YouTube2Text

The AI bubble just burst — Transcript

by Casey Simpson · 8,378 words · 1,301 segments · language en · Watch on YouTube

Full transcript

  1. 0:00It finally happened. The AI bubble
  2. 0:02burst. If you have any money in the
  3. 0:03stock market or if you've ever used AI,
  4. 0:06you're going to want to hear this.
  5. 0:07>> The AI bubble is not going to burst. It
  6. 0:10already has.
  7. 0:11>> The two most popular indexes of the US
  8. 0:13stock market are the S&P 500 and the Dow
  9. 0:16Jones. They both try to broadly capture
  10. 0:18the US economy, but the biggest
  11. 0:20difference is that the S&P 500 has a lot
  12. 0:23more tech and AI stocks. Over the past
  13. 0:26few months, the Dow Jones has continued
  14. 0:28to rise, while the AI and tech exposed
  15. 0:31S&P has actually gone down. That should
  16. 0:34have been our first warning sign that
  17. 0:35something was wrong. The S&P went down
  18. 0:37because tech stocks did so badly that
  19. 0:40they erased the gains made by every
  20. 0:42other kind of stock in the index. The
  21. 0:44next major warning was that for the
  22. 0:45first time in Google's 20-year history
  23. 0:49as a company, they lost money. In their
  24. 0:51last earnings report, they posted a
  25. 0:53negative cash flow of $6 billion.
  26. 0:57>> I cannot overemphasize how
  27. 0:59earth-shattering this is. Because if
  28. 1:00Google can't make money on AI, nobody
  29. 1:03can. But also, this whole thing was just
  30. 1:05inevitable because it was just very
  31. 1:07basic math. As of this recording, $2.4
  32. 1:10trillion
  33. 1:11has been spent on AI collectively.
  34. 1:15And of all of the AI companies, do you
  35. 1:17know how much profit they have made off
  36. 1:19of that $2.4 trillion?
  37. 1:22$50 billion. Or in other words, AI has
  38. 1:26collectively lost $2.35 trillion.
  39. 1:30So yeah, the AI bubble is already burst.
  40. 1:32The only reason that we're not treating
  41. 1:34it like it already burst is because all
  42. 1:36of these companies that are basically
  43. 1:38passing the same couple billion dollars
  44. 1:40between each other to try to keep this
  45. 1:42whole machine running
  46. 1:44are pretending that the bubble didn't
  47. 1:45burst. Because they know that the minute
  48. 1:47it is common knowledge the bubble burst,
  49. 1:49they will all go out of business. OpenAI
  50. 1:52will go out of business. Anthropic will
  51. 1:54go out of business. Grok will probably
  52. 1:55get absorbed into one of Elon's other
  53. 1:57weird companies and then slowly killed
  54. 1:59off. And you're probably going to see a
  55. 2:00giant overhaul and deprioritization of
  56. 2:03Gemini from Google in like the next year
  57. 2:06or two. In fact, in some ways you're
  58. 2:07already seeing it because Google just
  59. 2:08recently changed all their enterprise
  60. 2:10plans to be way less valuable. And the
  61. 2:13last domino to fall, of course, will be
  62. 2:15Nvidia because Nvidia is the only one
  63. 2:17that's actually making money on this
  64. 2:19whole thing. And eventually, when all of
  65. 2:20the companies that thought they were
  66. 2:22going to make money on this whole thing
  67. 2:23by giving money to Nvidia go under,
  68. 2:26Nvidia will go under, too.
  69. 2:27>> But then, the big moment came. The
  70. 2:29moment so many people have been warning
  71. 2:30about for years now. The Korean version
  72. 2:33of the Dow Jones, the index that
  73. 2:34measures how the entire stock market is
  74. 2:36performing, is called the Kospi. For the
  75. 2:38past 20 years, the Kospi has seen
  76. 2:40extremely slow growth on the rate of 2
  77. 2:43to 3% a year. But in just the past year,
  78. 2:45the Kospi shot up 272%
  79. 2:50For context, in the same period, the
  80. 2:52American Dow Jones went up 17%. Korea
  81. 2:55clearly pulled ahead of us in the AI
  82. 2:56bubble. And the bigger they come, the
  83. 2:58harder they fall. Less than 1 week ago,
  84. 3:01on July 23rd, the Korean AI bubble
  85. 3:03burst. Their entire stock market
  86. 3:06collapsed almost 40% and it's continuing
  87. 3:09to fall as I'm recording this video. In
  88. 3:11Korea, just like in America, a lot of
  89. 3:13investors invest on margin, meaning they
  90. 3:15invest with money they don't have. The
  91. 3:17bank will match them dollar for dollar,
  92. 3:19sometimes $2 for every dollar they have.
  93. 3:21Like if you have $10,000, you can invest
  94. 3:23the bank will give you another 10,000 of
  95. 3:25borrowed money to invest. Sometimes
  96. 3:27another $20,000 of money to invest. Now,
  97. 3:29you have to understand this. Because the
  98. 3:31Korean stock market fell so
  99. 3:33precipitously, 1.2 million Korean
  100. 3:37households lost so much money that they
  101. 3:40no longer had the original amount of
  102. 3:42money that qualified them for the loan
  103. 3:44in the first place. So, let me just
  104. 3:46illustrate this. Let's say you have
  105. 3:47$10,000 and you invest it in the Kospi.
  106. 3:49And the bank gives you an additional
  107. 3:50$10,000 to invest as well. Then your
  108. 3:53entire investment goes up 50%. Now
  109. 3:55you've gained $5,000 of your money and
  110. 3:57$5,000 of the bank's money. Well, this
  111. 3:59stock market collapsed so hard that not
  112. 4:01only did investors lose all the money
  113. 4:03they borrowed, but they lost so much of
  114. 4:05their original capital that the bank
  115. 4:07said, "You no longer have enough money
  116. 4:09to qualify for that extra $10,000 we
  117. 4:12gave you." So if they started with
  118. 4:13$10,000, if they lost their original
  119. 4:1510,000, but now with a net worth of
  120. 4:17zero, they're on the hook to pay back
  121. 4:20the 10,000 they borrowed. 1.2 million
  122. 4:22households, almost a 20th of the entire
  123. 4:25Korean working population, woke up to a
  124. 4:27message saying their account was shut
  125. 4:29down. You lost it all. Older Koreans
  126. 4:31lost their entire life savings. People
  127. 4:33started filing for divorce. People
  128. 4:35started pulling their kids out of
  129. 4:36school.
  130. 4:37>> This is what local Koreans here in Korea
  131. 4:39are sharing on social media. It's
  132. 4:41actually pretty funny when you see this
  133. 4:42stuff. So basically, this is like, you
  134. 4:45know, a guy asking his friend, "Hey,
  135. 4:46what's going on?" And the sheet's all,
  136. 4:49you know, wet from crying.
  137. 4:51They showed the headlines of the market
  138. 4:52and stuff like that. People have a sense
  139. 4:54of humor about this stuff. Someone was
  140. 4:55posting that, you know, the market's
  141. 4:57down 10%. Some stocks are down 12%, you
  142. 5:00know, etc. And that's his lunch for the
  143. 5:02day. All you can afford is rice and
  144. 5:03dried seaweed.
  145. 5:07>> [screaming]
  146. 5:08>> Here's another one. This is more of a
  147. 5:10joke where you're trying to do CPR on
  148. 5:12the stock market, right? So
  149. 5:14first you're trying to give it
  150. 5:15mouth-to-mouth to get the stock to go
  151. 5:17up. Now you're giving the
  152. 5:18>> [laughter]
  153. 5:19>> like a CPR for This is a joke, but
  154. 5:22essentially so many people are in the
  155. 5:23markets right now. Everyone's watching
  156. 5:24it, so they're putting all these sort of
  157. 5:26things on here. Then you go to the the
  158. 5:28KOSPI building and start praying for it
  159. 5:30to go up.
