YouTube2Text

Which Big AI Company Goes Bankrupt First? — Transcript

by Nanalyze · 2,115 words · 326 segments · language en · Watch on YouTube

Full transcript

  1. 0:00It's Nebius, okay? Neo Cloud Nebius will
  2. 0:03be the first AI stock to go bankrupt.
  3. 0:06Nah, nah, it's rage bait. I'm joking. Or
  4. 0:08am I?
  5. 0:09From Anthropic to Coreweave to Oracle to
  6. 0:12Manebius, we're going to look at the
  7. 0:14first AI stock that's likely to go
  8. 0:17bankrupt.
  9. 0:18is that so important? Because
  10. 0:20bankruptcies are the canary in the coal
  11. 0:23mine for any sort of financial crisis.
  12. 0:26Most you young bucks out there have no
  13. 0:28idea what real blood on the street looks
  14. 0:30like, but don't you worry. I'll talk
  15. 0:32about last week's blood bath in a
  16. 0:35second, but I have a very important
  17. 0:37point to make here. We believe that it's
  18. 0:39highly likely in the case of the AI
  19. 0:42bubble bursting that it's accompanied by
  20. 0:45large bankruptcies or even driven by
  21. 0:48large bankruptcies. Why? Because of
  22. 0:51something called circular financing, and
  23. 0:54it's getting worse and worse as the days
  24. 0:56go on. Nvidia is now being called the
  25. 0:59bank of AI for good reason. Why it's
  26. 1:02going to discuss that a little bit
  27. 1:03later, but let's get on to talking about
  28. 1:06the elephant in the room. Something
  29. 1:08terrible happened last week, and I can't
  30. 1:10stop thinking about it.
  31. 1:12Sydney Sweeney was filmed crying in
  32. 1:14character during the filming of the
  33. 1:16House Maid's Secret last week in New
  34. 1:18York. I hope that poor woman's okay.
  35. 1:21Another equally traumatic event that
  36. 1:23happened last week was the AI stock
  37. 1:26blood bath.
  38. 1:28Blood bath.
  39. 1:29As always, we kept our Analyzed Premium
  40. 1:32subscribers up to date on the situation
  41. 1:34and our Analyzed market open, but for
  42. 1:37the rest of you, take a look at this
  43. 1:39chart here. This is the Nasdaq 100 over
  44. 1:42the past five years, and it's not really
  45. 1:45that bad when you look at it in the
  46. 1:47context of a longer time frame, is it?
  47. 1:50Now yes, I know the crisis mainly
  48. 1:52affected high beta AI stocks. So take a
  49. 1:56look at this even better slide. These
  50. 1:58are the four AI stocks that hit the
  51. 2:00hardest during last week's blood bath.
  52. 2:03Sandisk, AMD, Micron, and Monebius. You
  53. 2:08know, this always gets me. Nobody bats
  54. 2:10an eyelid when Sandisk soars over 2,000%
  55. 2:14in less than two years. But heaven
  56. 2:17forbid it should drop 20% giving it an
  57. 2:20year-to-date return of only 200%.
  58. 2:23We need to stop slobbering over
  59. 2:26short-term price changes like a bunch of
  60. 2:28monkeys and focus on what systemic
  61. 2:31changes may have happened here. Most
  62. 2:33recently, there are concerns that the
  63. 2:36Democratic mid-term wins in the United
  64. 2:38States are going to slow down data
  65. 2:41center construction because of things
  66. 2:43like regulation, moratoriums, and the
  67. 2:45like. And probably most tangible would
  68. 2:47be the signs that are showing access to
  69. 2:50easy capital for AI expansion is
  70. 2:53starting to tighten.
  71. 2:55Evidence? Well, some AI IPOs are being
  72. 2:58pulled back due to lack of interest from
  73. 3:00investors like Firmus in Australia. You
  74. 3:03have Microsoft pulling back on some of
  75. 3:05their data center investments because of
  76. 3:07delays. And most importantly, I think is
  77. 3:09that new debt is becoming more costly
  78. 3:12because it's more costly to insure it.
