YouTube2Text

Micron’s Powerful AI Memory Advantage Could Drive a Major Rerating — Transcript

by Rick Orford - Trading Stocks and Options For All · 2,718 words · 467 segments · language en · Watch on YouTube

Full transcript

  1. 0:00Micron could be turning into one of the
  2. 0:03most must-buy AI stocks this year, but
  3. 0:06not for the reason that most investors
  4. 0:08are watching.
  5. 0:09The AI boom is creating what I'm calling
  6. 0:11a memory vacuum. Every new generation of
  7. 0:14AI chips needs dramatically more memory,
  8. 0:17and that demand is locking up the world
  9. 0:19supply of DRAM wafers faster than
  10. 0:23producers can expand it. The usual
  11. 0:25objection here is that memory is
  12. 0:27cyclical, and historically, it's
  13. 0:29absolutely true.
  14. 0:30But AI hardware is moving towards a
  15. 0:33roughly 2-year replacement cycle
  16. 0:36compared with the old 5-year PC cycle,
  17. 0:39and that could make Micron's revenue
  18. 0:42more predictable, more structural, and
  19. 0:44potentially worthy of a higher
  20. 0:46valuation. That's why in this video,
  21. 0:49you're going to learn what's creating
  22. 0:50this memory vacuum, why Micron may have
  23. 0:53pricing power it's never had before, and
  24. 0:56how this shift could affect Micron stock
  25. 0:59from here. But before we begin, if
  26. 1:01you're new to the channel, hello. My
  27. 1:02name is Rick Warford. I'm a Wall Street
  28. 1:04Journal best-selling author, been
  29. 1:05trading since 1999, and no, I'm not a
  30. 1:07financial advisor. But that's a good
  31. 1:09thing, cuz I'm right here with you
  32. 1:11breaking down the numbers so retail
  33. 1:13investors like us can make smarter, more
  34. 1:16confident decisions with our money.
  35. 1:18I want to thank The Motley Fool for
  36. 1:20sponsoring this video. The Motley Fool
  37. 1:23is a company that provides investing
  38. 1:25insight and stock recommendations for
  39. 1:27investors of all skill sets and risk
  40. 1:29levels. And you all know how much I love
  41. 1:31researching new stocks and trying to
  42. 1:33find the next best investment. So, I'm
  43. 1:37proud to partner with The Motley Fool to
  44. 1:39bring you 10 stock picks from their
  45. 1:41popular product Stock Advisor.
  46. 1:44Stock Advisor has beaten the market by
  47. 1:46almost six times. Go to fool.com/ricko
  48. 1:50to get your 10 stock picks right now.
  49. 1:53Okay, so Micron produces three major
  50. 1:57memory technologies. DRAM, and NAND, D,
  51. 2:01and NOR. DRAM is the working memory that
  52. 2:03we use in servers, PCs, and networking
  53. 2:06equipment. NAND D flash stores data in
  54. 2:10products like solid state drives and
  55. 2:12smartphones.
  56. 2:14But then we've got the high bandwidth
  57. 2:16memory or HBM, and that's become one of
  58. 2:19the most important components inside
  59. 2:21modern AI systems.
  60. 2:24This is specialized memory built to move
  61. 2:26enormous amounts of data to and from
  62. 2:29advanced processors at extremely high
  63. 2:31speeds. Now, modern AI chips from
  64. 2:34companies like Nvidia rely on HBM from
  65. 2:37suppliers including Micron to keep their
  66. 2:40GPUs fed with data.
  67. 2:42Because without enough memory bandwidth,
  68. 2:45even the most powerful AI processor
  69. 2:47can't perform at its full potential.
  70. 2:50That puts Micron in a sort of a critical
  71. 2:53position inside the AI supply chain. I
  72. 2:56mean, it's not building the GPU, but
  73. 2:59it's supplying one of the components
  74. 3:01that increasingly determines
  75. 3:04how much useful work that GPU actually
  76. 3:06does.
  77. 3:08AI's next major bottleneck may actually
  78. 3:10not be compute alone. I actually think
  79. 3:12it might be memory. And that's really
  80. 3:14the memory vacuum. Every time the
  81. 3:17industry builds a more powerful AI
  82. 3:19accelerator, it's going to need more
  83. 3:21memory capacity, more memory bandwidth,
  84. 3:24and more storage around it. Yet the
  85. 3:26supply of DRAM wafers can't instantly
  86. 3:29expand to match any of that demand.
