people장 홈📈 Stock Masters주식 유튜브를 Articles로 읽고 Author의 예측을 실제 가격으로 확인합니다 people장 Board 홈요청하기

한국어 · English

DO NOT BUY STOCKS! (Michael Burry’s Final Warning)

The Graham Stephan Show · 2026-06-09 · 유튜브 시청 206,733 · Board Views 7 · International

International

유튜브 썸네일

← 목록

저자: GrahamStephan · 외국인

요약 보기

마이클 버리가 엔비디아 GPU 임대 구조(Valor SPV, Apollo 자금, Athene 연금형 보험 상품)가 회계적으로 부풀려졌고 은퇴자 자금이 위험에 노출됐다고 주장. 스테판은 구조 자체는 합법적이며 엔비디아 매출 5.4B 중 자기투자 1.9B를 제외하면 실질 매출은 약 3.5B일 수 있다고 지적. GPU 감가상각 위험은 있으나 완전 손실 가능성은 낮다고 봄. 최선·현실·최악 시나리오를 제시하며 장기적으로 AI 수요는 성장할 것으로 전망.

다듬은 글

※ 자동 자막 원문(다듬기 전)

01. What's up guys? It's Graham here. So, we got something crazy going on in the stock market that apparently is going to impact a lot of people and it all has to do with fake numbers, Gazi, and uh believe it or not, Michael Burry. It all started with this tweet that I'm going to be referencing. I'll read through it because it went incredibly viral.

02. It's got millions of views and it's quite alarming. So, according to Michael Burry, he's alleged that billions of dollars in Nvidia chips are being hidden off the balance sheets and that American retirees are unknowingly funding the whole thing. A quick side tangent here, but we've all seen this circular financing going on in the stock market.

03. I'm going to put it up on screen here. It's incredibly confusing, but basically one company is like working with another and they're lending each other money and then setting up other companies to lend that company money which then funds other projects and some of it is just like creative accounting where they structure certain deals certain ways for tax advantages.

04. They got to do this. But this deal on the other hand is uh just as confusing. So, let me explain it. They say that Nvidia, who's the world's largest AI chip company, sold 5.4 billion worth of its most advanced GPUs to a company called Valor. It's important to note that Valor's not actually like a real company.

05. It's what these businesses set up is somewhat like a shell so they could distance themselves for tax purposes. Uh they're not putting it on their own books. It's basically like you saying, "Oh, I'm going to create this LLC and uh that LLC technically owns the assets. I don't own the assets, the LLC does." So, in a way, that's what they did.

06. Now, the purpose of this, like I said, is to be buying all these GPUs and Nvidia invested 1.9 billion of its own money directly into this company on top of the sale. Now, in terms of who's using these chips, it's all XAI. It's Grok. So, when you're talking to Grok, what you're really using is like a whole bunch of these GPUs that were set up and funded specifically for this purpose and this is important because again, it all kind of circles back at the end.

07. Now, Burry has alleged that a lot of these numbers are fake. See, the thing is, neither Nvidia nor xAI own those chips on paper. Valor, the shell company, holds legal title. This means that 5.4 billion in GPU assets do not show up in Nvidia's balance sheet as inventory. They also don't show up on xAI's balance sheet as assets.

08. They're invisible to both companies. Again, because they set up this LLC, so to speak, that holds all the assets. And then what happens is really interesting. Nvidia gets to book the 5.4 billion as a completed sale and record it as revenue. xAI gets full use of the chips without owning them. And the risk disappears into a shell company in the middle.

09. Again, it's kind of like me saying that, you know, I have my own assets. I have my own bank account that I use for like personal checking and savings. And then this computer here is actually owned by my LLC, my business. So, this is a business computer that I use specifically for running the business, even though me personally, I could use this computer for business purposes.

10. So, if the computer is not on my books, it's on the business's books. Now, in terms of where some of the risk is for everyday investors, retirees, and people who don't want anything to do with this, but are kind of left without any other choice, Michael Burry says that Valor needed 3 and 1/2 billion in debt to fund this structure.

11. Apollo provided it, and they're one of the largest asset managers on Earth with a trillion dollars under management and 834 billion specifically in private credit. So, what Apollo did is that they raised 3 and 1/2 billion dollars from investors. They packaged it into securities, and then they sold that to a company called Athene.

12. Now, Athene is the insurance company that Apollo had set up, and it sells fixed and indexed annuities, retirement savings products, and all of this to ordinary Americans. That means that when a retiree buys Athene, they believe their money is sitting in safe, stable investments, but that money is now inside the structure funding Elon Musk's AI data center.

13. Now here's where Michael Burry says is the risk. Athena holds $74 billion in reserves, but it's moved $200 billion in assets into a captive insurer based in Bermuda, meaning those assets sit outside of normal US insurance regulation. Now Burry says that every step of this structure is technically legal and publicly disclosed, but the entire thing was deliberately engineered across eight to 12 steps to move credit risk off balance sheets and away from any market pricing.