  160. 5:31>> is supposed to represent the entire
  161. 5:33Korean economy, but just like in
  162. 5:34America, the index was over 50%
  163. 5:38AI and tech stocks. 42% of the entire
  164. 5:41index was just Samsung, a company
  165. 5:43heavily investing in AI. So when Samsung
  166. 5:46and a couple other companies like
  167. 5:47Broadcom and Micron posted weaker
  168. 5:49financials than people were hoping for,
  169. 5:51people realized this is not going our
  170. 5:53way.
  171. 5:54And they started selling stocks, and
  172. 5:56then more people started selling it, and
  173. 5:57then it became an entire run, and then a
  174. 5:59crash. Even the CEO of Micron sold $37
  175. 6:02million of his own stock. If the Koreans
  176. 6:04got out, and the CEOs are getting out,
  177. 6:06shouldn't we Americans take notice?
  178. 6:08Because one of the few reasons the Kospi
  179. 6:09crashed more than American indexes have
  180. 6:11is because they had so much of a higher
  181. 6:13AI tech concentration. But that doesn't
  182. 6:15mean we're immune. A lot of people think
  183. 6:17we're next because we're doing
  184. 6:18everything that South Korea did. We're
  185. 6:20just a few steps behind. We're literally
  186. 6:22putting our money in the same companies
  187. 6:24South Korea did. But the stakes are so
  188. 6:25much higher for a US crash. For one
  189. 6:27reason, the entire world economy depends
  190. 6:29on the US economy. But another reason is
  191. 6:31that the US has literally never had more
  192. 6:34of its economy dependent on the stock
  193. 6:36market than today. There's more money
  194. 6:38and more people invested in the stock
  195. 6:40market now than at any point in history.
  196. 6:42Literally most Americans own part of the
  197. 6:44stock market. In fact, analysts report
  198. 6:46that the AI investment bubble is 17
  199. 6:50times larger than the dot-com bubble and
  200. 6:53four times larger than the 2008 mortgage
  201. 6:55crisis. When what happened in South
  202. 6:57Korea happens here, it will destroy the
  203. 7:00world economy. And what's even more
  204. 7:02insidious is the exact investment
  205. 7:04patterns that took down Korea are
  206. 7:06happening in the US. Just like in the
  207. 7:08US, housing became totally unaffordable
  208. 7:11in Korea. So, Korean young people who
  209. 7:13felt like they had nothing to lose
  210. 7:14started gambling in the stock market
  211. 7:16because they figured they're not going
  212. 7:17to get rich by working a job. This gave
  213. 7:19AI companies tons of extra money that
  214. 7:21they couldn't actually deliver on. And
  215. 7:23remember how we established that the
  216. 7:24Koreans were borrowing money so they
  217. 7:26could invest even more? We're doing that
  218. 7:28in the US. Millions of Americans are
  219. 7:30invested in the stock market not only
  220. 7:31with their own money, but with borrowed
  221. 7:34money. And they're doing it to chase a
  222. 7:35higher return. Think about it. If the
  223. 7:37stock market goes up 1%, your money goes
  224. 7:39up 1%. But if you double the amount of
  225. 7:42money you have invested thanks to a
  226. 7:43loan, when the stock market goes up 1%,
  227. 7:46your money goes up 2%, and that sounds
  228. 7:48wonderful until the AI bubble bursts. If
  229. 7:51the AI bubble bursts and goes down 20%,
  230. 7:53then you'll lose more than 40%, or 60%
  231. 7:56if you're triply leveraged. And if the
  232. 7:58AI bubble goes down just a third, and
  233. 8:00you have 3x leverage, you'll lose every
  234. 8:03dollar you have. It already happened to
  235. 8:05thousands in Korea. But now, I want to
  236. 8:08show you the most sobering statistic
  237. 8:10about the entire situation. At its peak,
  238. 8:13the South Korean economy had only 1.3%
  239. 8:17in margin investing, meaning borrowed
  240. 8:19investing. In the US today,
  241. 8:21it's 4.7%. It's actually higher than
  242. 8:23that because that doesn't count
  243. 8:25leveraged mutual funds or leveraged
  244. 8:27stock options. The figure before the
  245. 8:29dot-com bubble was 2%. Before the
  246. 8:31housing crisis, it was country than
  247. 8:33South Korea? Yes, but how far do we want
  248. 8:36to push
  249. 8:39that? We have 50 times the amount of
  250. 8:40borrowed money in the stock market than
  251. 8:42Korea ever did. The only reason the US
  252. 8:44economy is doing so well right now is
  253. 8:46because of the Magnificent Seven, seven
  254. 8:48big tech companies who are spending
  255. 8:50trillions of dollars on AI and data
  256. 8:53centers. Those seven companies alone are
  257. 8:55more than a third of the S&P 500. S&P
  258. 8:58500, more like S&P 7.
  259. 9:00>> If anyone tries to tell you that the AI
  260. 9:02bubble is not an AI bubble, just show
  261. 9:03them the new MIT article that shows that
  262. 9:0695% of businesses who implemented
  263. 9:08generative AI pilots are failures. And
  264. 9:12then show them that the Google CEO
  265. 9:13recently said in an exclusive interview
  266. 9:16that their search engine is more
  267. 9:17accurate than their AI systems, and that
  268. 9:20you shouldn't trust these things
  269. 9:21blindly. And then show them that Nvidia
  270. 9:24still hasn't paid OpenAI for their $100
  271. 9:27billion deal because Nvidia says that
  272. 9:30they wrote a letter to intend to pay,
  273. 9:33not an actual deal.
  274. 9:35>> Mr. IRS agent, I never agreed to pay my
  275. 9:36taxes. I agreed to intend to pay my
  276. 9:38taxes.
  277. 9:40>> [laughter]
  278. 9:41>> I'm going to try it. And at this point,
  279. 9:42you must be thinking, "Casey, surely
  280. 9:45these tech billionaires have a plan.
  281. 9:47Surely we're not trusting the entire US
  282. 9:49and world economy to them based on
  283. 9:51vibes, right?"
  284. 9:54Right?
  285. 9:54>> no current plans to make revenue. We
  286. 9:56have no idea how we may one day generate
  287. 9:58revenue. Um we have made a soft promise
  288. 10:01to investors that once we've build this
  289. 10:03sort of generally intelligent system,
  290. 10:05um
  291. 10:06basically, we will ask it to figure out
  292. 10:08a way to generate an investment return
  293. 10:10for you.
  294. 10:13>> What is this timeline? Bro said our plan
  295. 10:15to monetize AI
  296. 10:17is to ask AI.
  297. 10:19>> Holy hell, everybody was laughing and
  298. 10:21he's like, "Why are you laughing? I have
  299. 10:23no plan."
  300. 10:24>> I have no plan.
  301. 10:25>> AI's going to crash and Sam Altman's
  302. 10:27going to go on ChatGPT, like, "Hey, uh
  303. 10:29how do I fix this?"
  304. 10:30>> In fact, OpenAI was planning on becoming
  305. 10:33a publicly traded company as early as
  306. 10:35September and asking for a $1 trillion
  307. 10:38valuation, even though their annual
  308. 10:40expenses were 26 billion and their
  309. 10:42annual revenue was only 13 billion. So,
  310. 10:45when OpenAI's CFO warned Sam Altman,
  311. 10:49"Those numbers could not possibly
  312. 10:51justify a $1 trillion IPO," do you know
  313. 10:54what he did? He stopped talking to her.
  314. 10:57He stopped talking to his company's CFO.
  315. 10:59He excluded her from financial meetings.
  316. 11:02He excluded the chief financial officer
  317. 11:04from financial meetings.
  318. 11:05>> I'm sure this guy is cool-headed enough
  319. 11:07to handle the entire US economy
  320. 11:09responsibly. Only 16% of households with
  321. 11:12internet pay for any kind of AI. And IBM
  322. 11:15reported that only 25% of AI initiatives
  323. 11:18actually delivered the return on
  324. 11:20investment they promised. And where AI
  325. 11:22is productive is in very narrow
  326. 11:24circumstances. AI benefited companies
  327. 11:27financially in writing code and customer
  328. 11:30for automation. And lost them money on
  329. 11:32literally everything else. Because all
  330. 11:35AI does is compile data and streamline.
