엔비디아 저점 매수?, 젠슨 황 없이도 회사가 굴러갈지와 미국 정부가 어디까지 문제 삼을지가 중요합니다 #런치라이브핵심클립 — Transcript

Full transcript

1. 0:00And furthermore, US intelligence
2. 0:02authorities—and this part is even
3. 0:04scarier—reportedly knew about it five or
4. 0:07six months before it was announced. How
5. 0:09did they know? Well, they suspected this
6. 0:11person was a spy. So, they were
7. 0:13wiretapping this person, and it turned
8. 0:15out the person they kept meeting was
9. 0:17Jensen Huang.
10. 0:25Tencent’s market cap evaporated by $ 400
11. 0:27billion, with talks of the CEO stepping
12. 0:28down as the board launched an
13. 0:29independent probe into Chinese spy
14. 0:30allegations. The US Congress is also
15. 0:32pressuring for the truth to be uncovered
16. 0:33.
17. 0:34You all heard that, right? So, Jensen
18. 0:36Huang has a wife, and they were likely a
19. 0:39campus couple.
20. 0:41They were dating long before he founded
21. 0:43the company, and since his daughter
22. 0:46seems to be in her 30s or 40s, he must
23. 0:48be nearly in his 60s or 70s.
24. 0:52They’ve been together for almost 50
25. 0:53years, basically half a century, but
26. 0:55apparently, he committed adultery.
27. 0:59I don’t know why he would want to do
28. 1:00that. If I had that much money, would I
29. 1:03want to do that?
30. 1:05I’m not sure, but anyway, he committed
31. 1:07adultery, and the problem is it wasn't
32. 1:09just the affair; there were rumors that
33. 1:12the person involved was a Chinese spy,
34. 1:14and furthermore, US intelligence—this is
35. 1:16the scarier part—knew before it was
36. 1:18announced.
37. 1:24They suspected she was a spy.
38. 1:26So they were wiretapping her, and it
39. 1:28turned out the person she kept meeting
40. 1:30was Jensen Huang, and it was an affair.
41. 1:34The sharpness of those intelligence
42. 1:35authorities is quite frightening.
43. 1:38Anyway, because of this incident, the
44. 1:40stock price dropped significantly.
45. 1:42But it’s not like it fell solely because
46. 1:44of this.
47. 1:47There’s the debate over a "cliff" for
48. 1:49semiconductor and AI hardware companies
49. 1:52, or the "contraction" theory that the
50. 1:54peak has passed, which has soured
51. 1:57investment sentiment, and on top of that
52. 1:59, the Democrats won big in the midterms
53. 2:02and are blocking data center
54. 2:04construction.
55. 2:07Naturally, fewer GPUs will be sold, and
56. 2:09on top of that, rumors are circulating
57. 2:12about this affair, asking if they’re
58. 2:14getting divorced or if there will be a
59. 2:16division of assets.
60. 2:19For reference, it’s understandable why
61. 2:21the Democrats are blocking data center
62. 2:23construction.
63. 2:25It’s—well, this is suddenly a political
64. 2:28topic, but it’s related to the economy,
65. 2:30so let me quickly touch upon it: the
66. 2:32Democrats need to attack the Republicans
67. 2:34, right?
68. 2:36They need to regain power and boost
69. 2:38their approval ratings, after all.
70. 2:40But if you had to pick just one thing
71. 2:42that the Republican Trump did wrong, it
72. 2:44would be inflation—prices.
73. 2:46As for everything else—the stance on
74. 2:48Muslims, "America First," tariffs, war—
75. 2:51there are many factors, but voters might
76. 2:53say, "I don't know, I was neutral."
77. 2:56"But prices went up, and it’s so
78. 2:57annoying," is what they’re saying. So,
79. 2:59the Democrats have to attack inflation.
80. 3:01Attacking inflation to lower prices is
81. 3:04the Democrats 'path to victory.
82. 3:06There are two ways to lower inflation;
83. 3:09one is to stop the wars, right?
84. 3:13Only by stopping the wars can things
85. 3:14like oil prices come down. Gasoline
86. 3:16prices.
87. 3:17And then, we must eliminate tariffs.
88. 3:18Eliminating tariffs will lower import
89. 3:20prices.
90. 3:21But there is one more thing. Among the
91. 3:23various things they are suffering from
92. 3:26right now, when it comes to inflation,
93. 3:28we often think, "Well, it's just a few
94. 3:30extra dollars for electricity and water
95. 3:33, so what?" But the general consensus is
96. 3:35that in areas where data centers are
97. 3:37built, those costs have risen
98. 3:38several-fold, though it remains a
99. 3:40subject of debate. So, people are saying
100. 3:42, "Prices are up, and now electricity
101. 3:43and water are hard to afford. This is
102. 3:44all because of the data center."