  87. 3:32So, AI isn't simply creating another
  88. 3:34strong quarter for Micron. It may
  89. 3:36actually be locking up a growing share
  90. 3:39of the industry's available capacity for
  91. 3:41years to come. And that matters. That
  92. 3:44distinction matters because Micron has
  93. 3:46historically operated
  94. 3:49in one of the most cyclical parts of the
  95. 3:51semiconductor market. Hear me out here.
  96. 3:54Historically, strong demand encouraged
  97. 3:57memory suppliers to increase production.
  98. 4:00But then supply eventually caught up,
  99. 4:02prices fell, and profits weakened. Then
  100. 4:04producers reduced investment until
  101. 4:07demand recovered once again, and the
  102. 4:09cycle started all over.
  103. 4:11And that history explains why investors
  104. 4:13have rarely placed a premium multiple on
  105. 4:16Micron. They've learned not to assume
  106. 4:19that the peak of memory profits is going
  107. 4:21to last. But this cycle looks like it's
  108. 4:24developing differently.
  109. 4:26Micron's latest quarter wasn't just an
  110. 4:27earnings beat, it strengthened the
  111. 4:30argument that the memory industry has
  112. 4:32entered a new phase driven by expanding
  113. 4:35AI infrastructure investment, longer
  114. 4:37customer commitments, and structural
  115. 4:39constraints on supply.
  116. 4:42Micron expects industry demand for both
  117. 4:44DRAM and NAND to continue outpacing
  118. 4:46supply, and it also expects tighter
  119. 4:48supply conditions to extend beyond 2027
  120. 4:52into 2028, all driven by accelerating AI
  121. 4:55demand and structural supply
  122. 4:57constraints. And one of the biggest
  123. 4:59reasons is the replacement cycle for AI
  124. 5:03hardware.
  125. 5:04If AI hardware moves towards a roughly
  126. 5:07two-year replacement cycle, memory
  127. 5:10demand can refresh far more frequently
  128. 5:13than it did during the old PC era. And
  129. 5:16that could make Micron's revenue much
  130. 5:19less dependent on one short burst of
  131. 5:21demand,
  132. 5:22and more connected into a recurring
  133. 5:25infrastructure cycle.
  134. 5:27In other words, AI memory may still be
  135. 5:29cyclical, but the cycle itself could
  136. 5:32actually become shorter, stronger, and
  137. 5:34more predictable. And if that happens,
  138. 5:37investors may finally eventually stop
  139. 5:40valuing Micron like some ordinary
  140. 5:43commodity producer.
  141. 5:44That difference is very important. A
  142. 5:46temporary shortage creates a brief
  143. 5:49earning spike.
  144. 5:50A structural shortage supported by
  145. 5:53faster replacement cycles and multi-year
  146. 5:55commitments,
  147. 5:57well, this can create better visibility
  148. 5:58into future demand. And that is the
  149. 6:00foundation of the higher valuation
  150. 6:03argument. But before I get any deeper
  151. 6:06into what this could mean for Micron's
  152. 6:08valuation, I want to give you a quick,
  153. 6:10or I want to give my community a quick
  154. 6:12shout-out. It's one of the
  155. 6:13fastest-growing investing communities
  156. 6:15here on YouTube and we talk about
  157. 6:17specific companies, what they're doing,
  158. 6:19and even options trading. Now, best part
  159. 6:22is it's completely free to join. Just
  160. 6:25scan the QR code on the screen or check
  161. 6:27for the link in the description and I'll
  162. 6:28see you there.
  163. 6:30Now, the current data center numbers,
  164. 6:32they help explain why customers are
  165. 6:34trying to secure supply so far in
  166. 6:36advance. AI is driving massive growth in
  167. 6:40data centers and Micron expects industry
  168. 6:43data center DRAM and NAND D uh shipments
  169. 6:46in 2026 to more than double from just 2
  170. 6:49years earlier.
  171. 6:51Why is it? Well, the rise in a genetic
  172. 6:53AI is definitely one part of that shift.
  173. 6:55So, instead of using only racks filled
  174. 6:58with AI accelerators, these newer
  175. 7:00deployments are increasingly requiring
  176. 7:02dedicated CPU and storage racks.
  177. 7:06And these systems need to manage AI
  178. 7:08agents, execute programs, and store
  179. 7:10rapidly expanding amounts of contextual
  180. 7:12data.
  181. 7:13This means that data or the memory
  182. 7:15opportunity isn't limited to HBM
  183. 7:18attached directly to GPUs.