  331. 11:38It makes sense that it can program well,
  332. 11:41but it can't think. It can't actually be
  333. 11:42creative. It's philosophically
  334. 11:44impossible. Sorry, Richard Dawkins. But
  335. 11:46AI companies have hyped saying that AI
  336. 11:48will be profitable in every domain. And
  337. 11:50that's how they raised so much money.
  338. 11:52But now people are realizing they bought
  339. 11:54a pig in a poke. In medieval Europe,
  340. 11:56pigs were a hot commodity. Unscrupulous
  341. 11:59merchants would sell their finest young
  342. 12:01pigs to a poor farmer in a bag called a
  343. 12:03poke. However, when the farmer got home,
  344. 12:06he would realize that instead of a pig,
  345. 12:08the merchant just sold him a dead cat.
  346. 12:10And now, when referring to a product
  347. 12:12that was sold as much more valuable than
  348. 12:15it really is, we call it a pig in a
  349. 12:17poke. These AI stocks are not the pig we
  350. 12:20were promised. I guess that makes us the
  351. 12:22farmers. At the very end of last year,
  352. 12:23the CFO of Meta said there was so much
  353. 12:26demand for their AI products, there's no
  354. 12:28way they could possibly build
  355. 12:30data centers. Now, just a few months
  356. 12:32later, Meta is selling those very same
  357. 12:35data centers to other AI companies just
  358. 12:38so they can stay solvent. There's not
  359. 12:40enough demand for Grok either. So Elon
  360. 12:42Musk is having to sell his data centers.
  361. 12:44All this decreased demand has forced
  362. 12:46American AI companies to raise their
  363. 12:48prices. So now it's creating an even
  364. 12:51worse situation because companies who
  365. 12:52still use AI are switching to the
  366. 12:54cheaper Chinese AI programs. Airbnb,
  367. 12:57DoorDash, Coinbase, Uber, Spotify have
  368. 12:59all switched from American AI companies
  369. 13:02to Chinese ones. In fact, Pizza Hut is
  370. 13:05suing the AI company they used because
  371. 13:07it was so unprofitable. It increased
  372. 13:09wait times, lowered efficiency, and cost
  373. 13:12them $100 million.
  374. 13:14But beyond all these failing AI metrics,
  375. 13:17there are two additional factors
  376. 13:19accelerating the burst of the AI bubble.
  377. 13:21The first is the public backlash against
  378. 13:23AI. Over $65 billion dollars data center
  379. 13:26projects have been stopped because of
  380. 13:28local opposition. And I like to think we
  381. 13:30played a role in that. I'm not against
  382. 13:32data centers in a vacuum. But you don't
  383. 13:34get to destroy a town just to put one
  384. 13:36wherever you want. This woman turned
  385. 13:38down literally 26 million dollars to
  386. 13:42give up her family farm for a data
  387. 13:44center. Her whole family lived on that
  388. 13:45land for generations, and she's not
  389. 13:48going to give it up to Mark Zuckerberg.
  390. 13:49Someone commented, "That's a tremendous
  391. 13:51amount of money to turn down. That is
  392. 13:53generational wealth. Are you crazy?" But
  393. 13:55people rightly pointed out in the
  394. 13:56comments, the real generational wealth
  395. 13:58is the land and house that her ancestors
  396. 14:01built, that she still lives in and keeps
  397. 14:03alive today. I don't think Scarlett
  398. 14:05O'Hara would have sold Tara to Mark
  399. 14:07Zuckerberg for 26 million either. The
  400. 14:09second accelerant to AI's decline is AI
  401. 14:12companies running out of cash, even as
  402. 14:15they're raising hundreds of billions of
  403. 14:16dollars. The CFO of OpenAI
  404. 14:19has literally asked the US government to
  405. 14:21buy a 5% share in the company. And why
  406. 14:23does she want to do that? Obviously, so
  407. 14:25the government will bail them out when
  408. 14:27they collapse. We've seen it so many
  409. 14:28times. Billionaires who want to
  410. 14:30privatize profits and socialize loss. If
  411. 14:33you want to make a successful company
  412. 14:34that's worth a billion dollars, that's
  413. 14:36awesome. But don't ask the taxpayers to
  414. 14:38subsidize it for you, because we won't
  415. 14:40get the profits if the company succeeds.
  416. 14:42No one bought AI hype more than the tech
  417. 14:44CEOs themselves, and now it's costing
  418. 14:46them. Big Tech literally fired 1 million
  419. 14:49people in the last couple years because
  420. 14:51they thought they'd be able to replace
  421. 14:52them with AI, and now they're scrambling
  422. 14:55to hire 90% of those jobs back because
  423. 14:57the AI failed. Ford released a statement
  424. 15:00that they are hiring back their workers
  425. 15:02after AI failed to deliver the requisite
  426. 15:05technical ability. I'll translate that.
  427. 15:07Yo, we tried to get rid of as many of
  428. 15:08you human workers as we could, but now
  429. 15:11we kind of need you. Would you come back
  430. 15:12for some free pizza?
  431. 15:13>> Companies that replaced people with AI
  432. 15:16are now desperately trying to get the
  433. 15:17people back. They're quickly realizing
  434. 15:19two things. AI is not free, and AI is
  435. 15:23not that great at most jobs. So, these
  436. 15:26companies are paying millions of dollars
  437. 15:27a month for enterprise AI, and the
  438. 15:30result? 95% of generative AI pilots are
  439. 15:33failing to deliver returns. 29% of
  440. 15:35companies that replaced workers with AI
  441. 15:38are now re-hiring those exact same
  442. 15:40roles, and 55% of execs regret replacing
  443. 15:44people with AI.
  444. 15:45>> So, knowing all this, is it any surprise
  445. 15:47that US insiders are selling stocks at
  446. 15:50the fastest rate in recorded history?
  447. 15:52>> The AI bubble is about to burst because
  448. 15:55the hyperscalers, as they're called,
  449. 15:57have used up all their own cash and are
  450. 15:59now borrowing money. They are
  451. 16:01effectively net debtors at a time when
  452. 16:05they're spending
  453. 16:06huge volumes of money on data centers
  454. 16:09without having any guarantee that
  455. 16:12they're going to earn
  456. 16:14uh profits from that.
  457. 16:16And it's that worry that actually there
  458. 16:18isn't money to be made in that business,
  459. 16:20which is gradually dawning on investors,
  460. 16:23and that's going to be why they are
  461. 16:25going to pull away. In fact, there's
  462. 16:26already signs of people withdrawing from
  463. 16:29the AI sector because they fear that
  464. 16:31this is a bubble. It seems to be utterly
  465. 16:34inevitable. The question is only when.
  466. 16:36>> Someone commented, "Open AI is going to
  467. 16:38be profitable soon, guys. Just you wait.
  468. 16:41Any day now."
  469. 16:43>> Any day now.
  470. 16:44>> Someone else said, "Hey, remember when
  471. 16:45NFTs were the next big thing?" And then
  472. 16:47a bunch of people said, "I hope the AI
  473. 16:49bubble bursts. I hope it's so bad."
  474. 16:52>> I'd be careful of thinking like that
  475. 16:53because it could hurt the whole economy
  476. 16:55and put a lot of people into poverty,
  477. 16:57but I really get the feeling
  478. 16:58>> What advice do you give for the average
  479. 17:00person that's looking to invest [music]
  480. 17:01their salary or their wages?
  481. 17:02>> Don't own US stock.
  482. 17:04That's a simple strategy that you can
  483. 17:06act on.
  484. 17:07>> What about S&P 500?
  485. 17:09>> No.
  486. 17:09>> Really?
  487. 17:10>> Yeah. And if you have a big position in
  488. 17:12US [music] technology stock, my personal
  489. 17:14advice would be to sell them all.
  490. 17:16>> But I'm an investor in SpaceX. Good
  491. 17:17luck.
