YouTube2Text

Don't Buy Every AI Dip. Buy the Companies Controlling THIS — Transcript

by Milk Road Stocks · 2,061 words · 362 segments · language en · Watch on YouTube

Full transcript

  1. 0:00I think the market is like severely
  2. 0:02overreacting to this because there is a
  3. 0:05difference between
  4. 0:06saying that the new data center
  5. 0:08construction needs better oversight and
  6. 0:10saying that companies can no longer
  7. 0:12build data center
  8. 0:14data centers. And those are not the same
  9. 0:17thing because I was looking at Hinojos
  10. 0:19own website and the priorities she
  11. 0:22outlined are quite specific. She called
  12. 0:25for stronger labor protection, requiring
  13. 0:28developers to employ skilled American
  14. 0:30union workers and giving local
  15. 0:33communities a greater voice in whether
  16. 0:36these facilities should be built near
  17. 0:39them. She also proposed greater
  18. 0:41transparency including restrictions on
  19. 0:43non-disclosure agreements that can
  20. 0:45prevent residents from understanding the
  21. 0:48details of what's actually being
  22. 0:49developed in the communities. And two of
  23. 0:52her biggest priorities relate to
  24. 0:55electricity and water. She wants data
  25. 0:58center developers to demonstrate that
  26. 1:00surrounding households won't face higher
  27. 1:03electricity or water cost because of new
  28. 1:05projects, you know.
  29. 1:08And she wants developers to essentially
  30. 1:10take more responsibility to financing
  31. 1:12their own infrastructure rather than
  32. 1:15depending on tax paying dollars or
  33. 1:18subsidies.
  34. 1:19And finally she proposed stronger
  35. 1:21protections for
  36. 1:22you know waters local water supplies and
  37. 1:25particularly in situations where large
  38. 1:28industrial water contracts would
  39. 1:30threaten the resources available to
  40. 1:34residents. Is any of this crazy? Like I
  41. 1:37don't think what she's asking is crazy
  42. 1:39at all. I think these are pretty
  43. 1:42reasonable ask. Like if you want to
  44. 1:45build data centers in our community,
  45. 1:48you have to work with the community a
  46. 1:50little bit more better. You have to
  47. 1:52demonstrate that your projects won't
  48. 1:56place an unfair financial burden on the
  49. 1:58people living there. I mean, that is
  50. 2:00what everyone should realistically do,
  51. 2:04but some of these media outlets like to
  52. 2:06spin this, you know, spin these
  53. 2:09takes as, oh, they're shutting data
  54. 2:11centers down. No, I think
  55. 2:14data centers provide an enormous amount
  56. 2:16of jobs in the community.
  57. 2:18It's it's growing like it's the largest
  58. 2:22part part of our GDP like growth as of
  59. 2:25now.
  60. 2:26If
  61. 2:27the data center build out
  62. 2:29gets slower,
  63. 2:31I think we're going to severely because
  64. 2:33severely be
  65. 2:36potentially going towards a negative GDP
  66. 2:38growth and if you have two two
  67. 2:41consecutive months of negative GDP
  68. 2:42growth, congratulations, you're on
  69. 2:45you're on recession right now. You have
  70. 2:47a recession in your hands. So nobody
  71. 2:48wants that. I don't think Democrats or
  72. 2:52Republicans want a recession in their
  73. 2:53hands because
  74. 2:55now Democrats and Republicans will point
  75. 2:57fingers at each other and say, oh, you
  76. 3:00caused it, you caused it, but in reality
  77. 3:02both of them will cause
  78. 3:05cause this if if if the build out don't
  79. 3:08continue. So it is essential for this
  80. 3:11build out
  81. 3:12to take place and I think companies need
  82. 3:15to work with these communities to build
  83. 3:18these data center out. Now, here's why I
  84. 3:21don't think the AI build out is going to
  85. 3:24slow down anytime soon
  86. 3:26because AI build out is already well
  87. 3:29underway. We all know that. Amazon,
  88. 3:31Google, Microsoft, Meta
  89. 3:34collectively are spending billions,
  90. 3:36okay, billions of dollars. Look at these
  91. 3:39look at these three CAPEX estimates, you
  92. 3:42know. I think this just got revised just
  93. 3:44this week alone or last week.