103. 3:46And so, "Ban the data centers" is
104. 3:48currently the Democratic Party's stance
105. 3:51. Because of that, companies like Nvidia
106. 3:54, TSMC, Samsung Electronics, and Hynix
107. 3:56are bound to suffer. Now, let's take a
108. 3:58look at the article in detail.
109. 4:01As the scandal involving Nvidia CEO
110. 4:02Jensen Huang expands into allegations of
111. 4:04Chinese espionage, Nvidia, the world's
112. 4:06largest AI semiconductor company, faces
113. 4:07a management crisis. With the Nvidia
114. 4:09board of directors launching an
115. 4:10independent investigation, there is even
116. 4:11talk of CEO Huang potentially stepping
117. 4:13down. As the U.S. Congress also begins
118. 4:14an investigation, some speculate that
119. 4:15the influence of CEO Huang, who has
120. 4:16maintained close ties with the Donald
121. 4:18Trump administration, may be waning.
122. 4:21Ah.
123. 4:23Moving on to the next section, it goes
124. 4:24like this.
125. 4:27First off, Nvidia's Jensen did admit to
126. 4:29it.
127. 4:30"I had an affair.""I'm sorry.""But it
128. 4:32won't affect the company." That's how he
129. 4:35addressed it for now.
130. 4:37The person he had the affair with is
131. 4:38Alice Huang. Well, there are a lot of
132. 4:39Huangs. This Alice Huang is reportedly
133. 4:42an executive at a company called
134. 4:45MegaSpeed.
135. 4:47But this MegaSpeed company buys Nvidia
136. 4:50GPUs for distribution, and the suspicion
137. 4:52is, did they bypass normal trade
138. 4:55practices and just funnel all those GPUs
139. 4:57to China?
140. 4:59Current Nvidia GPUs, specifically the
141. 5:01latest models, are not allowed to be
142. 5:03sold to China. It's illegal, so the
143. 5:05question is, did they find a way to
144. 5:06circumvent that?
145. 5:08That is the suspicion being raised.
146. 5:12What is the problem here? The problem as
147. 5:14I see it.
148. 5:15I don't think the issue is suddenly not
149. 5:18being able to build data centers, or
150. 5:21asset division, or stake ratios; I think
151. 5:23the real problem is CEO Huang's
152. 5:26management capabilities, which I will
153. 5:28discuss next. Now, let's look at that
154. 5:30part. Assessments suggest that CEO
155. 5:32Huang's position has also become
156. 5:33uncertain. Bloomberg analyzed that while
157. 5:34the possibility of CEO Huang stepping
158. 5:35down was once unimaginable, it has now
159. 5:37emerged as a risk factor that investors
160. 5:38must consider. The problem is that
161. 5:40Nvidia is overly dependent on CEO
162. 5:41Huang's leadership. Nvidia has a unique
163. 5:44management structure where nearly 60
164. 5:45executives report directly to CEO Huang
165. 5:47.
166. 5:48Since co-founding Nvidia in 1993, CEO
167. 5:50Huang has led the company for 33 years,
168. 5:52and no official succession plan has been
169. 5:55disclosed.
170. 5:56It’s a very famous philosophy that
171. 5:59Nvidia’s Jensen receives direct reports
172. 6:03from 60 executives.
173. 6:05Usually, large companies prioritize
174. 6:09bureaucracy and CEO efficiency, but
175. 6:13Jensen, despite his age, remains
176. 6:16incredibly energetic, and this fast
177. 6:20decision-making structure works because
178. 6:24Nvidia isn't actually a hyper-massive
179. 6:28corporation.
180. 6:30They just have a high market cap because
181. 6:32they make a lot of money, and he has
182. 6:34often stated that this structure fits
183. 6:36their management environment.
184. 6:39It’s a management style that gets a lot
185. 6:41of attention in academia, and it implies
186. 6:44that those 60 people are all on equal
187. 6:46footing.
188. 6:49It’s just that no one with exceptional
189. 6:51leadership ability has been verified yet
190. 6:52. Jensen has to conduct those 60 people
191. 6:56like an orchestra, and while the
192. 7:00individual components haven’t changed,
193. 7:04you never know what kind of decisions
194. 7:07might be made.
195. 7:10That’s the problem that arises.
196. 7:14Recently, while studying and investing
197. 7:17in Big Tech and AI, I’ve come to feel
198. 7:20that strong leadership or, if direct
199. 7:23involvement isn't possible, a perfect
200. 7:26delegation process is extremely
201. 7:28important.
202. 7:30For example, it’s like this.
203. 7:33Before ChatGPT was released to the
204. 7:35public in November 2023, Google’s
205. 7:38internal LLM model was supposedly much
206. 7:41better than ChatGPT at the time.