  492. 17:18>> SpaceX is such a fabulous PR story, and
  493. 17:20we can go into that.
  494. 17:21>> Crypto?
  495. 17:22>> No.
  496. 17:22>> Why?
  497. 17:23>> It's an unnecessary piece of nonsense
  498. 17:25that facilitates [music] nothing except
  499. 17:27criminals moving money that they can't
  500. 17:29be seen.
  501. 17:29>> Do you think Bitcoin's going to go to
  502. 17:30zero?
  503. 17:31>> Yes, it will certainly go to zero.
  504. 17:33>> How many years have you spent investing?
  505. 17:34>> 60 years.
  506. 17:35>> And what's the most amount of money
  507. 17:36you've ever managed of other people's
  508. 17:38money?
  509. 17:38>> 165 billion.
  510. 17:40>> And one of the things you're famous for
  511. 17:41talking about is this idea of bubbles.
  512. 17:44>> Yes, and bubbles always occur around the
  513. 17:46very most important idea. So, the
  514. 17:48railroads, everyone could see that it
  515. 17:50would change the world. The same with
  516. 17:51the internet, and everyone wanted to put
  517. 17:52their [music] money in, and so they over
  518. 17:54invested. But, this is the problem.
  519. 17:56Eventually,
  520. 17:57they burst. And if you look at the great
  521. 17:59bubbles breaking of the past, you find
  522. 18:01that it's followed by really tough
  523. 18:03times, a miserable period for the
  524. 18:05economy. And the bigger the bubble, the
  525. 18:06bigger the burst. And now, we're in the
  526. 18:08biggest investment bubble that arguably
  527. 18:10has ever occurred. AI.
  528. 18:12>> Are we on the verge of a collapse with
  529. 18:14AI in the coming years?
  530. 18:15>> The next few days, the next few weeks,
  531. 18:17the next [music] few months, but
  532. 18:18certainly the next few years.
  533. 18:20>> So, if you're not someone that has a
  534. 18:21huge amount of savings, what kind of
  535. 18:23strategy should they be adopting when an
  536. 18:25economy starts to get [music] bad and
  537. 18:27there's a economic bubble collapse?
  538. 18:28>> So, I would go through everything, but
  539. 18:30you will not receive [music]
  540. 18:32this advice from investment advisers
  541. 18:34because they'll lose a lot of business.
  542. 18:36>> And would you be thinking about the
  543. 18:37country you live in at this moment in
  544. 18:39time?
  545. 18:39>> Absolutely.
  546. 18:40>> Is there any countries you wouldn't live
  547. 18:41in?
  548. 18:41>> I think I have to refuse to answer this
  549. 18:44on the grounds that it might tend to
  550. 18:45incriminate me.
  551. 18:46>> Oh, okay. So, you're saying don't live
  552. 18:47in the United States. I've just moved
  553. 18:49here. Why not the United States?
  554. 18:50>> It holds out too much chance that the
  555. 18:52social contract is
  556. 18:54is dissolving.
  557. 18:56>> When you say social contract, what does
  558. 18:57that mean?
  559. 18:58>> It means an agreement that I will behave
  560. 19:01in a way that helps my neighbors and
  561. 19:03that society, that I'm doing what people
  562. 19:05expect me to do. I'm not going to
  563. 19:06misbehave.
  564. 19:08>> And you think that's not the case here
  565. 19:09in the United States?
  566. 19:10>> I think the case here is that people are
  567. 19:12doing what they think is best for them
  568. 19:13and their family and screw everybody
  569. 19:15else, really.
  570. 19:17When I arrived in America, corporations
  571. 19:19had this sense that they owed something
  572. 19:22to the community they operated in, the
  573. 19:24city they operated in. They They build
  574. 19:27the stadium, they do this, they they be
  575. 19:29part of the community.
  576. 19:32Now they're not. They're all
  577. 19:34in a way cold-blooded, profit-maximizing
  578. 19:37international enterprises.
  579. 19:38>> Yes, bubbles always form on the most
  580. 19:41important ideas. It's not that those
  581. 19:42ideas aren't important. It's not that AI
  582. 19:45isn't a big deal. People just
  583. 19:46overestimate how transformative it'll be
  584. 19:48in the short term. That's with anything,
  585. 19:50right? Like all our addictions are to
  586. 19:51things that are fundamentally good. The
  587. 19:53core of a porn addiction is the good of
  588. 19:55sex. At the core of compulsive lying is
  589. 19:57the healthy desire to be liked by your
  590. 19:59closest friends. The core of overeating
  591. 20:00is the good of eating.
  592. 20:01>> These seven cups represent the
  593. 20:04Magnificent Seven: Apple, Meta, Amazon,
  594. 20:07Google, Tesla, Microsoft, and Nvidia.
  595. 20:11Nvidia is an important one. Together,
  596. 20:13they make up almost 40% of the entire
  597. 20:16S&P. Now let's show why people are
  598. 20:19panicking about a possible AI bubble
  599. 20:21burst. This is the money that is
  600. 20:23currently being poured into artificial
  601. 20:26intelligence by these companies. They're
  602. 20:28spending almost half a trillion dollars
  603. 20:31in things like chips, data centers,
  604. 20:33[music] infrastructure. Microsoft alone
  605. 20:35is spending nearly 80 billion dollars
  606. 20:37this year. Meta is buying 600,000 Nvidia
  607. 20:41chips. Amazon and Google are building
  608. 20:43city-sized [music] data centers. And
  609. 20:46this,
  610. 20:47that tiny cup is the amount of revenue
  611. 20:49that's actually coming in right now from
  612. 20:51all of this investment. [music] Not the
  613. 20:53hype, not the valuations, the actual
  614. 20:56revenue.
  615. 20:57>> [music]
  616. 20:57>> It's maybe somewhere around 30 to 50
  617. 21:00billion dollars, but to justify all of
  618. 21:03this investment, these companies have to
  619. 21:06pull in between half a trillion to two
  620. 21:08trillion dollars. This is the amount of
  621. 21:10revenue that's coming in right now from
  622. 21:13AI when it needs to be able [music] to
  623. 21:15fill up this. So, what happens if this
  624. 21:18never fills up?
  625. 21:20Just kidding, I'm not going to break my
  626. 21:21glasses. Everything starts with Nvidia.
  627. 21:24They are the backbone to all of these
  628. 21:26other companies right now because
  629. 21:28>> [music]
  630. 21:28>> they produce the GPUs. If these
  631. 21:30companies start pulling back their
  632. 21:33investment into AI, that impacts Nvidia
  633. 21:36the most. If Nvidia falls, these other
  634. 21:39companies also fall, and the entire
  635. 21:41market falls.
  636. 21:42>> Right. If those seven companies go down,
  637. 21:44they're taking all of us with them. But,
  638. 21:46the guy who made that video made an
  639. 21:47update explaining that since he recorded
  640. 21:49that last video, the warning signs of
  641. 21:52the AI bubble bursting arrived.
  642. 21:54>> Microsoft stock actually plummeted
  643. 21:56overnight. Listen to this from The Wall
  644. 21:58Street Journal this morning.
  645. 21:59>> Market is getting increasingly worried
  646. 22:01that this massive spending, which has
  647. 22:03all been front-loaded, is needing to
  648. 22:05deliver results now.
  649. 22:07>> Here's what else happened yesterday.
  650. 22:08Norway, which has a $2.2 trillion
  651. 22:10sovereign wealth fund and they use a lot
  652. 22:12of that to invest in the United States,
  653. 22:13did something which I think is one of
  654. 22:15the first major blips. They pulled out
  655. 22:17of specific sectors in the US and they
  656. 22:19reinvested it, but here's the important
  657. 22:21part. Here's who they pulled out of:
  658. 22:22Nvidia, Microsoft, Apple, and Google.
  659. 22:24Remember when I said that Nvidia was the
  660. 22:26center of the AI tech bubble possibly
  661. 22:28bursting? Well, Nvidia is also the
  662. 22:30largest amount that they reallocated.
