Brad Gerstner: No AI Bubble, Semis Eat the Nasdaq & AI's Take Off Problem — Transcript

Full transcript

1. 0:00Our favorite fifth bestie in the world,
2. 0:03the one, the only Brad Bersner from
3. 0:06Altimter. Brad has had unbelievable
4. 0:08career starting five companies. So, he's
5. 0:11got a very different mentality than your
6. 0:13sort of classic hedge fund guy.
7. 0:15>> He's an amazingly successful guy and
8. 0:17he's put up a tremendous amount of
9. 0:19money.
10. 0:19>> Every child in America, all 70 million
11. 0:23kids under the age of 18 deserve to have
12. 0:25one of these accounts. This would not be
13. 0:27a law if Brad Gerster did not pursue it
14. 0:31with absolute dogged determination.
15. 0:34>> This is not a program. This is a
16. 0:36platform. It is the largest unlock of
17. 0:39direct philanthropy in the history of
18. 0:40the country.
19. 0:41>> I think the antidote to more socialism
20. 0:43is more capitalism.
21. 0:46Please welcome Brad Gersner.
22. 0:54Woo.
23. 0:56Wow.
24. 0:58Let's go.
25. 1:02Let's go.
26. 1:04Thank you guys. And thank you for so
27. 1:06much love yesterday, especially on the
28. 1:08Trump accounts. So many people came up.
29. 1:12Everybody gets what this means for
30. 1:14America. We're in a battle for the soul
31. 1:16of America. 70 million kids are going to
32. 1:19be made direct owners in America. That
33. 1:22is how we beat the scourge of capital or
34. 1:25socialism. We make every child a
35. 1:27capitalist.
36. 1:32>> And also
37. 1:34thank you to all the people yesterday
38. 1:36who came up who took the CAC scan, the
39. 1:39heart scan outside from the center for
40. 1:41heart attack prevention which we
41. 1:43started. There's no doubt based on these
42. 1:45results we uh you know we're going to
43. 1:48save some lives even yesterday. Okay.
44. 1:51This is the highest ROI thing you can do
45. 1:54in healthcare. Every cardiologist I talk
46. 1:57to does this for themselves, their
47. 1:59family, and their friends. It's a
48. 2:00hundred bucks, 15 minutes. Get it done.
49. 2:03If we turn this into the mammogram for
50. 2:05the heart, we'll save 50,000 lives a
51. 2:07year in this country. It's like ending a
52. 2:09Ukraine war in America every year. So,
53. 2:11go get your CAC scan done if you have
54. 2:13it. But today is not about those two
55. 2:14moonshots.
56. 2:16Today is kind of a throwback to what I
57. 2:18used to do on the pod with these guys,
58. 2:20which is a market check, a tech check,
59. 2:22state of the market. Where are we? Where
60. 2:24are we going? What do we have to believe
61. 2:26to be true in order for the market to
62. 2:28continue to work? We're going to do a
63. 2:30bit of a speed round here. So, bear with
64. 2:32me. Get your cameras out. Get your notes
65. 2:35out. Some good some good chart candy in
66. 2:37here for you guys. So markets up 15%
67. 2:41this year, up 39% since January of last
68. 2:43year, despite all the concerns about
69. 2:45tariffs, despite all the concerns about
70. 2:47geopolitics,
71. 2:49despite all the concerns about AI
72. 2:51regulation.
73. 2:52The scoreboard, we have a lot of people
74. 2:55in the bestie group who said gold was
75. 2:57going to be off the charts this year.
76. 2:58It's flat. Bitcoin's down 10%. But look,
77. 3:01we have Nvidia revenue up 2x.
78. 3:03Hyperscaler capex up 2x. Open AI and
79. 3:06Enthropics valuation up 2x. SpaceX up
80. 3:09two and a halfx in a pretty nasty
81. 3:11backdrop.
82. 3:13This is not about multiple expansion.
83. 3:16This is an earnings driven market
84. 3:19expansion. We've seen multiple
85. 3:21contraction this year. Earnings are up
86. 3:2326% of course driven a lot by AI
87. 3:26infrastructure, but the multiple on the
88. 3:28NASDAQ and the S&P is actually down.
89. 3:31Look at Nvidia trading at 14 times next
90. 3:34year's fully taxed gap earnings. This is
91. 3:37no bubble like it was in 2000.