207. 7:45The department in charge said, "Let's
208. 7:48launch this, it's really good," but the
209. 7:50reason they couldn't was that in
210. 7:52Google’s decision-making structure, the
211. 7:54advertising division sits at the very
212. 7:56top.
213. 7:59Because advertising makes so much money
214. 8:01, no one could challenge the power of
215. 8:03that division.
216. 8:05And there was a concern—which still
217. 8:07exists—that if LLMs grew too large, it
218. 8:09might hurt their ad revenue.
219. 8:12Those concerns were filtered all the way
220. 8:14up through their bureaucratic
221. 8:15decision-making structure, and the
222. 8:17project was dropped.
223. 8:19As a result, Google initially fell
224. 8:22behind ChatGPT in AI, although they have
225. 8:26caught up a lot now.
226. 8:30Also, as I often mention, people like
227. 8:33Tesla’s Elon or Meta’s Mark are still
228. 8:36young.
229. 8:39And because they are the founders,
230. 8:41largest shareholders, and CEOs making
231. 8:44decisions directly, I believe they can
232. 8:47still make the fast, disruptive
233. 8:49innovations that are possible in smaller
234. 8:52companies despite their current scale.
235. 8:56I think that part is really crucial
236. 8:58these days. In the past, when Warren
237. 9:01Buffett talked about looking at
238. 9:02management, people would say, "Why look
239. 9:04at the managers?""You're the chairman so
240. 9:06you look, or you're a manager so you
241. 9:07look; I don't care about that." I used
242. 9:09to think I should just look at the
243. 9:10financial statements, but my thinking
244. 9:11has changed a lot lately.
245. 9:14When looking at an executive, the key is
246. 9:16whether they can make decisions—which
247. 9:18can be seen as dogmatic in a bad way, or
248. 9:21very swift and accountable in a good way
249. 9:23—and that was Nvidia’s greatest strength
250. 9:26.
251. 9:27That strength was Jensen Huang.
252. 9:30But now, we don't know what will happen
253. 9:33moving forward.
254. 9:37Even if it's true that CEO Huang managed
255. 9:40the company well, regardless of whether
256. 9:42he committed adultery or acted as a spy
257. 9:45, if a decision is made that someone who
258. 9:48threatens national secrets or security
259. 9:50shouldn't be running the company, it
260. 9:53might be good for the greater cause, but
261. 9:56it could be very bad for Nvidia
262. 9:58shareholders.
263. 10:00Since those are factors we cannot grasp
264. 10:03at the moment, it is understandable that
265. 10:05the stock price is falling.
266. 10:09Looking at additional reports, Jensen
267. 10:11has been arguing that the U.S. should
268. 10:14supply advanced semiconductors to
269. 10:17Chinese companies.
270. 10:19He asked for an easing of export
271. 10:21restrictions. He argued that we must
272. 10:23supply them so that by flooding China
273. 10:26with our AI chips early, we can crush
274. 10:28their own desire to research, develop,
275. 10:30or launch their own AI semiconductors.
276. 10:33He said that would be better for the
277. 10:34U.S. in the long run. He said all that,
278. 10:36but now they are saying, "Wait, you're
279. 10:37actually Taiwanese."
280. 10:40"And you've actually committed adultery
281. 10:42.""With a Chinese person.""You're
282. 10:43passing secrets to a spy." It turns into
283. 10:45a situation like that. Then the problem
284. 10:47becomes: what if Jensen's words don't
285. 10:49carry any weight anymore?
286. 10:51This is, in fact,
287. 10:55the factor behind the recent significant
288. 10:57drop in Nvidia's stock price, and it's a
289. 11:00difficult situation.
290. 11:02I actually like contrarian investing—
291. 11:05buying when stock prices have fallen to
292. 11:07make a profit—but at the core of that,
293. 11:10there must be a belief that it is a good
294. 11:13company with an excellent leader.
295. 11:15So, Nvidia is obviously a good company.
296. 11:17They have a global monopoly on data
297. 11:20center GPUs, but if it had been a
298. 11:22different issue—for example, a tariff
299. 11:24issue, or if it had just ended with an
300. 11:27affair—or a divorce, I think I would
301. 11:29have invested.
302. 11:31But the current problem is, how far will
303. 11:34the U.S. government take this?
304. 11:38If they decide that something is more
305. 11:40important than Nvidia's management, it
306. 11:43would be a very unfortunate development
307. 11:45for Nvidia shareholders.
308. 11:49They say his daughter, Madison Huang, is
309. 11:52the "number two," but on the surface,
310. 11:55she seems to only play the role of an
311. 11:58assistant, so it’s hard to tell if she
312. 12:01could actually coordinate things or if
313. 12:04the company would run well without
314. 12:07Jensen.