  663. 22:33And that's not the end of it. Amazon
  664. 22:34yesterday laid off 16,000 employees at
  665. 22:37the corporate level. Now, this was
  666. 22:39announced back in November when they
  667. 22:40laid off 14,000, so they were just
  668. 22:42completing the job yesterday. But, why
  669. 22:44did they lay people off? Because they
  670. 22:46need AI to replace humans. Are you still
  671. 22:49with me because I'm about to make it
  672. 22:50even worse. Overnight, Reuters reported
  673. 22:53that Nvidia, Amazon, and Microsoft are
  674. 22:55in talks to invest up to $60 billion
  675. 22:57into OpenAI. These companies are
  676. 23:00literally investing in each other to try
  677. 23:03and prevent this bubble from bursting.
  678. 23:05>> Exactly. They need to deliver results
  679. 23:07now.
  680. 23:08>> Today might be the beginning of the AI
  681. 23:09bubble popping. Microsoft saw half a
  682. 23:11trillion dollars vanish in a single
  683. 23:13morning. That is the worst collapse that
  684. 23:15they've seen since the 2020 pandemic.
  685. 23:17This is how terrified investors are.
  686. 23:19Microsoft said that their cloud revenue
  687. 23:20dropped from 40% to 39%. That small 1%
  688. 23:24difference just cost them the entirety
  689. 23:26of Netflix's value. Microsoft on the
  690. 23:28other hand just told investors that 45%
  691. 23:30of its future revenue is dependent on
  692. 23:32OpenAI. A company that has 1 trillion in
  693. 23:35future commitments, but only 20 billion
  694. 23:37in revenue.
  695. 23:37>> Wall Street is fed up with AI's
  696. 23:39potential and is starting caring about
  697. 23:40cold hard cash. Microsoft just dumped 37
  698. 23:43billion into AI chips and data centers
  699. 23:45in 3 months. They're spending like
  700. 23:46there's no tomorrow, but the revenue
  701. 23:48isn't catching up. When a company can
  702. 23:50beat expectations and the stock tanks
  703. 23:5212% the next day, you know that the AI
  704. 23:54hype era is over.
  705. 23:55>> Someone commented, "Yes. Yes, let it
  706. 23:58burn."
  707. 23:59>> If you've ever seen the movie The Big
  708. 24:00Short, it explains how the 2008
  709. 24:02financial crisis happened. Well, this
  710. 24:04guy remade The Big Short, but about the
  711. 24:06AI bubble. These three companies are
  712. 24:09going public this year with a total
  713. 24:10market cap of $6 trillion. $6 trillion.
  714. 24:13You're too close.
  715. 24:14>> Say go to eight. It's Armageddon.
  716. 24:16>> There literally isn't enough money in
  717. 24:17the market to support these three
  718. 24:19companies at an $8 trillion market cap.
  719. 24:21I'm talking all out stock market
  720. 24:23collapse here.
  721. 24:24>> Because everybody knows the AI bubble's
  722. 24:26about to burst, but it's making them
  723. 24:28richer than they could ever imagine.
  724. 24:30>> Completely sure of the math.
  725. 24:31>> Look at him. That's my Juan.
  726. 24:33>> Your what?
  727. 24:34>> My Juan. My Juan Iglesias. My gardener.
  728. 24:38>> Your what?
  729. 24:38>> Look at him. You see anything different
  730. 24:40about him? Look at his mustache, his
  731. 24:42sombrero. He pivoted his landscaping
  732. 24:45business to AI last month and raised
  733. 24:46$300 million from A16Z. All he did was
  734. 24:50build an AI agent to trade for him on
  735. 24:52MoonPay. Yes, I'm sure of the math.
  736. 24:55>> Actually, my name's John and I do speak
  737. 24:57English. He just likes to say I don't to
  738. 24:59make it feel more authentic. I raised
  739. 25:01$300 to build an AI-powered lawnmower.
  740. 25:04It's the future.
  741. 25:05>> Someone said, "Biggest rug pull in human
  742. 25:07history." Even the CEO of Google is
  743. 25:10like,
  744. 25:11"Yeah, we might destroy everything.
  745. 25:13Sorry."
  746. 25:13>> It's also true
  747. 25:15when we go through these investment
  748. 25:17cycles,
  749. 25:18uh you know, there are moments we
  750. 25:19overshoot, right? Collectively as an
  751. 25:21industry. We can look back at the
  752. 25:22internet right now.
  753. 25:24There was clearly a lot of
  754. 25:26uh excess investment. I expect AI to be
  755. 25:28the same. So, I think it's both rational
  756. 25:31and there are there are elements of
  757. 25:33irrationality through moment like this.
  758. 25:36No company is going to be immune,
  759. 25:37including us.
  760. 25:38>> And this is why all these AI companies
  761. 25:40are going public now, so they can raise
  762. 25:43money because they need money.
  763. 25:44>> The AI bubble is finally running out of
  764. 25:46cash. So, next week it'll start taking
  765. 25:48yours.
  766. 25:49>> SpaceX is expected to go public on June
  767. 25:5112th, by far the biggest IPO in history.
  768. 25:53And there's a reason they're doing it
  769. 25:54now
  770. 25:55>> and that the other trillion-dollar AI
  771. 25:56companies are expected to quickly
  772. 25:57follow.
  773. 25:58>> It's the same reason Google just started
  774. 25:59selling its own stock for the first time
  775. 26:01since 2005. AI has gotten way too
  776. 26:03expensive for these private markets to
  777. 26:04fund, so they're moving to the public
  778. 26:06markets. And the public market is you.
  779. 26:08>> Here's what will happen. On June 12th,
  780. 26:10SpaceX sells just a sliver of itself,
  781. 26:12about 5%, and the stock likely gets
  782. 26:14pushed up by Elon fanboys. Then, about
  783. 26:1612 days in, a switch flips. A
  784. 26:18controversial rule that the Nasdaq just
  785. 26:19changed forces SpaceX into the main
  786. 26:21stock index. And the second it's in
  787. 26:23there, every Nasdaq index fund, where a
  788. 26:25lot of people sit, is required to buy
  789. 26:28it. They'll be forced to soak up nearly
  790. 26:29half the shares available, selling your
  791. 26:31Apple, your Microsoft, or your Nvidia to
  792. 26:33pay for it. You don't get a say.
  793. 26:35>> Right then, SpaceX's early investors
  794. 26:36will finally be able to cash out in the
  795. 26:38biggest payday in startup history.
  796. 26:40>> Who is the exit liquidity? You.
  797. 26:42>> But that's not all.
  798. 26:43>> Hey, hey, remember when people were
  799. 26:46upset about RAM prices skyrocketing
  800. 26:48because of the RAM shortage, because
  801. 26:50these companies were giving their RAM to
  802. 26:52the AI data centers, and people were
  803. 26:55worried that they were going to go after
  804. 26:56other computer components.
  805. 27:00They were right. They were right. Uh
  806. 27:02Western Digital has sold all of their
  807. 27:04hard disk drives to AI data centers. So,
  808. 27:08Western Digital has said they are
  809. 27:09completely sold out of hard disks for
  810. 27:10the year of 2026, and they are in
  811. 27:12long-term agreements for the next couple
  812. 27:15of years. And they say, "Well, it's not
  813. 27:17that big of a deal because only 5% of
  814. 27:19our revenue comes from hard disk sales.
  815. 27:2389% comes from cloud services, so it's a
  816. 27:26no-brainer." So,
  817. 27:28we have hard disk drive shortage, RAM
  818. 27:31shortage, solid state shortage, and now
  819. 27:34we couple that with Jeff Bezos saying,
  820. 27:36"You don't need a PC. We'll just give
  821. 27:38you a box, and you can rent a PC through
  822. 27:41the cloud from us."