92. 3:40NASDAQ, S&P, Socks, Nvidia, all trading
93. 3:44well below their average multiples.
94. 3:47Okay, MAG 7 basically in line with its
95. 3:51average multiple, but not everybody is
96. 3:53winning. Okay, at the bottom here,
97. 3:55consumer discretionary, software,
98. 3:57financials, huge sections of the market
99. 3:59have barely moved, right? This is a
100. 4:02market that is being driven by the
101. 4:04largest capex buildout, the largest
102. 4:06super cycle in the history of
103. 4:08technology.
104. 4:11Semiconductors are 70% of the NASDAQ's
105. 4:14return.
106. 4:1670% of the return. That's both good and
107. 4:18bad, and we'll get into that.
108. 4:22So, who's making the money? The makers
109. 4:23of the tokens are making the money and
110. 4:26the buyers of the tokens are basically
111. 4:28going along for the ride
112. 4:33because of the tightness in the
113. 4:35infrastructure market. We have massive
114. 4:38public companies that look like venture
115. 4:40capital returns.
116. 4:42Dell up 5x up 9x in just 18 months.
117. 4:49I love this chart. In blue you have the
118. 4:52hyperscaler capex.
119. 4:54In orange you have the free cash flow of
120. 4:57the semiconductor companies. Do you
121. 4:59notice anything?
122. 5:02Their capex is almost dollar fordoll
123. 5:04free cash flow to the infrastructure
124. 5:06companies.
125. 5:08Okay. Where were we at the start of the
126. 5:10year? This is a really important
127. 5:14framework to get your head around. You
128. 5:16may remember a certain podcast I did
129. 5:19with Sam Alman and Satcha in October of
130. 5:21last year. And I asked Sam a very basic
131. 5:24question that was on everybody's mind.
132. 5:26How can you commit to a trillion dollars
133. 5:28in capex when you have $13 billion of
134. 5:31gap revenue?
135. 5:33I thought this would be a way to clear
136. 5:35up some confusion in the market.
137. 5:37Instead, he told me to sell my shares.
138. 5:42But then in the beginning of December we
139. 5:44have opus 4.5 and claude code and then
140. 5:48in January anthropics revenue was two
141. 5:51billion and then in February it's 4
142. 5:53billion and in March it was 11 billion.
143. 5:56Do you notice anything that happened
144. 5:57there in March? People started figuring
145. 5:59out what anthropics revenue was. We had
146. 6:02a an exclamation point answer to the
147. 6:05question will the AI revenue showed up?
148. 6:07It showed up in a massive way and that's
149. 6:09why we got this historic run in April
150. 6:11and May. We ripped off the bottom
151. 6:14because the fuse was lit by Anthropic's
152. 6:16monthly revenue. And then in in June and
153. 6:20July, we had some consolidation. Why?
154. 6:22Because Anthropic came out and said,
155. 6:24"Oh, our annual run rate revenue is 65."
156. 6:26People thought it was 75. So they had to
157. 6:29revise down their estimates a little
158. 6:30bit. And we had some concerns about open
159. 6:33source. Is open source catching up? you
160. 6:35know, is Anthropic going to continue to
161. 6:36be able to generate those revenues? And
162. 6:39so, we've kind of moved sideways since
163. 6:40then.
164. 6:44It's hard to get your head around the
165. 6:46fact that these revenues have never
166. 6:49happened before in the history of
167. 6:51capitalism,
168. 6:53right? We are on parabolic double
169. 6:55exponential curves around these
170. 6:57revenues. So the collective run rate,
171. 7:00let's just call it of the top three
172. 7:01labs, Anthropic, Open AI, and SpaceX is
173. 7:05about a hundred billion based upon all
174. 7:08the rumors that are out there coming out
175. 7:10of July.
176. 7:12I think they need to collectively get to
177. 7:15at least $180 billion by the end of the
178. 7:17year, add another $80 billion across
179. 7:20those three labs just to keep the AI
180. 7:23trade intact.
181. 7:27So this is the single most important
182. 7:29data point in the market today, right?