14. Or in this case, Nvidia books the revenue, Apollo collects the fees, xAI gets the computing power, and retirees are sitting at the bottom of a 16x leveraged Bermuda insurance structure holding $100 billion in assets with no market price that retirees don't know they're getting into. Now again, if you're confused by this whole thing, the setup is that basically xAI needs a whole bunch of GPUs to run.

15. But if they were to go and buy those GPUs, it's very expensive. They don't know if they're going to need to upgrade these chips over time. So what they do instead is they rent them. So what Nvidia does is say, "Hey, we're going to set up a company for you to be able to rent from this company so that you could use all the chips that you need to without actually buying them yourself." So Nvidia and xAI and all these companies get together and they go and they set up this other company that holds the chips and holds everything needed.

16. xAI could rent from that company, Nvidia could fund that company, and then they could bring in outside investors who just want a safe stable return on their money kind of like buying a real estate building except xAI is just renting what's inside that real estate building. And then what they do is once they have that deal structured, let's just say it pays 10% a year for 10 years just as an example, they could then go and package that, sell it off to an insurance company, and then that insurance company goes and they say, "Oh, well now we we this product that's makes 10% a year.

17. We're going to be giving this to people who want a safe, stable return on their money. We could sell it out, and that's where everyday investors get looped in. And if everything goes well, they get a safe, stable return on their money backed by GPU chips in a very high-demand sector. That's really all that's going on here.

18. And the benefit for XAI, they get to rent them. Nvidia gets to make a whole bunch of chips and book a big sale. And investors, again, if everything goes well, just get a return on their money. Now, before I break this down even further, there is a rebuttal that I saw here that I do want to go over because it wouldn't be fair just to show one side and not the other.

19. Now, really quick, I just want to say that when it comes to the market and everything that's going on and rising prices on top of that, one of the places that I feel it personally is just when it comes to food. Like when I go and get a meal, all of a sudden it's 20, 25, $30 for food that just tastes meh.

20. Thankfully though, it doesn't have to be that way. Food does not have to be insanely expensive or taste mediocre. All thanks to our sponsor Cook Unity. No joke, this has been my go-to food for years. And if you're unfamiliar, Cook Unity is a chef-to-you meal delivery platform that connects you with hundreds of fresh, fully prepped meals from award-winning chefs.

21. Even better, but it tastes like this is restaurant-quality food delivered right to your front door. Except with this, you could choose from over 300 meals nationwide every week across more than 25 diets and global cuisines. And there's over 180 chefs on the platform, including Food Network stars, James Beard Award winners, and Michelin starred chefs.

22. For me, this has been one of the easiest ways of just being able to operate more efficiently that I don't have to spend all this time trying to figure out what I'm going to eat. Or even worse, like sometimes throughout the day, I'm just snacking on crap and it ends up as just empty calories. When instead, I could just make one of these.

23. There's no meal prep, there's no thinking, there's no cleanup required. It's ready in a few minutes, and I know I'm eating something healthy. I'm not even kidding. I referred these to friends and family, and everyone raves about them. You could even probably check in the comments with a few people who have tried this out, and like every single person that I've referred to Cook Unity ends up sticking with them, and they love it.

24. My favorites have so far been the roasted salmon in lobster sauce by Chef Andre Mendes, mustard-crusted salmon by Ludo Lefebvre, and the crispy tilapia over Greek salad by Chef Ruben Garcia. So, if you want to see what all the fuss is about, just try it by going to cookunity.com/graham with the link down below in the description, and use the code graham50 at checkout to get 50% off of your first order.

25. Like I said, at 50% off, there's no reason not to try it. It's half off. It is a great deal. That link is down below. They're truly one of my favorite sponsorships I've ever done. Enjoy. Again, that link is down below. Thank you so much. Now, let's get back into the video and talk about what's happening next.

26. So, according to this guy, OX Sammy, he's worked as a big four auditor for a decade. The transaction is real, and the figures check out. All of this was publicly disclosed. But, Michael Burry's thread oversells two things. On Nvidia's revenue, selling it to a special purpose vehicle, like that middleman company, is totally fine.

27. The question is that whether or not control is actually transferred. If the middleman company bears the risks and rewards, Nvidia books the sale legitimately. The real issue, though, is because Nvidia invested in that middleman company $1.9 billion. If they're counting that into $5.4 billion of revenue that they say they received, well, that's kind of like just paying themselves.

28. So, in reality, their revenue should or might be more like 3 and 1/2 billion. Now, you know, again, we're talking five-something billion, three-something billion. It's still a lot of money. But again, if you're counting that as a sale, then that's not quite a sale. It's like me saying that, you know, I sold this coffee to myself.

29. So, I charged myself $10 for this coffee and, you know, I put $5 in. I I can't say that I made, you know, $10 selling this coffee because five of the dollars I just like kind of paid for the coffee, but I owned the coffee. So, it doesn't quite work like that. So, if that is the case, then yes, this could be a bit overstated.

