favorite mine。

Josh Kale:THIS DOES LOOK LIKE A NEW TREND, AND IT IS STILL ONE STEP AHEAD OF MOST PEOPLE. OVER THE PAST 12 MONTHS, YOU'VE BEEN LOOKING FOR A BOTTLENECK OF AI, RARE METALS, MEMORY, RAM, AND SO ON. THESE JUDGEMENTS ARE NOT WRONG, BECAUSE THE WAVE DID OCCUR。

But now, the directions that are considered bottlenecks are gradually being rationalized. There is a better understanding of the business patterns, market space and future revenues of these companies, so much of the value has actually been priced. The next round is more concerned about where the money will continue to flow。

You just mentioned land, electricity, housing, physical infrastructure, which seems to be the right direction. Because if we think about AI, which is really the most important thing, the answer is increasingly like physical capacity-building. You see XAI, or more precisely SpaceX, which is now on the market, its income core is not the rocket itself, but the AI infrastructure。

The value of this deal with Anthropic, Google has exceeded the sum of Starlink, Starship and the entire satellite operation. There is clearly a great need and great value in this. So the question is, who actually built these things

SpaceX is obviously an answer. Its stock price since last night has reached $230, corresponding to a valuation of about $3.1 trillion. We're going to do a SpaceX issue this week because it's an exaggerating move, just completing the acquisition of Cursor, and now it's worth $3 trillion, and Elon is making more money a day than Warren Buffett's whole career。

Who's gonna get the next bonus

Josh Kale: We're concerned about which companies are best at making this hardware infrastructure and developing the machines. In combination with the direction of Leopold and the overall trend, we feel that the next step in funding will be here. So, Ejaaz, what kind of company does this wheel actually fall on

Ejaaz Ahamadeen:Many will be infrastructure companies that don't sound sexy. The name you have often mentioned in the past month is Marvell. At the Computex Congress in Taiwan, China a few weeks ago, Jensen Huang directly said that this would be the next trillion-dollar company。

And just three months before he said that, NVIDIA had just invested $1.5 billion in Marvell. I don't know if it's an inside deal or a market manipulation, because when he said it, it went up 70 percent。

I THINK IT'S EASY TO JUDGE AI INFRASTRUCTURE RIGHT NOW; BUT IF YOU COMPARE IT WITH THE FINANCIAL CRISIS OF HISTORY, FOR EXAMPLE, IN 2008, THE TASTE OF HIGHLY LEVERAGED, FINANCIAL ENGINEERING AND SYSTEMIC MANIPULATION, THIS ROUND IS NOT ENTIRELY PRESENT。

There are two key differences. First, the products produced today by these companies are really being bought. Neither the Internet bubble nor the financial crisis were so solid real needs. Secondly, subject to the laws of physics, there is no limitless leverage at all, because the system is stuck in human capacity and capacity-building。

Even if you raise more money, you will not be able to complete the data centre quickly enough and expand enough memory chip capacity, nor will you be able to expand the grid, the power lines and the related infrastructure immediately. There are not so many people on the ground, and approval, supervision and procedures are stopping you。

So I think it gives investors an advantage. Now that you know that the hottest chip and shovel sales are too crowded, the money goes to electricity, data networks, such as Astera Labs, and to other related links。

What you really want to think about is when these contracts will come true, when these round mills will actually be built, when SpaceX rockets will send the AI satellite up to the sky, or even when solar energy will start training the AI model。

THE TIMELINE DETERMINES THE PACE OF THE BET. AT LEAST I'M DOING IT MYSELF, AND THIS IS NOT AN INVESTMENT PROPOSAL. I SEE THIS BECAUSE WE HAVE SEEN FOR OURSELVES OVER THE PAST YEAR AND A HALF HOW MONEY GOES FROM PAN-AI STOCKS TO SEMICONDUCTOR AND INFRASTRUCTURE TRANSACTIONS。