183. 7:32Is Anthropics monthly revenue is
184. 7:34OpenAI's monthly revenue going to be 4
185. 7:36billion or 8 billion? It's almost hard
186. 7:38to get your head around. Most of us
187. 7:40who've been in venture capital for a
188. 7:42long time, if you added a billion, if
189. 7:44you got to a billion dollars in software
190. 7:46revenue over like four or five years,
191. 7:49you were in the top 5% of software
192. 7:51companies. But this is what the world is
193. 7:54now pricing in. So why is this so
194. 7:56important? It's this slide. If you're
195. 7:59going to build a a trillion half dollars
196. 8:02a year in capex, somebody has to pay for
197. 8:05it, right? Microsoft's not paying for
198. 8:08it. They're building it to rent it.
199. 8:10Google's not paying for it. They're
200. 8:12building it to rent it. Amazon's
201. 8:14building it to rent it. Well, who is the
202. 8:16person renting it? We have to have the
203. 8:18offtake revenues in order to pay that
204. 8:20rent. So if we exit this year around,
205. 8:23you know, let's call it 200 billion of
206. 8:25run rate revenue. That's below the the
207. 8:27years there. I think you have to go from
208. 8:29200 to 450 to 800 or a trillion dollars
209. 8:34just to keep up. The blue bar is the
210. 8:37expected capex just from the MAG 5 and
211. 8:40the orange bar is the offtake revenue,
212. 8:42the gap offtake revenue that we need to
213. 8:45see in order to keep this trade intact
214. 8:48over the course of the next few years.
215. 8:49Otherwise, we can't build this much
216. 8:51capex.
217. 8:53So, the labs are increasingly,
218. 8:56you know, like the conversation I had
219. 8:58with Sam last October, they're
220. 9:00aggressively expanding compute. Why? All
221. 9:02the things you heard yesterday, we're
222. 9:04heading into recursive cycles. They're
223. 9:06seeing, you know, incredible demand for
224. 9:08the product. And so, they're building
225. 9:10out compute. This year the total amount
226. 9:12of uh compute added so 2026 about 19
227. 9:16gawatt and about seven of those
228. 9:18gigawatts went to the two leading labs.
229. 9:21Next year this is semi analysis so Dylan
230. 9:23Patel forecast that the work that we're
231. 9:26going to add 43 gawatt and that about 14
232. 9:29gawatt are going to go to the leading
233. 9:32labs and we'll come back to this
234. 9:33question of can we really stand up 43
235. 9:37gawatts of compute next year. Notice
236. 9:39that the amount that we're adding next
237. 9:41year is as much as the total compute we
238. 9:43have in the United States this year, the
239. 9:46cumulative compute we have in the United
240. 9:49States this year. So that's what the
241. 9:51market is anticipating.
242. 9:53What happens if that doesn't happen? And
243. 9:56by the way, if you added those bars up,
244. 9:58by 2028, to David Sax's point yesterday,
245. 10:01over half of the total compute in the
246. 10:04country is controlled by two labs.
247. 10:07So it does the TAM exist? Does the TAM
248. 10:11exist? Right? So if we look at the total
249. 10:13TAM of knowledge work, right? That this
250. 10:16is a massive massive category. You got
251. 10:18consumer plus ads plus coding and all
252. 10:21these white collar workflows, millions
253. 10:24of enterprises. I would argue it's the
254. 10:25largest TAM in the history of the world.
255. 10:28You only have to get to about 4% of that
256. 10:30TAM or 1.2 trillion in order to pay for
257. 10:33the capex. So I don't think it's a TAM
258. 10:35issue. Um, obviously Jensen was on our
259. 10:39was on the pod and he he he talked two
260. 10:41years ago that inference was going to 1
261. 10:43billionx and remember all the people
262. 10:45saying he's full of there's no way this
263. 10:47can 1 billionx. That's exactly what
264. 10:49we've done in the a age of agents,
265. 10:52right? We've had this exponential token
266. 10:54growth this year. 47 quadrillion tokens
267. 10:58that are going to be produced. So it's
268. 11:00not a question again of demand. Codeex
269. 11:02users have grown 40x in the last eight
270. 11:05months and knowledge work at enterprises
271. 11:08if you look at the median amount that
272. 11:10enterprises are spending about 17x over
273. 11:13the course of the last 18 months and
274. 11:14I've talked to many people in the
275. 11:16audience here small businesses medium
276. 11:18businesses large businesses businesses
277. 11:20like alimter we can't operate our
278. 11:22business without buying AI so this is
279. 11:26not overly surprising to me and when we
280. 11:29think about the productivity dividend to
281. 11:30the economy me. You know, if you look at
282. 11:332015 to 2025, EPS growth was about 10%.