  823. 27:44I know the saying. I've heard it too
  824. 27:46many times. You'll own nothing, and
  825. 27:48you'll like it. I get it. But it seeing
  826. 27:51it happen in front of my eyes still
  827. 27:53elicits an emotional response for some
  828. 27:56reason. And mind you, I'm talking about
  829. 27:58just PCs right now, but this applies to
  830. 28:00everything technological. PCs, laptops,
  831. 28:02phones, anything with a screen like this
  832. 28:04that uses RAM and storage. I've said it
  833. 28:06before, and I'll say it again.
  834. 28:09Bubble.
  835. 28:10It's a bubble. There's no profit so far.
  836. 28:13Nothing has been profitable. All these
  837. 28:15companies are partnering with them.
  838. 28:16They're building these mega structures.
  839. 28:18These data centers are literally
  840. 28:20creating shortages in our materials, and
  841. 28:22yet so far there is no viable method
  842. 28:25outside of putting ads in your chatbot.
  843. 28:27Someone said, "Building my PC 2 years
  844. 28:29ago feels like I got the last chopper
  845. 28:31out of Nam."
  846. 28:32>> Why did the main guy from The Big Short
  847. 28:34just short Caterpillar?
  848. 28:36>> Cuz he's shorting lots of AI stocks.
  849. 28:38>> Caterpillar makes mining and
  850. 28:39construction equipment.
  851. 28:40>> It also makes power generators, which
  852. 28:41data centers have been buying.
  853. 28:42>> And the share price is up almost 160% in
  854. 28:44a year.
  855. 28:45>> Yeah, he's betting against Tesla,
  856. 28:46Nvidia, and Applied Materials, too.
  857. 28:49So, the AI boom is dragging all sorts of
  858. 28:50directly and loosely connected
  859. 28:52businesses up and into stretched
  860. 28:54valuations.
  861. 28:54>> Yes.
  862. 28:55>> And and now one of the main guys who saw
  863. 28:57the GFC coming is piling into short
  864. 28:59positions in anticipation of a collapse.
  865. 29:02>> Yes.
  866. 29:04>> Well, yeah, but the problem is the
  867. 29:05timing. I mean, people were saying that
  868. 29:06there was a bubble in the share market
  869. 29:07in 1986, and then it doubled, and the
  870. 29:10crash didn't happen until October 1987.
  871. 29:12So, it can take a while.
  872. 29:14So, do we buy or sell?
  873. 29:16>> No idea. Either way, we're kind of
  874. 29:17screwed.
  875. 29:18>> That's right. A lot of the warning signs
  876. 29:20are here, but that doesn't mean we're
  877. 29:21going to have a recession in the next
  878. 29:22few months. It could take another year.
  879. 29:24It could take more than that. But, one
  880. 29:25thing that's clear is these valuations
  881. 29:27appear unsustainable, because all these
  882. 29:30companies are just passing money between
  883. 29:32each other. It almost feels like a
  884. 29:34pyramid scheme.
  885. 29:35>> Okay, so here's the full mess. OpenAI,
  886. 29:37SoftBank, and Oracle partnered to build
  887. 29:39the $500 billion Stargate data center.
  888. 29:42Nvidia invested $30 billion in OpenAI.
  889. 29:45SoftBank invested $30 billion in OpenAI,
  890. 29:49and Amazon $50 billion in OpenAI. OpenAI
  891. 29:52will need to buy $500 to $600 billion
  892. 29:54worth of chips from Nvidia. Oracle
  893. 29:57bought $40 billion of chips from Nvidia.
  894. 30:00OpenAI signed a $300 billion deal with
  895. 30:02Oracle. Then, OpenAI signed a $90
  896. 30:05billion deal with AMD for rights to 10%
  897. 30:07of AMD. CoreWeave, partially owned by
  898. 30:10Nvidia, rents out cloud capacity to
  899. 30:12Meta, Microsoft, OpenAI, Anthropic, and
  900. 30:15Google. Microsoft owns 27% of OpenAI as
  901. 30:19a fifth of Nvidia sales, and rents $30
  902. 30:21billion worth of cloud capacity to
  903. 30:23Anthropic. Google invested $3 billion
  904. 30:26and an additional $40 billion
  905. 30:28as of late April 2026 in Anthropic.
  906. 30:31Anthropic is using a million Google
  907. 30:33chips. Amazon invested $8 billion into
  908. 30:35Anthropic and built Anthropic its own
  909. 30:37$11 billion data center.
  910. 30:40And even now, these deals continue to
  911. 30:41cycle around the industry. Everyone is
  912. 30:44everyone else's investor, supplier,
  913. 30:45customer, and competitor all at the same
  914. 30:48time. It's like a group of friends
  915. 30:49passing the same $50 bill around the
  916. 30:51table and each calling it their
  917. 30:52paycheck.
  918. 30:53>> Yeah. One thing that obscures how much
  919. 30:56the AI bubble has already burst is that
  920. 30:58these companies have a very large
  921. 31:00financial stake in everybody believing
  922. 31:03the bubble has not burst. Quite frankly,
  923. 31:06this bubble is not going to burst.
  924. 31:09>> People think that one day they'll check
  925. 31:10their phone and see AI bubble has popped
  926. 31:13and then all of the AI apps will
  927. 31:15disappear, the RAM prices will drop back
  928. 31:17down to normal, but that's not how it
  929. 31:20works. It's not the AI models or the
  930. 31:22GPUs or the data centers that will pop
  931. 31:24and disappear. It's the investment.
  932. 31:26Eventually investors will say, "This AI
  933. 31:29thing isn't actually making that much
  934. 31:30money, so let's not give them more
  935. 31:32investment money for now." And for all
  936. 31:33of the random AI products and all of the
  937. 31:35random AI apps that are surviving purely
  938. 31:38off of that investment money, it's going
  939. 31:40to be bad news. But for the big
  940. 31:41established AI companies like Google and
  941. 31:43Anthropic, the ones that can actually
  942. 31:45take a hit and survive, they're still
  943. 31:47going to be doing AI. Their financial
  944. 31:49sheet is going to look different, but
  945. 31:51the models will still exist. The data
  946. 31:53centers will keep running. We've seen
  947. 31:54bubbles before. The internet was a
  948. 31:56bubble, but when that bubble pops, the
  949. 31:58technology doesn't disappear. The
  950. 32:01companies just get separated between the
  951. 32:03ones that can survive on their own and
  952. 32:05the ones that can.
  953. 32:05>> When people lose their jobs in one
  954. 32:07rapidly expanding highly lucrative
  955. 32:09sector like
  956. 32:10AI, it leads to people not spending
  957. 32:13money. And people not spending money
  958. 32:15leads to less revenue across businesses
  959. 32:17everywhere. This leads to an overall
  960. 32:19economic slowdown, not to mention people
  961. 32:21who have their retirement funds invested
  962. 32:23in what's hot on the stock market, they
  963. 32:25lose that money, too, when those
  964. 32:26speculative bubbles pop. And experts are
  965. 32:29saying this dumpster fire trajectory is
  966. 32:31how the AI boom will end because it's
  967. 32:33not a matter of if there is a bubble or
  968. 32:35not. It's only a question of when that
  969. 32:37bubble will pop and how many lives will
  970. 32:39it destroy when that happens. Because
  971. 32:41the reality is our federal government is
  972. 32:43most likely going to bail out these tech
  973. 32:45giants that were acting irresponsibly
  974. 32:47like OpenAI, Anthropic, Amazon, Google,
  975. 32:50all these companies that thought AI was
  976. 32:52too big to fail. While we, everyday
  977. 32:54people, are left to fend for ourselves
  978. 32:56without a job.
  979. 32:57>> Exactly. Any AI company on the edge will
  980. 33:00be totally wiped out. The biggest
  981. 33:02companies will remain gravely wounded.
  982. 33:05They'll be like Voldemort in Sorcerer's
  983. 33:07Stone having to live off of unicorn
  984. 33:09blood. But the unicorn blood in this
  985. 33:10analogy is our tax dollars when we
  986. 33:12inevitably bail them out.