  184. 7:20AI can increase demand across the
  185. 7:22broader data center market but and
  186. 7:24including standard DRAM and NAND D
  187. 7:26storage. And Micron now expects global
  188. 7:29server shipments to grow at a high teens
  189. 7:32percentage rate. And that's up from a
  190. 7:34previous forecast for low double-digit
  191. 7:37growth.
  192. 7:38And that's where Micron's strategic
  193. 7:40customer agreements or SCAs become so
  194. 7:43important. These are the multi-year
  195. 7:45supply agreements that allow
  196. 7:47hyperscalers and other customers to
  197. 7:49secure the memory capacity that they
  198. 7:52expect to need.
  199. 7:53And Micron said it's signed 16
  200. 7:56agreements that represent roughly 20% of
  201. 7:59its DRAM volume and a third of its NAND
  202. 8:03D volume over their applicable periods.
  203. 8:06And multiple market commentators have
  204. 8:07described Micron's capacity of as as
  205. 8:10effectively sold out through the end of
  206. 8:12the year.
  207. 8:13Now, whether investors call it sold out,
  208. 8:15allocated, or committed, the basic point
  209. 8:18is that customers are competing
  210. 8:20for constrained amount of future supply.
  211. 8:23And that creates some pricing power
  212. 8:26that Micron has very rarely enjoyed. A
  213. 8:29hyperscaler built spending billions of
  214. 8:31dollars on AI can't afford to leave
  215. 8:33those systems unfinished just because it
  216. 8:36failed to secure enough memory.
  217. 8:38So, customers, well, they commit
  218. 8:40earlier, they sign longer agreements,
  219. 8:43and they accept stronger pricing to
  220. 8:46protect their own build-out plans. And
  221. 8:49Micron believes its earnings boom may
  222. 8:50actually still have a lot of room to
  223. 8:52run.
  224. 8:52HBM and DRAM are becoming strategic
  225. 8:55components of the AI build-out rather
  226. 8:57than products that customers casually
  227. 8:59purchase whenever they need them.
  228. 9:02And it also helps explain why management
  229. 9:05is still optimistic after such
  230. 9:07record-breaking performance.
  231. 9:09Micron thinks that it can achieve
  232. 9:11another record quarter
  233. 9:13guiding towards $50 billion in fiscal
  234. 9:16fourth quarter revenue at a margin a
  235. 9:18gross margin around 86%.
  236. 9:21Management also thinks that it can carry
  237. 9:23that momentum into the following quarter
  238. 9:25after that.
  239. 9:27In Q3, Micron's DRAM revenue reached
  240. 9:30about 31.3 billion, and that was about
  241. 9:3376% of total revenue. And that helped
  242. 9:36push total revenue to 41 and 1/2
  243. 9:39billion, up 346%
  244. 9:41from the previous year.
  245. 9:43Operating margin on a GAAP basis surged
  246. 9:45to 80.4%,
  247. 9:47meaning Micron generating more than 80
  248. 9:50cents of operating income for every
  249. 9:52single dollar of revenue they took in.
  250. 9:55Those numbers make the company look much
  251. 9:57less like a traditional manufacturer,
  252. 9:59and more like a cloud provider with
  253. 10:01software-like margins, right?
  254. 10:04And the next wave of demand is probably
  255. 10:06going to extend beyond AI data centers.
  256. 10:08Micron expects AI-enabled smartphones,
  257. 10:11premium PCs, and new consumer devices to
  258. 10:14become future growth drivers, all as AI
  259. 10:17goes into them.
  260. 10:19And Micron also expects industrial
  261. 10:21automation, autonomous vehicles, and
  262. 10:23robotics to all go to support long-term
  263. 10:26demand for memory and storage. And that
  264. 10:28brings the story
  265. 10:30all the way back for to the stock.
  266. 10:33Micron shares have traded between $103
  267. 10:36and $1,255
  268. 10:38over the last 52 weeks. And during that
  269. 10:40period, the stock is up almost 700%.
  270. 10:44The stock rallied ahead of third-quarter
  271. 10:46earnings as investors anticipated, they
  272. 10:48guessed there was going to be another
  273. 10:50blowout quarter driven by AI memory
  274. 10:52demand, and there was.