283. 11:38That was 6% revenue plus about 38 bips
284. 11:41of margin expansion every year out of
285. 11:42the NASDAQ. So, here's the question. Can
286. 11:45we turn the 38 bips to 100 bips of
287. 11:47margin expansion because of AI? The
288. 11:50answer is obviously yes. Every company I
289. 11:52talk to, Uber says we're going to grow
290. 11:5420%, we're not going to grow headcount.
291. 11:56Snowflake says we're going to grow 30%,
292. 11:58we're not going to grow headcount.
293. 11:59That's what's happening. That is margin
294. 12:01expansion. The single largest cost input
295. 12:03to every one of these companies are
296. 12:05humans and engineers. It's not that
297. 12:07they're going to fire everybody. They're
298. 12:08just not going to hire them at the rate
299. 12:09that they hired him before. And then of
300. 12:11course, we're going to have consumer
301. 12:12agents in everybody's pocket. You may
302. 12:14remember this bet I had with Bill. You
303. 12:16know, when are we going to be able to
304. 12:17book a hotel using your consumer agent
305. 12:20in your pocket? I think we've just
306. 12:21gotten that with Muse and with Instinct.
307. 12:24This could be another trillion dollar
308. 12:26category, but it's definitely going to
309. 12:27consume massive tokens. Here are the
310. 12:31three risks and challenges. And sorry
311. 12:32I'm going so quick, but I want to keep
312. 12:34it keep it moving. Regulation. We heard
313. 12:37a lot about this yesterday. Power. And
314. 12:40then what's going on with interest
315. 12:42rates? So, you know, this is the
316. 12:43regulation tugof-war,
317. 12:45right? Um, and I I've heard from a lot
318. 12:48of people, they're like, you're on both
319. 12:49sides of the issue. Here's the fact,
320. 12:52okay, the answer is not going to be on
321. 12:55one end or the other. We're going to
322. 12:58have to have common sense, pragmatic
323. 12:59solutions that get my mom, my sister,
324. 13:02and the br brother off the cliff, right?
325. 13:05We need to give comfort and confidence
326. 13:07to the people who elect our
327. 13:09representatives that it's safe. And so,
328. 13:11you heard Elon yesterday give a great
329. 13:13suggestion around peer review. So, I'm
330. 13:15confident that we're going to get there,
331. 13:17but it's kind of messy the sausage
332. 13:18making along the way. But we have a
333. 13:21prior history of excess regulation. When
334. 13:24people get scared, when activists start
335. 13:27pushing an agenda, we shut down 67
336. 13:30fision reactors in this country. We
337. 13:32unilaterally disarmed against China.
338. 13:35It's been a disaster for the country.
339. 13:37Right? All the clean energy we could
340. 13:39have gotten. Instead, we've gotten
341. 13:41non-clean emissions because we had a
342. 13:43group of activists who were hellbent on
343. 13:45shutting down nuclear. can't allow this
344. 13:47to occur to AI.
345. 13:53So regulation's a threat. It's a risk.
346. 13:57Atoms and energy are hard. So getting
347. 14:00back to can we stand up 43 gawatt of
348. 14:02compute. Our total compute in the
349. 14:04country is less than 40 gawatt. Doing
350. 14:06this in one year, we've got to overcome
351. 14:08permitting and local opposition. You
352. 14:10guys see all of that. grid
353. 14:12interconnection delays, skilled labor
354. 14:14shortages, power equipment is sold out.
355. 14:16It's the largest buildout in the history
356. 14:18of the country. I would suggest Dylan's
357. 14:20forecast to 43 gawatt next year is too
358. 14:23aggressive. I don't think we're going to
359. 14:25get there. I think the total amount
360. 14:27we're actually going to stand up is
361. 14:28somewhere closer to 25 gawatt. And I
362. 14:31think of that 25 gawatt, half of it will
363. 14:34be for anthropic and open AI.