  79. 3:15This is something we measure with tools
  80. 3:17called credit default swaps. This is a
  81. 3:20really cool metric that any investor out
  82. 3:23there can use to monitor the situation.
  83. 3:26So here we've listed out some of the
  84. 3:28hyperscalers and I'll just show you
  85. 3:30really quickly how to read this. So
  86. 3:32let's take Microsoft as an example here.
  87. 3:34It costs $5.60
  88. 3:38per year to insure Microsoft against
  89. 3:41default over the next five years. And
  90. 3:44for most large AI companies out there,
  91. 3:47you can find these instruments. What you
  92. 3:49can then do is see what the cost of
  93. 3:52these credit default swaps has been
  94. 3:54doing over time. For example, how far
  95. 3:57are they off their all-time high, right?
  96. 3:59If they're close to the all-time high,
  97. 4:01shows you that risk is still pretty high
  98. 4:03or perceived as being high. And we can
  99. 4:06expand this out to other names and of
  100. 4:09course the name that's been in the news
  101. 4:11lately, Coreweave. So it costs $110 a
  102. 4:15year to protect against default for
  103. 4:17$1,000 in Coreweave debt. That's very
  104. 4:20high. And when we look at the change of
  105. 4:22these credit default swaps over the past
  106. 4:24month, you can see they've all been
  107. 4:26rapidly appreciating. In other words,
  108. 4:28this debt is getting riskier as time
  109. 4:30goes on. Nvidia, look at that
  110. 4:33appreciation in the risk for Nvidia's
  111. 4:35debt. What's going on there? Well, why
  112. 4:37it's going to talk a bit more about that
  113. 4:39in a second. But the other thing that we
  114. 4:41can do with the credit default swap
  115. 4:43that's quite interesting is that you can
  116. 4:45calculate an implied probability of
  117. 4:48default. In the case of Coreweave,
  118. 4:49that's about a 60% likelihood they
  119. 4:52default over the next five years. That's
  120. 4:55crazy, right? And when you consider the
  121. 4:57likelihood of default, there can be
  122. 5:00problems that are not so obvious to
  123. 5:02investors. It's not just a matter of
  124. 5:03looking at their debt and then their
  125. 5:06cash flows to see if they can make their
  126. 5:08interest payments. You also need to
  127. 5:09consider things like contractual
  128. 5:11commitments, covenants for their debt
  129. 5:14that may be quite obscure, off-balance
  130. 5:16sheet debt that you can't even see. And
  131. 5:19it's the job of credit ratings agencies
  132. 5:22to look through all this stuff. For a
  133. 5:24good number of AI stocks out there, we
  134. 5:27can look to credit ratings agencies for
  135. 5:30signs of trouble. We can see how they
  136. 5:32change their outlook over time. Now, I
  137. 5:34know what a lot of you are thinking
  138. 5:36right now. Too many Sydney Sweeney jokes
  139. 5:38lately. How about not enough Sydney
  140. 5:41Sweeney jokes?
  141. 5:43>> How about that? How about that? How
  142. 5:44about that?
  143. 5:45>> Now, I know what a lot of you are
  144. 5:46thinking. Credit ratings agencies
  145. 5:48weren't of any use during the 2008
  146. 5:51financial crisis. And you're exactly
  147. 5:53right. A sufficiently sized cataclysmic
  148. 5:57crisis doesn't give investors time to
  149. 6:00respond. That's the entire point of it.
  150. 6:02Because of the circular nature of the AI
  151. 6:05economy, a crash is likely to be sudden
  152. 6:07and powerful. Neo clouds are going to be
  153. 6:10the first to fail because they rely on
  154. 6:13debt for their lofty valuations. You see
  155. 6:16that happening right now. This simply
  156. 6:18captures current sentiment because it
  157. 6:20becomes more difficult for them to raise
  158. 6:22money. Those growth stories start to dry
  159. 6:24up.
  160. 6:25It comes down to this. The core business
  161. 6:28for these neo clouds is not generating
  162. 6:30positive operating cash flows or
  163. 6:32earnings sufficient to service their
  164. 6:35debt. The debt servicing is simply
  165. 6:37coming from customer advances or new
  166. 6:40financing or even selling equity.