Josh Kale:IF YOU KEEP LOOKING DOWN AT THIS COMBINATION, YOU'LL ACTUALLY FIND THAT THE STORY IS CLEARLY WRITTEN INTO HIS HOLDING STRUCTURE. WHAT IS HIS LARGEST CONFIGURATION BY CATEGORY? IT'S POWER AND ENERGY. THE SECOND IS MEMORY, FOLLOWED BY CLOUD AND GPU MINERS, THE MOST SOLID INFRASTRUCTURE。

He wants to hold new cloud manufacturers like CoreWeave, and he wants to hold miners who have switched to cloud power。All he wanted was those physical infrastructure, because he thought it was the real bottleneck。As you mentioned earlier, there are, of course, many more detailed elements, such as physical construction, hardware manufacturing and the construction of the data centre itself, which are extremely difficult。

If that was the biggest bottleneck, it might even be a permit. Then who's going to solve these problems? SpaceX wants to move the data centre into space, Tesla wants to solve the manpower problem with a human robot. But both things are far away. In the short and medium term, there are significant gaps in the way in which Leopold is being held。

Optical modules and fibre-optic advantages

Josh Kale: I would like to add one more detail that we did not start with. For those who want to dig deeper and look for more excess earnings, many of his clues are actually hidden in optical and lower-level technology. Ejaaz, you've been working on this lately. Can you tell us what he's thinking

Ejaaz Ahamadeen:And if you look at these barns on his screen, CoreWeave and Iron are basically the top of the new cloud service. Simplely understood, they're kind of like Amazon Web Services, except that AWS provides cloud services to Internet companies that provide ready-made GPU infrastructure to AI。

They help you get the GPU, the network, the deployment, so that AI doesn't have to worry about the infrastructure at the bottom of it, and it can train the models and make the calculations. CoreWeave and Iron have been one of the largest concentration positions since he built the warehouse, with the highest returns。

It is also noteworthy that he still holds these two companies in the largest positions today. This also points to the problem: in his view, the transaction is far from over. Further, he had privately invested in Core Science, a company that could help release CoreWeave ' s infrastructure supply capacity. In a sense, he's adding another layer of leverage to CoreWeave。

Apart from that, look at companies like Cohen and Lumentum, which are essentially suppliers of fibre-optic and optical connections. If it is to be explained in the simplest terms, there is a need for communication between semiconductors and GPUs, usually through a large number of copper wires。

The problem isWHEN THE GPU IS GROWING IN SIZE, THE COPPER LINE WILL BECOME HOTTER, ENERGY WILL BECOME MORE EXPENSIVE AND INEFFICIENT, AND FIBRE-OPTICS WILL BE THE NEXT COURSE OF UPGRADING。It allows for faster data transmission, more cost-effective, and more money for companies that provide reasoning and training. So you'll findHis bets have always been very infrastructure-oriented, both for optical and electricity-related companies。It may not sound sexy enough, but in my opinion, that's where the money is really falling。

Josh Kale:Copper is also interesting to me, because I've only recently realized how critical it is in the transmission of data at close range。A lot of short-range transmissions with high bandwidth, copper is almost the only material you really want to use. Only when it starts not to apply, such as too far away or too much heat, does it reach fibre-optics, so the market is now in strong demand for copper and fibre-optic combinationsAnd that's why it's interesting to watch the deal。

Copper futures have recently taken a strong toll, essentially because they are needed by everyone. It is the most critical base material in short-range high bandwidth, and fibre-optics is the next step。

From the bottom, the material line has always been interesting. The bottom of all the bottoms is actually what raw materials are needed at the core to acquire intelligence. Copper is one, lithium is one, and there's a lot more. We should really do a separate material topic. Maybe Leopold hasn't reached that level yet, but we can see the next wheel first。

Josh Kale:If you keep going to the bottom, you can even go straight to the copper mine and see how these things are made. But back to the core, I think that the next round of rotation is really going from what looks like smaller bottlenecks to really difficult things, that is, hardware and large data centres。