  94. 3:47Meta, Google, Amazon, Oracle, CoreWeave,
  95. 3:49Nebius, SpaceX
  96. 3:51could increase from approximately two or
  97. 3:55did increase from approximately 280
  98. 3:57billion in 2024 to 951 billion in 2026.
  99. 4:02By 2027 that projected figure is going
  100. 4:06to reach 1.37
  101. 4:08trillion.
  102. 4:10Okay, and by 2028 we're talking about
  103. 4:121.5 trillion
  104. 4:15going into these data centers. Now, I
  105. 4:17think there is a reasonable possibility
  106. 4:19that spending projects could continue
  107. 4:21moving higher. Why? Because we're still
  108. 4:26relatively early in the development of
  109. 4:28AI.
  110. 4:29Think about how much AI has changed over
  111. 4:32the past few years. We have essentially
  112. 4:35gone from large language models that
  113. 4:37could answer questions
  114. 4:39and generate text to increasingly
  115. 4:41sophisticated AI agents that runs for
  116. 4:44days.
  117. 4:45And every new generations of these
  118. 4:47system requires enormous computing
  119. 4:50resources. And as more businesses
  120. 4:53integrate AI into their everyday
  121. 4:54operations,
  122. 4:56the computing power required to serve
  123. 4:58those application could increase term
  124. 5:00significantly.
  125. 5:02>> We're going to pause for a second.
  126. 5:04The AI trade is changing. The bottleneck
  127. 5:06stocks are either running out of steam
  128. 5:08or getting ready for another leg higher
  129. 5:10while the application layer is finally
  130. 5:11starting to show some real profits. Our
  131. 5:13analysts were early to the last wave and
  132. 5:15now they are buying for the next one.
  133. 5:18Catch your next call inside Milk Road
  134. 5:19Pro for just a dollar at the link below.
  135. 5:21>> And we're still so early. That's what
  136. 5:23I'm saying. We're probably in the second
  137. 5:26or the third inning of this AI
  138. 5:28infrastructure build out. And
  139. 5:30uh as investors, the profits may not
  140. 5:33come till the sixth, seventh, or the
  141. 5:34eighth or even the ninth inning of the
  142. 5:36majority of the profits. We're starting
  143. 5:38to see the profits come in, but the
  144. 5:40infrastructure has to get built out
  145. 5:42before before the profits come in. So
  146. 5:45another reason I think that we're not
  147. 5:48these this CAPEX number is going to go
  148. 5:49higher is because competition. We all
  149. 5:52know hyperscalers are competing against
  150. 5:54each other. If Microsoft decides to
  151. 5:57significantly reduce its AI
  152. 5:58infrastructure while Amazon continues to
  153. 6:00expand AWS,
  154. 6:02Microsoft essentially risk lose
  155. 6:05losing customers that need access to you
  156. 6:07know, additional computing capacity to
  157. 6:10Amazon. If Google slows down its
  158. 6:12investment in AI infrastructure while
  159. 6:14Microsoft and OpenAI keep pushing
  160. 6:16forward, Google risks losing ground in
  161. 6:19enterprise AI and cloud computing. If
  162. 6:22and if Meta decides to
  163. 6:25you know, reduce spending on AI models,
  164. 6:27somebody else is going to take take
  165. 6:29their position. These companies
  166. 6:31understand that and the because of the
  167. 6:33potential long-term market is so
  168. 6:35enormous, the TAM is like so so huge
  169. 6:39because we're talking about human
  170. 6:41workers are, you know, essentially
  171. 6:43getting replaced by AI. So the TAM is a
  172. 6:46lot here.
  173. 6:48So
  174. 6:49I don't think this CAPEX is going to
  175. 6:52slow down. Yes, it might slow down like
  176. 6:55in terms of like the dollar amount
  177. 6:56because it it's just a large it's just a
  178. 6:59numbers game. It's going to get smaller,
  179. 7:02right? It can't keep going higher every
  180. 7:04year.
  181. 7:05But what you have to understand is like
  182. 7:08companies are still spending
  183. 7:10billions and trillions of dollars to
  184. 7:12build out this AI infrastructure.