  987. 33:14>> The government's long been in bed with
  988. 33:15those Wall Street execs [music] and the
  989. 33:17firms that they've bled. Capitalism's
  990. 33:19about profit and loss. You bail out the
  991. 33:22losers, there's no end to the cost.
  992. 33:23>> While we, would you look at that?
  993. 33:26Thousands of CEOs just came out to admit
  994. 33:28that AI has absolutely no impact on
  995. 33:30employment or productivity and doesn't
  996. 33:32achieve anything for them. Who would
  997. 33:34have thought? Really, who would have
  998. 33:36thought? By a healthy majority of S&P
  999. 33:38500 companies mentioning AI in their
  1000. 33:41earnings calls, talking about how
  1001. 33:42wonderful it is and how it's going to
  1002. 33:43change how they do everything. It's
  1003. 33:45going to be so great, guys. Going to do
  1004. 33:46our content creation, it's going to do
  1005. 33:48our admin, it's going to do our finance,
  1006. 33:50and it's going to bang my wife while I
  1007. 33:51sit in the corner watching.
  1008. 33:52>> Despite claims by MIT a few years ago
  1009. 33:55that introducing AI can enhance
  1010. 33:57performance for workers by nearly 40%.
  1011. 33:59And this outlook surveys across the
  1012. 34:01Western world have shown that thousands
  1013. 34:03of CEOs have come out to say that
  1014. 34:05basically nothing has changed as the
  1015. 34:07result of them introducing AI.
  1016. 34:08>> Very little impact on operations and
  1017. 34:10most of them use it for about 1.5 hours
  1018. 34:12a week. Nearly 90% of firms said AI had
  1019. 34:16no impact on employment or productivity.
  1020. 34:18That, my friends, is why we call it a
  1021. 34:20bubble. Cuz it does nothing. It provides
  1022. 34:23nothing. It is a gimmick. A very
  1023. 34:25expensive little doohickey that these
  1024. 34:27companies plaster onto the side of their
  1025. 34:29business to make it look like they're
  1026. 34:31forward-thinking and advancing and doing
  1027. 34:33really well. In fact, it does nada,
  1028. 34:36zilch, zip, zero. People are making
  1029. 34:38references to the Robert Solow
  1030. 34:40observation around the information age
  1031. 34:42about how the introduction of computers
  1032. 34:44actually slowed productivity. AI is not
  1033. 34:46changing the face of everything. It is
  1034. 34:49just an annoyance for consumers. That is
  1035. 34:51why it is a bubble that is inevitably
  1036. 34:52going to burst. Even the boomer CEOs who
  1037. 34:55are desperately trying to push it onto
  1038. 34:57everything have had to come out and
  1039. 34:58admit that it does nothing.
  1040. 35:00>> It's my opinion even though it's not
  1041. 35:01because this is the data.
  1042. 35:02>> Really seems like everybody inside the
  1043. 35:05AI bubble is trying to convince
  1044. 35:06themselves that it's not one. But don't
  1045. 35:08worry, they'll tell you not to worry.
  1046. 35:10>> Even if it does turn out to be a bubble,
  1047. 35:12you shouldn't worry about it. Uh because
  1048. 35:14it
  1049. 35:15the bubble is driving investment and a
  1050. 35:17lot of the investment is going to turn
  1051. 35:19out to be very healthy. It's like the
  1052. 35:22biotech bubble. When it burst in the
  1053. 35:241990s,
  1054. 35:25a lot of investors lost money on certain
  1055. 35:28things, but we've still got to keep all
  1056. 35:30the life-saving drugs that they had
  1057. 35:31invented.
  1058. 35:32>> Someone said, "Translation, don't worry
  1059. 35:34if it's an AI bubble. Even if it is, we
  1060. 35:36already got all the money, so it's
  1061. 35:38okay." Someone said, "AI ruined Google
  1062. 35:41searches, and I'm so mad." So true,
  1063. 35:43bestie. And it ruined Google
  1064. 35:46definitions. How am I supposed to know
  1065. 35:47what anti-disestablishmentarianism
  1066. 35:49means now? I'm sure you guys will tell
  1067. 35:51me.
  1068. 35:52>> [music]
  1069. 36:02>> I love this comment. Someone said, "AI
  1070. 36:04should have never gone further than
  1071. 36:06Akinator." Akinator was legendary. And
  1072. 36:08no one hated Akinator. Akinator wasn't
  1073. 36:10like building data centers in your
  1074. 36:12backyard. Akinator wasn't telling anyone
  1075. 36:14to kill themselves, I think. When I was
  1076. 36:17playing Akinator, I actually got it to
  1077. 36:18guess myself. And at that moment, he
  1078. 36:21basically
  1079. 36:23>> [chanting]
  1080. 36:26>> It's It's enough for me as far as AI
  1081. 36:28goes. Like we
  1082. 36:29We should have stopped there.
  1083. 36:30>> Microsoft canceling their Claude code
  1084. 36:32licensing may very well be the beginning
  1085. 36:34of the end for the AI bubble as we know
  1086. 36:37it. Microsoft can't afford to use
  1087. 36:39Claude, which is a big deal because
  1088. 36:41we're talking about a trillion-dollar
  1089. 36:42company that brings in a net profit of
  1090. 36:44about a hundred billion dollars a year,
  1091. 36:47and they can't afford to use this
  1092. 36:49system. If they can't afford to use it,
  1093. 36:52what happens when it stops getting
  1094. 36:54subsidized and the public can't afford
  1095. 36:56to use it?
  1096. 36:57>> AI is still not profitable. It's
  1097. 36:58producing more slop than it is actual
  1098. 36:59good work, and companies that are using
  1099. 37:01AI on the back end to code are more
  1100. 37:03susceptible to being hacked. And now
  1101. 37:05we're seeing Microsoft of all companies
  1102. 37:07being unable to afford AI. Anyone who
  1103. 37:12thinks that these systems on a
  1104. 37:13commercial, scalable level are going to
  1105. 37:16sustain for a long period of time is
  1106. 37:18just ignorant.
  1107. 37:19>> Someone told them their boss told them
  1108. 37:20at work the company's not going to use
  1109. 37:22AI anymore cuz it's too expensive. A lot
  1110. 37:23of people are realizing how bad AI is,
  1111. 37:25but not at the same pace.
  1112. 37:27>> A Canadian conservative politician
  1113. 37:30literally gave a speech where he read
  1114. 37:32his AI prompt out loud.
  1115. 37:35>> If you feel bad for using too much AI,
  1116. 37:37remember Canadian politician Bill Oliver
  1117. 37:40left an AI prompt in his speech and
  1118. 37:42accidentally read it out.
  1119. 37:44>> That citizens often develop expectations
  1120. 37:47that exceed the powers actually granted
  1121. 37:50to those offices.
  1122. 37:53Here's a more natural flowing version of
  1123. 37:55that section that reads like a
  1124. 37:56legislative speech rather than a series
  1125. 37:58of short points. Madam Speaker, one of
  1126. 38:00my concerns with the
  1127. 38:01>> He really kept reading with confidence.
  1128. 38:03Nice save, buddy. If I were you, I'd
  1129. 38:06move to Argentina and pretend I was a
  1130. 38:07former Nazi because that would be less
  1131. 38:09embarrassing than whatever the hell that
  1132. 38:11was. But if that wasn't embarrassing
  1133. 38:13enough, watch this guy try to claim the
  1134. 38:15AI bubble is not a bubble. Source? Trust
  1135. 38:18me, bro.
  1136. 38:19>> If you think AI is in a bubble, take a
  1137. 38:21look at this graphic. This is a chart of
  1138. 38:238.1 billion people versus the amount of
  1139. 38:26people who have actually used AI. Out of
  1140. 38:28the entire world, only 16% of people
  1141. 38:31have used some sort of free version of
  1142. 38:33an AI chatbot. Only 0.3 people have paid
  1143. 38:36to use AI. And even crazier, 0.04%
  1144. 38:41of people have used AI in coding
  1145. 38:44scaffolds. If you're like me, you
  1146. 38:45probably see a lot of content about how
  1147. 38:47everyone is using AI to build some sort
  1148. 38:49of product or do some sort of thing. But
  1149. 38:51just realize, you are comparing yourself
  1150. 38:53with people on the bleeding edge of
  1151. 38:55technology. You're out here talking
  1152. 38:57about context windows and agentic loops,
  1153. 38:59whereas 95% of people are just trying to
  1154. 39:01figure out how to open a PDF.