364. 14:38Um I think that's enough to generate the
365. 14:41revenue. Remember Anthropics revenue
366. 14:43reportedly this year 100 billion 110
367. 14:46billion. If they do that they're doing
368. 14:48it with a gigawatt and a half of
369. 14:49compute. So if they add another four or
370. 14:52five gawatts of compute that's certainly
371. 14:54enough to add another 100 billion in
372. 14:56revenue. So I don't think we need more
373. 14:58gigawatts to get to the revenue targets
374. 15:01for next year. But this is my hunch that
375. 15:03we're not going to get to the 43.
376. 15:05And then of course we're going to hear
377. 15:07more tomorrow. I think rate hikes are
378. 15:08coming. I think it's now over 90% chance
379. 15:10that we're going to have rate hikes
380. 15:12tomorrow. Why does this matter? Because
381. 15:13all of this now borrowed money, right?
382. 15:16We're borrowing money in order to stand
383. 15:17up these data centers. So the hurdle
384. 15:19rate for that money is going up. And um
385. 15:23so that's not only a challenge for data
386. 15:25centers, but remember, you know, as
387. 15:27Warren Buffett says that, you know,
388. 15:29interest rates are to stocks what
389. 15:32gravity is to matter,
390. 15:34right? Right? If you can earn 5 1.5 or
391. 15:366% on your money without taking equity
392. 15:39risk, then you know it's going to be a
393. 15:41challenge for stocks. So, here's where I
394. 15:43think we are and we're going to end on
395. 15:45this slide and and then then we'll bring
396. 15:47the guys out and chop it up a little
397. 15:48bit. But this is the flight path. This
398. 15:50is how I think about managing the
399. 15:51portfolio, right? We're up about the
400. 15:55NASDAQ's up about 15%. But as I sit here
401. 15:58and think about the risk, do I want to
402. 15:59be small, medium, or large? This is how
403. 16:02I think about the fan of potential
404. 16:03outcomes. If the monthly AI lab revenues
405. 16:06are closer to that$8 billion number, I
406. 16:09think it's takeoff. Okay. I think we are
407. 16:11going to see an IPO this year. Um, and
408. 16:14so I'm paying very close attention to
409. 16:17what in fact are those monthly revenue
410. 16:19numbers. I think the trends are intact,
411. 16:21but we will see. That's going to be the
412. 16:22single most important thing as to
413. 16:24whether or not uh between now and the
414. 16:26end of the year uh we we have liftoff.
415. 16:28Second one is rates, the election, oil
416. 16:30prices. Obviously, rates are following
417. 16:32oil prices to a certain extent. So, what
418. 16:35happens there? You know, if rates were
419. 16:37to go to five and a half on the 10-year,
420. 16:38that's going to be a big burden uh you
421. 16:41know, on on the equity market. Then
422. 16:43finally, the regulation, the anthropic
423. 16:45IPO, we saw a trade down yesterday
424. 16:47because people are concerned that maybe
425. 16:49we're going to have a halt or or
426. 16:50postponement. That would obviously be a
427. 16:52major issue. I don't think that's going
428. 16:54to happen, but again, that would be a
429. 16:56concern. So that's the fan of potential
430. 16:58outcomes. From here we're up 15%.
431. 17:01I think we could go higher uh you know
432. 17:03through the balance of the year. Uh or
433. 17:05we could go lower. I'll leave you with
434. 17:07this.
435. 17:08The period of 2023 to 2025, you only had
436. 17:12to get one thing right. That AI was
437. 17:15going to be the biggest super cycle in
438. 17:16the history of technology. And you
439. 17:18needed to shove your chips into the AI
440. 17:19trade. That's it. Right? If you were in
441. 17:22the AI trade, you made money. That is
442. 17:24not where we are in 2026. Everybody
443. 17:26knows about AI. It's all priced now.
444. 17:29It's about facts and circumstances. Stay
445. 17:32mentally flexible. Follow the facts.
446. 17:35Don't yolo. Okay? Don't go forex like
447. 17:38our friend up north who uh gave all his
448. 17:40money to Citadel, right? Um Forex
449. 17:44levered in this market, you know, very
450. 17:47dangerous. So, in my estimation, we're
451. 17:50medium position. We're mentally
452. 17:52flexible. If we see those revenues come
453. 17:54in big for these next few months and we
454. 17:57see oil prices retreat, we're going to
455. 17:59put more chips on the table. If not,
456. 18:02we'll, you know, we'll reserve the right
457. 18:03to go even smaller. With that, thank you
458. 18:05all. Thanks for having me.