  185. 7:15And
  186. 7:16as I said, revenue is going to be a big
  187. 7:18deciding factor on all this. Revenue
  188. 7:20hasn't really come in yet.
  189. 7:23Okay, we're still early. So it will
  190. 7:26eventually start to come in because
  191. 7:28yesterday alone we had couple great
  192. 7:31earnings
  193. 7:32from companies like AMD.
  194. 7:35AMD reported
  195. 7:37significant growth in their data center
  196. 7:39businesses. I think they were at one
  197. 7:40point they were up like 6 or 7% after
  198. 7:42hours.
  199. 7:43AMD generated $13.4 billion in total
  200. 7:47revenue representing a 45%
  201. 7:50year-over-year growth in its data center
  202. 7:52business, generated approximately 8 and
  203. 7:551/2 billion dollars, up 97%
  204. 7:58from previous last year.
  205. 8:00It doubled in 12 months.
  206. 8:02And AMD isn't the only one benefiting
  207. 8:04from this expansion. Arista Networks has
  208. 8:07been
  209. 8:08reported strong demand for high
  210. 8:10performance networking equipment. They
  211. 8:11also had their earnings yesterday.
  212. 8:13Massive beat. Supermicro, another name
  213. 8:17experienced significant growth tied to
  214. 8:19AI server developments.
  215. 8:22There's so many earnings that came out
  216. 8:24yesterday all pointed that AI demand is
  217. 8:28stronger than ever before. And of course
  218. 8:30the stocks are down or a little bit flat
  219. 8:32because of weakness in semis and that
  220. 8:34that's fine.
  221. 8:36And even look at GPU rentals, okay?
  222. 8:39Nvidia B200 and H100 GPUs have had
  223. 8:43significantly
  224. 8:44like went up during the year and it's
  225. 8:46still continued to go up.
  226. 8:48And H100 is isn't even the Nvidia's
  227. 8:51newest generation. It's it's so old.
  228. 8:54If like these older computing hardware
  229. 8:57is still commanding rental high prices,
  230. 8:59suggest that there may be more
  231. 9:01underlying demand for capacity than
  232. 9:03compute is available. Okay? I I you guys
  233. 9:06need to understand it. Understand this.
  234. 9:09This is compute need for compute is
  235. 9:11going up. It's not slowing down.
  236. 9:15Now let's let me bring it back to the
  237. 9:16semiconductor sell off because I think
  238. 9:18investors this is where investors could
  239. 9:20be missing the opportunity.
  240. 9:22I think the market seems to be assuming
  241. 9:24that tighter restrictions on data center
  242. 9:26will translate into lower demand for
  243. 9:29semiconductors, but I don't think it's
  244. 9:31true. Even if data center approvals
  245. 9:34become difficult, that doesn't mean
  246. 9:35semiconductor demand disappears. It's
  247. 9:38this existing this facilities still need
  248. 9:40more powerful GPUs,
  249. 9:42bandwidth memory networking equipment,
  250. 9:45yada yada yada. And if developers can
  251. 9:47build any new facilities as they were
  252. 9:49originally planned,
  253. 9:50they have a greater incentive to
  254. 9:53maximize the computing capacity of the
  255. 9:54infrastructure they already have.
  256. 9:57Think about it. If Amazon or Microsoft
  257. 9:59already has an operational data center
  258. 10:01with access to electricity, cooling
  259. 10:04systems, yada yada yada, that facility
  260. 10:07becomes increasingly valuable if new
  261. 10:09construction becomes more difficult.
  262. 10:11So if you're a neo cloud provider that
  263. 10:13has
  264. 10:15these available capacity, your
  265. 10:18your spot price just went up
  266. 10:20because if there is less compute to go
  267. 10:23around,
  268. 10:24who benefits? People do that have
  269. 10:26computing capacity.
  270. 10:28And now look at this chart.
  271. 10:30This is the AI stack right here.