  275. 10:54And when Micron reported, it exceeded
  276. 10:56both expectations for revenue and
  277. 10:58earnings. And management highlighted
  278. 11:00growing demand for DRAM and HBM, along
  279. 11:03with the long-term customer agreements,
  280. 11:05all to support future volume. But after
  281. 11:08the stock reached some new highs, I
  282. 11:11think some investors locked in some
  283. 11:13gains. Even after record results, many
  284. 11:15investors still see Micron as a cyclical
  285. 11:17memory company whose earnings growth may
  286. 11:20not last. And I get it, that skepticism
  287. 11:23is exactly why the valuation debate is
  288. 11:25so important. Micron trades at about 9.2
  289. 11:29times earnings.
  290. 11:30So, investors are paying roughly $9.20
  291. 11:33for every dollar of earnings that the
  292. 11:35company generates. That is an incredibly
  293. 11:37low multiple for a company growing this
  294. 11:40fast. But, the market isn't valuing
  295. 11:43Micron based only on its current
  296. 11:45earnings. It's
  297. 11:47discounting the possibility
  298. 11:49that those earnings are near a cyclical
  299. 11:52peak. That's the thing. Today's profits
  300. 11:55are being driven by tight DRAM supply
  301. 11:58and high HBM demand. Major producers are
  302. 12:01working to increase memory production.
  303. 12:03So, investors expect supply to
  304. 12:05eventually catch up.
  305. 12:07And they think that when that happens,
  306. 12:09pricing power could weaken.
  307. 12:11But, the bullish argument is that the
  308. 12:13market is probably using the wrong
  309. 12:17historical template.
  310. 12:19If AI hardware is replaced every 2 years
  311. 12:22instead of 5,
  312. 12:23demand will renew before the industry
  313. 12:26has fully worked through the previous
  314. 12:27wave.
  315. 12:28You know, if customers they continue
  316. 12:30signing these multi-year agreements, a
  317. 12:33larger portion of Micron's volume
  318. 12:35becomes visible in advance. And if AI
  319. 12:38keeps consuming more memory per system,
  320. 12:41new supply may actually be absorbed more
  321. 12:44quickly than investors expect.
  322. 12:47Those three forces could make Micron's
  323. 12:50revenue much more structural. Structural
  324. 12:53doesn't mean perfectly stable. No, but
  325. 12:55it does mean that demand is being
  326. 12:57supported by a long-term change in how
  327. 13:00computing systems are designed and
  328. 13:03replaced rather than by a like a
  329. 13:06temporary inventory restocking cycle.
  330. 13:09And that distinction is what could
  331. 13:11justify a higher price-to-earnings ratio
  332. 13:14over time. I mean, the market doesn't
  333. 13:16need to value Micron like Nvidia for the
  334. 13:19stock to benefit.
  335. 13:20It only needs to decide that Micron
  336. 13:22deserves something higher
  337. 13:24than a deeply discounted cyclical
  338. 13:26multiple. And that's the potential
  339. 13:29re-rating that I'm talking about. This
  340. 13:31is when investors decide a company's
  341. 13:33future earnings are more dependable
  342. 13:37and they become willing to pay a higher
  343. 13:39multiple for each dollar of profit. Now,
  344. 13:42if you've been watching Micron, let's
  345. 13:44say you like the stock but you think
  346. 13:46just too expensive right now, especially
  347. 13:49after a massive run. But maybe you'd be
  348. 13:51happy to own it 15 or 20% below where it
  349. 13:54is today.
  350. 13:55Check this video up here because I talk
  351. 13:57about how you can sell a put option on
  352. 14:00an AI stock you'd want to own and get
  353. 14:03paid to wait for it to potentially come
  354. 14:05down to your price. But get this, if it
  355. 14:07does come down to your price, you get to
  356. 14:09buy it but you got paid to buy it as
  357. 14:11well. And Wall Street is already very
  358. 14:14bullish on Micron. The consensus among
  359. 14:1640 analysts, they rate the stock a
  360. 14:18strong buy and that rating is held over
  361. 14:21the last quarter.
  362. 14:22Micron's average score is 4.68 out of
  363. 14:25five and it's actually up from 4.62 just
  364. 14:28three months earlier. And the price
  365. 14:30targets, they suggest is between 52 and
  366. 14:33104%
  367. 14:34potential upside. Saying analysts are
  368. 14:36saying the stock could double over the
  369. 14:38next year from here.
  370. 14:40And I got to say that is definitely
  371. 14:42possible. I mean, the compute shortage
  372. 14:44has definitely benefited companies like
  373. 14:46Nvidia, Broadcom. But those processors
  374. 14:50can't deliver any of their full
  375. 14:51potential without enough memory.