30. Now, when it all comes to this, I got to say as just a dude who's like weirdly obsessed with like deal structure like this that's really complicated that no one understands, a lot of it to me is just like clever accounting where they're trying to get stuff off the books and they're trying to optimize this because the biggest issue that I see is that these chips that are used for computing AI age really fast.

31. It It's kind of like buying a brand new iPhone knowing that 3 years from now it's going to be selling for like 40% of the price. It's going to be way slower. There's going to be all these new features that come out. And with the way things are going right now, 2 years from now like all these chips could be completely irrelevant.

32. And for XAI, they don't want to like buy these chips that are depreciating so quickly cuz they lose value. Now, there is a chance of course that the chips don't advance as quickly as we think and they hold more value, but realistically it's kind of like just buying an iPhone. It's going to be worth less in a few years because guaranteed a better iPhone is going to come out.

33. So, for XAI, it makes sense to want to rent these chips instead. For Nvidia, makes sense that they just want to pump out as many of these chips as possible because it helps their bottom line, too. And for the retirees, you know, yes, they're investing a portion of their pension in something like this.

34. But then the risk really becomes is XAI going to stop paying on these GPUs? Is that middleman company going to go bankrupt? Let's just say then they do. What is the value of those GPUs? 50%? 60%? 70%? That's why I think the chance of a total loss with that is probably not going to happen. Can't be a total loss because the GPUs are probably going to hold some amount of value.

35. Again, it's like you know, the iPhone. Like even the iPhone 5 still has some level of value. Someone's still using the thing. It's not ideal, but it it's still it's not going to sell for zero. It's not like you have to pay someone to take off the iPhone from you. So, it's always going to have value.

36. Realistically, XAI is only going to expand and they're going to keep making money. And so, my guess is that the realistic risk for retirees is there. It's not zero, but it's also not going to be like this scenario where we're we're walking into something that like obviously is going to hit a wall at some point.

37. As of now, we don't know when that wall might may be or if there is even a wall to begin with. You know, also keep in mind that Athena, the insurance company, is also huge. So, them investing 3 and 1/2 billion dollars actually kind of pales in comparison to the amount of assets they have, which is over 400 billion dollars.

38. So, that's why when you look at this in the big picture, it's kind of like investing 1% of your net worth or your investable assets into a deal. And if that 1% doesn't do that well, well, then you got 99% to fall back on. And again, that chance of the 1% going to zero probably not going to happen. It's probably going to you know, worst case maybe lose half, but even if it loses half, you got 99.5% of your other investments that should be working for you.

39. Like this could be a bucket in the bigger category, and then you have a whole bunch of other stuff that you could use if you need it. That's why I think that we have three realistic scenarios for this. We have we'll call the best-case scenario is that XAI makes a ton of money. They rent these chips for a very, very long time.

40. Investors make an even better return on their money. This whole middleman company keeps growing, gives investors even better returns. Twitter wants, you know, even more GPUs and the whole deal just continues to grow. That's the best case scenario. The realistic scenario is that these GPUs slowly lose their value, but like say I just keep paying them.

41. And that just means that investors get the returns that they've been promised, that they're stated, that you know, XAI has enough money, enough users to keep this deal going. Uh AI demand is going to remain just as robust over the next few years and really nothing happens. It's a bit of a concern. Michael Burry kind of shocks the market, but long-term everything just kind of turns out as planned.

42. The worst case scenario is that these chips lose a ton of value because they're progressing so fast. Uh XAI is not making as much money as they thought they would. They have trouble paying the rent, uh leasing these machines out. Maybe they stop and then all of a sudden there's a glut of these GPUs that maybe a lot of people don't want to take over.

43. Maybe they don't want them. So, XAI stops paying the rent on those and then the retirees, the investors, the pensions, all of a sudden they take a hit and a portion of their investment loses money. And again, as long as the AI train keeps going, they're going to keep expanding. And just the way I see it, AI demand is probably going to just continue growing from now.

44. Even if the market seems to fall off a cliff or something, I just think development is just getting started. This is probably going to keep going, but that is the biggest risk that I see is just like if everything goes to crap, then a portion of the fund also goes to crap and then it also depends what else they're investing in beyond this that I have no clue on.

45. So, in terms of all of this, please let me know what you think down below in the comments section. I will do my best to read and respond to as many of you as I can. And also, don't forget hit the like button, subscribe. Don't forget to check out our sponsor Cook Unity down below in the description. I genuinely love the food.

46. They are so good. So, if you want to make me happy and your stomach happy, check out Cookie Unity, get 50% off with the code gram50 at checkout. You may as well do that for the first one cuz it's 50% off, so just enjoy that. Thank you so much and until next time.

원문 보기 · 글 내려 주세요 신청

← List으로
공유페이스북밴드

이 Articles에 대해 요청(숨김) 수정 요청

Comments 0