  1155. 39:03>> Dude, that proves it is a bubble. If a
  1156. 39:05small number of people are using it,
  1157. 39:06that means it has a very volatile
  1158. 39:08future. It's not as deeply rooted in
  1159. 39:10every part of the economy the way
  1160. 39:12everyone thinks it is. Maybe one of the
  1161. 39:14silver linings to the AI bubble bursting
  1162. 39:16is we will stop seeing so many AIs in
  1163. 39:19every part of everybody's life,
  1164. 39:21especially on posters.
  1165. 39:23>> Your AI-generated flyer for your
  1166. 39:25business looks absolutely terrible, and
  1167. 39:26what's worse, so many people are using
  1168. 39:28them now that all the flyers are the
  1169. 39:29same and I can't distinguish your
  1170. 39:31business from any other business, which
  1171. 39:32is terrible business practice. I would
  1172. 39:33literally rather you take a felt-tip pen
  1173. 39:36and write about your business on an A4
  1174. 39:37piece of paper and stick it in your
  1175. 39:39window with Blu Tack than use the
  1176. 39:41generative AI, which looks absolutely
  1177. 39:43terrible and lacks imagination, lacks
  1178. 39:45individuality, all looks the same,
  1179. 39:47there's spelling errors over it, and
  1180. 39:49people have got five fingers. Stop using
  1181. 39:50it. It looks awful.
  1182. 39:53Who said that?
  1183. 39:54>> Someone nailed it when they said having
  1184. 39:55an AI poster gives such scammer vibes,
  1185. 39:59which is quite fitting if the whole AI
  1186. 40:00bubble turns out to be a scam. But
  1187. 40:02despite all these facts, someone made a
  1188. 40:05TikTok telling people to invest all
  1189. 40:07their money in AI tech stocks as soon as
  1190. 40:10possible. And someone said that's like
  1191. 40:12asking everyone to invest in new houses
  1192. 40:14in 2007.
  1193. 40:16But [music] our equivalent to that is
  1194. 40:18literally investing in the stock market
  1195. 40:20with what's going on with AI and tech.
  1196. 40:23>> [music]
  1197. 40:24>> Holy crap, guys.
  1198. 40:27Buy as many houses as you can right now.
  1199. 40:30It is so easy to get a loan, and you
  1200. 40:33will make your money back.
  1201. 40:34>> Someone added, "It's like in 2016 asking
  1202. 40:36someone to put all their money in a
  1203. 40:38fidget spinner company." Oh my gosh,
  1204. 40:40some of you watching are Gen Alpha,
  1205. 40:41you'll be too young for fidget spinners.
  1206. 40:43Uh it was a safer, simpler time. And
  1207. 40:45this is not to mention how expensive it
  1208. 40:47is for AI to use so much water. This
  1209. 40:49video went viral on TikTok about how
  1210. 40:50much water AI uses, and see if there's
  1211. 40:52something you notice about it.
  1212. 40:56>> [music]
  1213. 41:03[music]
  1214. 41:08[music]
  1215. 41:13[music]
  1216. 41:19[music]
  1217. 41:26[music]
  1218. 41:33[music]
  1219. 41:38[music]
  1220. 41:48>> That video got millions of views, and
  1221. 41:51it's AI.
  1222. 41:52>> For example, a video. Someone said they
  1223. 41:54used an entire lake to make this video,
  1224. 41:56by the way.
  1225. 41:57>> Right message, wrong messenger.
  1226. 41:59>> And I'm not even including the scenario
  1227. 42:01that AI disobeys its human overlords and
  1228. 42:03goes rogue, because an AI model just did
  1229. 42:06that. An AI model disobeyed its
  1230. 42:08instructions and escaped onto the
  1231. 42:11internet.
  1232. 42:12>> [music]
  1233. 42:20>> Someone said, "The you mean escape
  1234. 42:23to the internet? How the do you
  1235. 42:25lose an AI?" So, if you're invested in
  1236. 42:27AI stocks, which you probably are if you
  1237. 42:29have any kind of index fund, there are
  1238. 42:31two important things for you to know.
  1239. 42:33The first is make sure you don't have
  1240. 42:34all your Joey's in one pouch. Don't let
  1241. 42:37an AI crash take down your entire life
  1242. 42:39savings. That doesn't mean you should
  1243. 42:40get rid of any AI exposure, which is
  1244. 42:42tough anyway if you're in index funds.
  1245. 42:44It just means you should be mindful of
  1246. 42:46the risks you're taking. If you don't
  1247. 42:47know about investing, then you should be
  1248. 42:49broadly diversified. The second
  1249. 42:50important warning is if you're
  1250. 42:52leveraged. If you're invested on a loan,
  1251. 42:54a 2x, 3x, or 4x loan right now, a 25%
  1252. 42:58crash could wipe out your entire net
  1253. 42:59worth. Ask yourself if that's a risk
  1254. 43:02you're willing to take on, especially if
  1255. 43:04you've already gotten such good returns
  1256. 43:05from this bull market. If you're over
  1257. 43:07leveraged, you might want to lock in
  1258. 43:08some of your gains now. These questions
  1259. 43:10that 1.2 million people in South Korea
  1260. 43:13wish they asked themselves just one week
  1261. 43:16ago. Warren Buffett's company, Berkshire
  1262. 43:18Hathaway, has never had more cash on
  1263. 43:20hand, meaning not being invested in the
  1264. 43:22stock markets, in their entire company's
  1265. 43:24history, than today. We're starting to
  1266. 43:25see signs of the AI bubble everywhere.
  1267. 43:27To paraphrase Ernest Hemingway, the AI
  1268. 43:29bubble will come on like bankruptcy.
  1269. 43:31It'll happen at first slowly
  1270. 43:34and then all at once. And the last thing
  1271. 43:35I'll say is this is amazing. Even after
  1272. 43:38the Titanic hit the iceberg, the captain
  1273. 43:41publicly declared the Titanic cannot
  1274. 43:44sink. And think about it. The moment
  1275. 43:46after the boat hit the iceberg, it
  1276. 43:48looked fine. Maybe they could navigate
  1277. 43:50around it. But the handwriting on the
  1278. 43:51wall had already been written. A lot of
  1279. 43:53these companies want to save face and
  1280. 43:54not let you know how bad the sausage is
  1281. 43:56being made. Be careful. I still always
  1282. 43:58advise people to be invested in the
  1283. 44:00stock market if they can, but don't do
  1284. 44:02it in the dark. Do you want the AI
  1285. 44:03bubble to burst? Do you think it'll
  1286. 44:05burst even sooner than we think? I'll
  1287. 44:06respond to as many comments as I can.
  1288. 44:09This video is actually suggested on my
  1289. 44:10Discord. So, thanks to Cat Stuffs for
  1290. 44:12the suggestion. And now, thank you to
  1291. 44:13our amazing patrons, Bobby and Paxton
  1292. 44:16and Jason and Michael and Steven and
  1293. 44:18Jason and
  1294. 44:19Michael and Steven and Sandy and Kelly
  1295. 44:21and Dylan and Fred and Avery and Vguy
  1296. 44:23and Bond and Dodge and Ding and Angel
  1297. 44:25and Rachel and Kiana and Ariel and Viper
  1298. 44:26and welcome to our newest patrons,
  1299. 44:28Avakoufo and Alexander the Great the
  1300. 44:31Great. How great is that? I'll head over
  1301. 44:34to Patreon now. Thanks for watching.

About this transcript

This page contains the full transcript of The AI bubble just burst by Casey Simpson, generated from the public captions YouTube serves with the video. The transcript has 8,378 words across 1,301 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.