  272. 10:33Broadcom companies like Nvidia, AMD,
  273. 10:35Broadcom, Credo, Light, AAAOI
  274. 10:39could continue benefiting. Let's start
  275. 10:41with Nvidia. We all know Nvidia remains
  276. 10:43one of the most suppliers in the entire
  277. 10:45ecosystem. Regardless of whether
  278. 10:47Microsoft or Google or another company
  279. 10:49starts to slow down the AI race, Nvidia
  280. 10:51continues selling their accelerator
  281. 10:53networking technologies to help power
  282. 10:56their infrastructure. And then AMD is in
  283. 10:59the same position. Broadcom is another
  284. 11:01company I think particularly interesting
  285. 11:04because it's exposure to custom silicon
  286. 11:06and networking. And Marvell. Marvell is
  287. 11:09thinking about is going to acquire Light
  288. 11:11and they're going to that's going to
  289. 11:12create a massive moat in their business
  290. 11:14as well. So these are all some of the
  291. 11:17companies that are benefiting from the
  292. 11:19AI infrastructure demand.
  293. 11:21And then
  294. 11:22as I mentioned like Credo, Lumentum, Ply
  295. 11:25Electronics are like helping remove one
  296. 11:28of the bottlenecks in the industry,
  297. 11:29which is like networking. And then we
  298. 11:31have stuff like memory, companies like
  299. 11:33Micron, SK Hynix, Samsung has
  300. 11:36you know benefited from growing demand
  301. 11:38for high bandwidth memory and other
  302. 11:39memory products.
  303. 11:41So you get it. You have to understand
  304. 11:43there's so many bottlenecks right now
  305. 11:44like power, memory, substrates,
  306. 11:48um CPU, GPU like there's the entire
  307. 11:52supply chain is constrained right now.
  308. 11:55And you have to understand demand for
  309. 11:57all this is only going higher.
  310. 12:00And that brings me to the broader
  311. 12:02opportunity. I think the investors are
  312. 12:04still trading the AI infrastructure
  313. 12:07as though every company faces the same
  314. 12:09regulatory and financial financing
  315. 12:11risks, but the reality is much more
  316. 12:13complicated.
  317. 12:14A semiconductor selling into multiple
  318. 12:17hyperscalers is very different from a
  319. 12:19neo cloud.
  320. 12:20A or a lower tier, I should say lower
  321. 12:24tier neo cloud that borrows billions of
  322. 12:25dollars to fund build a new facility in
  323. 12:27Texas.
  324. 12:29And even within the neo cloud industry,
  325. 12:31I've told you guys this Core Weave and
  326. 12:33um Nevius still remains the strongest.
  327. 12:36They can actually get the funding. And
  328. 12:38so I think this this sell off could
  329. 12:40actually create some very attractive
  330. 12:42points. And um but I would definitely
  331. 12:45focus on companies with strong demand,
  332. 12:47real customers,
  333. 12:49reasonable financing needs and a
  334. 12:51credible path flow to generating cash
  335. 12:54flow.
  336. 12:55I think there's a lot of opportunities
  337. 12:57that you can buy here. This is a buy the
  338. 12:59dip moment. I'm going to buy some
  339. 13:02companies here. And if you want to know
  340. 13:04what companies I'm buying, you know, you
  341. 13:05can just subscribe to Milk Road for just
  342. 13:07just it's just a dollar. You can come
  343. 13:08check it out.
  344. 13:09>> Thanks for listening to Milk Road. If
  345. 13:11you enjoyed the show, make sure you like
  346. 13:12and subscribe. [music] And if you're
  347. 13:13struggling to find winners in the
  348. 13:15market, that's exactly what Milk Road
  349. 13:16Pro is built for. [music] Our analysts
  350. 13:18have called some of the biggest winners
  351. 13:20early and Pro lets you see what they're
  352. 13:22buying next, every trade they make, and
  353. 13:24the research behind every position.
  354. 13:25Check out Milk Road Pro at the link
  355. 13:27below. Everything you hear on Milk Road
  356. 13:29is for informational purposes only.
  357. 13:31These are our personal opinions, not
  358. 13:33financial advice, and we may own some of
  359. 13:35the investments we talk about. Always do
  360. 13:36your own research and make the decisions
  361. 13:38that are right for you. See you next
  362. 13:40time.

About this transcript

This page contains the full transcript of Don't Buy Every AI Dip. Buy the Companies Controlling THIS by Milk Road Stocks, generated from the public captions YouTube serves with the video. The transcript has 2,061 words across 362 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.