  376. 14:54As demand for compute rises, that
  377. 14:56bottleneck increasingly shifts towards
  378. 14:59HBM, DRAM, and storage. And Micron may
  379. 15:03be one of the largest beneficiaries
  380. 15:05because it's right in the middle of that
  381. 15:07constraint.
  382. 15:09So, the bullish case comes down to three
  383. 15:10connected ideas. First, AI systems
  384. 15:13require more memory with each new
  385. 15:15generation.
  386. 15:16Second, supply constraints and long-term
  387. 15:19agreements are giving Micron some
  388. 15:20unusual visibility and some very strong
  389. 15:23pricing power.
  390. 15:25And third, a faster hardware replacement
  391. 15:28cycle could make demand more recurring
  392. 15:31and justify a higher valuation. Of
  393. 15:33course, the bearish case is still very
  394. 15:35real. Memory producers will eventually
  395. 15:38add capacity. Supply conditions probably
  396. 15:41will improve. And if supply catches up
  397. 15:44faster than expected prices and margins
  398. 15:47could fall. Micron itself expects
  399. 15:49industry supply to begin improving in
  400. 15:522028. But the size of the current
  401. 15:55shortage
  402. 15:56suggests that normalization may take
  403. 15:58years rather than quarters to complete.
  404. 16:01And before any supply fully catches up,
  405. 16:04of course, demand is still going to
  406. 16:05expand and still could expand again
  407. 16:08through more powerful AI accelerators,
  408. 16:11AI-enabled devices, and more intelligent
  409. 16:13machines. And that's why I don't think
  410. 16:15the Micron debate can be reduced to
  411. 16:18simply memory is cyclical. Yes, memory
  412. 16:21is cyclical.
  413. 16:22But cycles can change.
  414. 16:24A 5-year cycle can turn into a 2-year
  415. 16:27cycle. Very different, right? Multi-year
  416. 16:30customer agreements, they're also very
  417. 16:32different from short-term purchasing
  418. 16:35agreements. And capacity that's
  419. 16:37effectively committed well in advance
  420. 16:40well, that creates a very different
  421. 16:42pricing environment. Micron doesn't need
  422. 16:45to stop being cyclical to become a more
  423. 16:48valuable business.
  424. 16:49It just needs its revenue to become more
  425. 16:52predictable.
  426. 16:53Its customer relationships to be more
  427. 16:55durable.
  428. 16:56And its pricing power to last longer
  429. 16:59than the market currently expects. If
  430. 17:01those changes hold
  431. 17:04Micron could be turning from a
  432. 17:06boom-and-bust memory supplier into one
  433. 17:08of the most important infrastructure
  434. 17:10companies in the AI economy. And that
  435. 17:13would make today's low earnings multiple
  436. 17:16look much less like a warning and much
  437. 17:18more like an opportunity. So, if you
  438. 17:20think the memory vacuum will persist, AI
  439. 17:23if you think AI hardware will keep
  440. 17:25refreshing quickly, and if you think
  441. 17:26that long-term agreements will protect
  442. 17:28Micron's earnings,
  443. 17:30then the stock may still have some
  444. 17:32meaningful upside despite its enormous
  445. 17:34run.
  446. 17:35But, if you think that earnings have
  447. 17:37already peaked and supply will actually
  448. 17:39catch up sooner than later, perhaps
  449. 17:42waiting for a better entry point or for
  450. 17:44more evidence may be the smarter move.
  451. 17:48But, now I want to turn it over to you.
  452. 17:49Do you think AI has permanently changed
  453. 17:53Micron's memory cycle, or or is this
  454. 17:55still just another boom that ends when
  455. 17:58supply catches up? And did I get any of
  456. 18:01this right? Did I forget anything? Let
  457. 18:04me know all of that and more in the
  458. 18:05comments below. And while you're there,
  459. 18:07if you found the video helpful, don't
  460. 18:09forget to like and subscribe because it
  461. 18:10really does help others find the video.
  462. 18:12It supports the channel, and it makes
  463. 18:13sure that you don't miss out on my next
  464. 18:15deep dive. Well, that's it for me today.
  465. 18:18Want to thank you so much for watching.
  466. 18:19Hope to see you again here on YouTube or
  467. 18:21in my community. Bye for now.

About this transcript

This page contains the full transcript of Micron’s Powerful AI Memory Advantage Could Drive a Major Rerating by Rick Orford - Trading Stocks and Options For All, generated from the public captions YouTube serves with the video. The transcript has 2,718 words across 467 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 translate the transcript into 100+ languages.

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, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.