  167. 6:42Whenever we bring this up that neo
  168. 6:44clouds are likely to be the first to
  169. 6:47fail, nobody disputes that. They simply
  170. 6:50focus on reasons why the AI trade's
  171. 6:52going to keep going. This is the compute
  172. 6:54demand story. Fair enough, but for neo
  173. 6:57clouds in particular, when they can't
  174. 6:59access capital, that growth story dries
  175. 7:02up. Hyperscalers on the other hand, they
  176. 7:04have very strong cash flows that they
  177. 7:07can fund build outs with. They don't
  178. 7:09need to take on debt even though some
  179. 7:11are. And hyperscalers are actually major
  180. 7:14customers for some neo clouds and that's
  181. 7:17a real problem, right? You never want to
  182. 7:19have your competitor as your customer.
  183. 7:22We talked about this in an earlier piece
  184. 7:24this year which looked at problems in
  185. 7:26the neo cloud sector. Now since that
  186. 7:28piece, neo clouds have lost half their
  187. 7:31value except for Nebius which has a
  188. 7:33stronger balance sheet and is more
  189. 7:35resilient to turmoil. The first AI
  190. 7:38company to go bankrupt is likely to be a
  191. 7:41neo cloud and it's likely to be
  192. 7:43CoreWeave because they're the most
  193. 7:45likely to struggle when it comes to
  194. 7:47refinancing their debt or paying their
  195. 7:49interest obligations. Of the neo clouds,
  196. 7:52Nebius is probably the best positioned
  197. 7:55at the moment because of their strong
  198. 7:56balance sheet. So ask yourself this, why
  199. 7:59did this recent bloodbath have you so
  200. 8:01worried? Is it because you're
  201. 8:03overexposed to the AI trade? Well, yes,
  202. 8:06because everyone is.
  203. 8:08Simply holding the S&P 500 gives you
  204. 8:10anywhere from 35 to 45% exposure to this
  205. 8:14tech sector and large tech companies,
  206. 8:17which aren't all classified under the
  207. 8:19GICS technology sector. So let's talk a
  208. 8:22little bit more about Nvidia, which is
  209. 8:24presently our largest position that
  210. 8:27we've been holding for over a decade
  211. 8:28now.
  212. 8:29>> Putting aside Jensen Huang's recent
  213. 8:31scandal, today I want to look at some
  214. 8:32concerns raised by Hunter Brook, which
  215. 8:34is an investigative journalism company.
  216. 8:36They came up with this really cool piece
  217. 8:37called the Comprehensive Guide to the
  218. 8:39Bank of Nvidia, which made some really
  219. 8:40cool points. Essentially, they're saying
  220. 8:42that Nvidia has evolved from just a
  221. 8:44chip-making company to a company that
  222. 8:46supports the entire AI industry. There
  223. 8:49have been these concerns going around
  224. 8:50about circular financing, right? You've
  225. 8:51seen the giant bubble chart where Nvidia
  226. 8:54supplies a customer, they have an
  227. 8:56investment in that customer. Hunter
  228. 8:57Brook has an even more complex and
  229. 9:00comprehensive chart here that shows
  230. 9:02Nvidia with relationships to the
  231. 9:03hyperscalers, Microsoft, Google, Amazon
  232. 9:06Web Services, to the LLM providers,
  233. 9:08Meta, Anthropic, OpenAI, to the likes of
  234. 9:11Tesla and SpaceX, to the likes of AI
  235. 9:13startups, Neo Clouds, Nebius, CoreWeave.
  236. 9:15All of these are interconnected, and
  237. 9:17they're saying that it's not circular
  238. 9:18financing, it's reflexive. So the
  239. 9:21reflexive property in math, going back
  240. 9:23to our high school math days, means that
  241. 9:26one thing equals itself, meaning that
  242. 9:29the value of Nvidia is dependent on the
  243. 9:31value of its customers, of its
  244. 9:32investments, of its suppliers. So
  245. 9:34lately, all of the value of these
  246. 9:37investments have been growing together.