Whoever has the capacity to build a data centre will make money。We've seen how much SpaceX earns because of how much data needs. Those who have faster access to more data centres and those who have adequate electricity and GPU will make the most money. That's basically the direction Leopold's in now。

Is there a foam

Josh Kale: In conclusion, we don't think we're in the bubble breakup phase. Leopold's position is more like a wheel than a full retreat. So, shouldn't we just keep following him

Ejaaz Ahamadeen:I admit it was the first time I saw his 13F, and my first reaction was that it was too much for this guy to see the world's most valuable, and the demand was in 2029. But now that I see this financing, I'm beginning to think that if NVIDIA continues to raise external debt in the future, and even sells shares in the future, if this trend continues, Leopold may be right again。

If that were the case, his fund could eventually exceed the world ' s top traders and best investment funds. He's really been winning. It's hard to believe。

Josh Kale:But one more thing is important. His past life has almost always been more than anything, and he has never really experienced the test of mass sale. Bill Ackman, we mentioned earlier, made 30 times more returns and lived on the market for 30 years, but it's different。

It would be even more terrible if he could really sustain this growth and learn when to push the key, how to manage the risk, and how to protect himself with a hedge. Now we're beginning to see the prototype of that ability. And the $9 billion empty is not just going to take $9 billion in cash to make it empty, but it's going through options and leverage, not one-to-one nakedness. In any event, it is worth continuing to observe。

Energy is the core bet

Josh Kale: If you pick a stock from his entire portfolio that you want to buy, which one do you want

MY OWN ANSWER IS THE ENERGY UNIT. I'VE ALWAYS WANTED ENERGY BECAUSE EVEN IF THE DEMAND FOR AI SLOWS DOWN, ENERGY ITSELF IS STILL A GLOBAL DEMAND, AND IT'S ONLY GETTING HIGHER。

Even if we don't look at AI at all, we need more energy, more electricity. Companies like Bloom Energy that boost power supply and transmission are the most exciting direction I've ever been in, because they're the ones that look like hedge bets. The single trend that will continue to rise in any situation is our demand for energy, electricity and power, companies that I would most like to do in the long run。

Ejaaz Ahamadeen:My answer is a bit cheating。What I'd like to talk to most of is the companies that Jensen pitches, and at the same time intersects with Leopold's logic. The closest I've got is Marvell. It's not a publicly held company of Leopold, but it's very close to his bets on fibre optics and electricity, and Jensen has already invested $1.5 billion in real money and silver。

I've observed a phenomenon that, as long as Jensen went to a company through NVIDIA, whether it was Intel, CoreWeave or something, it was going up. So I'm probably in here right now. I myself hold some CoreWeave because Jensen and Leopold are extremely good at it。

Josh Kale:Marvell has increased by 270 per cent in the last six months. It's probably a really good rule of experience: people like Jensen, even people like Trump, who have a lot of influence, if they say they're going to buy a stock, you probably should really take a look at it。

In the past, it has been proved on many occasions that such signals often have very large fulfilling space. Both Intel and Marvell show that on the one hand they do understand what they are saying, and on the other hand they are capable of influencing the results of these companies. So this business is really crazy。

I hope it continues. For the time being, it is also likely to continue. At least now we are all too much or optimistic and will continue to judge every day with change。

Josh Kale: Is there anything you want to add to Leopold's portfolio update

Ejaaz Ahamadeen:I'd really like to hear what those who doubt it. If you listen to the analysis that we have just done, you feel that we are completely wrong, or that there is some misinterpretation, and you are welcome to point it out directly。

YESTERDAY I WATCHED THE $25 BILLION FINANCING NEWS THAT I WAS TRYING TO STAB. BUT IT DOES MAKE SENSE TO SEE FINANCIAL LOGIC ALONE。

Why not borrow this cheap, almost risk-free money? It would obviously be more reasonable to expand with someone else's money than to sell your own equity, because that would allow you to retain more future gains。

Original Link