  247. 9:39However, when the opposite happens, when
  248. 9:42demand for AI compute begins to
  249. 9:44underwhelm supply of AI compute, right?
  250. 9:46Because we're in a massive build out
  251. 9:48right now, the opposite is going to
  252. 9:49happen and you're going to see the
  253. 9:51reflexive financing in action. Let me
  254. 9:53give you an example. Nvidia has an
  255. 9:55investment in OpenAI. OpenAI is a
  256. 9:58customer of Nvidia.
  257. 9:59Nvidia has an investment in CoreWeave.
  258. 10:01CoreWeave is a customer of Nvidia.
  259. 10:03OpenAI has an investment in CoreWeave.
  260. 10:05There's this loop, and Hunter Brook
  261. 10:07calls it orbital financing as opposed to
  262. 10:09circular financing, which implies some
  263. 10:11motion throughout this orbit. As soon as
  264. 10:13one component of this orbit, of this
  265. 10:15reflexive financing loop starts to
  266. 10:18struggle, it impacts the rest of the
  267. 10:20group.
  268. 10:20>> So far, we've only talked about public
  269. 10:23companies going bankrupt, but honestly,
  270. 10:25a overly hyped private company going
  271. 10:27bankrupt can be a catalyst that pops the
  272. 10:30AI bubble as well. So I think right
  273. 10:32behind Neo Clouds being the most likely
  274. 10:35to go bankrupt first, you probably have
  275. 10:38startups that have overextended
  276. 10:40themselves. And a good reason for that
  277. 10:42is simply just look at how the venture
  278. 10:46capital community is just throwing money
  279. 10:48at AI hand over fist. No business model,
  280. 10:51no customers, no website, no problem.
  281. 10:54Here's a billion dollars. Now, one other
  282. 10:56thing I recently saw that's quite
  283. 10:58interesting is Polymarket, so that's a
  284. 11:01prediction market, says there's a 27%
  285. 11:04chance that the AI bubble bursts before
  286. 11:07June of 2027. Well, immediately you ask,
  287. 11:10what does that mean? Well, it's quite
  288. 11:13interesting. So they say that three at
  289. 11:16least of these events need to happen in
  290. 11:19a 90-day window for that bet to resolve
  291. 11:22to AI bubble has burst.
  292. 11:25They are Nvidia down 50% from an
  293. 11:27all-time high, Sox, the semiconductor
  294. 11:30ETF down 40% from an all-time high,
  295. 11:34Open AI or Anthropic declaring
  296. 11:36bankruptcy,
  297. 11:38Open AI being acquired,
  298. 11:40H100 GPU rental prices falling to a
  299. 11:44dollar or lower for five consecutive
  300. 11:46days,
  301. 11:47or any major AI hardware supplier seeing
  302. 11:51their price, their stock price collapse
  303. 11:5350% or more. And you can see the list of
  304. 11:56names here. Again, three or more of
  305. 11:58these events need to happen during a
  306. 12:0090-day window. So by all measures, the
  307. 12:03AI bubble is not bursting at all. This
  308. 12:06is just normal price volatility. One or
  309. 12:09more bankruptcies is major bankruptcies
  310. 12:12is likely to accompany any real AI
  311. 12:15crisis. And the first company to go
  312. 12:17bankrupt is likely to be CoreWeave based
  313. 12:19on what everything points to today. Just
  314. 12:22remember, no matter how strong demand is
  315. 12:25for AI compute, there are always unknown
  316. 12:28unknowns that can happen, right? Think
  317. 12:30like a COVID-style moratorium on AI
  318. 12:33because that whole cybersecurity
  319. 12:35situation gets out of hand. That's why
  320. 12:37you should always be aware of how much
  321. 12:39tech risk you have. And for those of you
  322. 12:42thinking about doubling down on Neo
  323. 12:44Clouds, you ought to give this next
  324. 12:46piece a watch which talks about some of
  325. 12:48the risks in that space. Thanks so much
  326. 12:50for taking the time to watch this today.

About this transcript

This page contains the full transcript of Which Big AI Company Goes Bankrupt First? by Nanalyze, generated from the public captions YouTube serves with the video. The transcript has 2,115 words across 326 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.