1
00:00:00,080 --> 00:00:13,480
Speaker 1: Bloomberg Audio Studios, Podcasts, radio news. Bloomberg Tech is alive

2
00:00:13,560 --> 00:00:17,360
Speaker 1: from coast to coast with Caroline Hyde and New York

3
00:00:17,680 --> 00:00:20,800
Speaker 1: and Vla Low in San Francisco.

4
00:00:22,880 --> 00:00:26,000
Speaker 2: This is Bloomberg Tech coming up in video. Earnings on deck,

5
00:00:26,079 --> 00:00:29,000
Speaker 2: a huge moment to the world's most valuable company in

6
00:00:29,080 --> 00:00:32,320
Speaker 2: the stock at the center of the AI Boom plus

7
00:00:32,360 --> 00:00:34,680
Speaker 2: SpaceX inches closer to launch.

8
00:00:34,720 --> 00:00:35,239
Speaker 3: It could be.

9
00:00:35,240 --> 00:00:39,239
Speaker 2: Today that must company unveils it's IPO filing and we

10
00:00:39,280 --> 00:00:42,400
Speaker 2: take a deep dive into soft banks bet on open

11
00:00:42,440 --> 00:00:46,879
Speaker 2: AI insiders worried the Masoschi sun is too tight to

12
00:00:46,920 --> 00:00:50,479
Speaker 2: the AI giant. Right now in video is everything that

13
00:00:50,479 --> 00:00:53,880
Speaker 2: the Bloomberg Tech audience is clicking on reading about, talking

14
00:00:53,880 --> 00:00:55,880
Speaker 2: about We're up two percent, taking our year to D

15
00:00:55,880 --> 00:00:59,640
Speaker 2: eight game twenty percent. We expect eighty percent top line growth,

16
00:01:00,040 --> 00:01:02,880
Speaker 2: eighty five percent growth in earnings per share. But there

17
00:01:02,880 --> 00:01:06,480
Speaker 2: are questions about China and how much momentum this AI

18
00:01:06,600 --> 00:01:07,440
Speaker 2: story has.

19
00:01:07,760 --> 00:01:09,720
Speaker 3: Let's get the stock story for the world's.

20
00:01:09,440 --> 00:01:14,640
Speaker 2: Most valuable company with Bloomberg Tech Equasies Reporter Carmen Ryanikey Friday,

21
00:01:14,840 --> 00:01:19,479
Speaker 2: Monday Tuesday, we were almost in correction territory within video

22
00:01:19,520 --> 00:01:21,199
Speaker 2: at the heart of what was happening in chip stocks.

23
00:01:21,240 --> 00:01:23,959
Speaker 2: You write today that in Videa's earnings will either make

24
00:01:24,360 --> 00:01:29,080
Speaker 2: or will break the rally and chips, why, Yeah, So.

25
00:01:29,160 --> 00:01:32,080
Speaker 4: It's been so important in video overall. Right, we know

26
00:01:32,120 --> 00:01:34,520
Speaker 4: it's sort of the king maker of the AI trade.

27
00:01:34,640 --> 00:01:37,000
Speaker 4: It's the stock that all investors are watching really to

28
00:01:37,040 --> 00:01:39,679
Speaker 4: see the demand going forward and if that story is

29
00:01:39,720 --> 00:01:42,680
Speaker 4: still strong. And so this huge rally that we've seen

30
00:01:42,760 --> 00:01:45,280
Speaker 4: in the chip making space really from this March thirtieth

31
00:01:45,360 --> 00:01:49,240
Speaker 4: low up through last week, it's had incredible momentum. But

32
00:01:49,320 --> 00:01:51,640
Speaker 4: in video is really the thing that investors are going

33
00:01:51,680 --> 00:01:53,880
Speaker 4: to watch to see if this can go forward. We

34
00:01:54,040 --> 00:01:56,240
Speaker 4: got a little bit of a rest fit over the

35
00:01:56,280 --> 00:01:58,160
Speaker 4: last few days, we saw a sell off, but we're

36
00:01:58,240 --> 00:02:00,840
Speaker 4: kind of back up that rally is going stronger today.

37
00:02:00,960 --> 00:02:04,040
Speaker 4: So how Nvidia talks about the future demand here is

38
00:02:04,080 --> 00:02:06,200
Speaker 4: going to be paramount to that entire sector.

39
00:02:07,040 --> 00:02:09,120
Speaker 2: Calm and the m Live and macro View team go

40
00:02:09,240 --> 00:02:13,840
Speaker 2: back and say, what typically happens post earnings for Nvidia,

41
00:02:13,880 --> 00:02:17,040
Speaker 2: even if it is a blowout, and this is the answer.

42
00:02:17,520 --> 00:02:20,560
Speaker 2: It's a pretty mixed picture, right. There is no guarantee

43
00:02:20,840 --> 00:02:23,960
Speaker 2: that even if Nvidia has a bond storming print that

44
00:02:24,080 --> 00:02:26,560
Speaker 2: in the days that follow, even the month that follows,

45
00:02:27,280 --> 00:02:29,920
Speaker 2: the market will go with it exactly.

46
00:02:29,960 --> 00:02:32,400
Speaker 4: So, something that we've seen over the last few earnings

47
00:02:32,440 --> 00:02:35,119
Speaker 4: reports for in video is that shares at actually fall

48
00:02:35,200 --> 00:02:37,800
Speaker 4: in the day after results, even if they're very strong,

49
00:02:37,840 --> 00:02:40,000
Speaker 4: even if it's a blowout report. Some of that is

50
00:02:40,040 --> 00:02:42,720
Speaker 4: pretty typical around earnings. You get a little bit of

51
00:02:42,760 --> 00:02:45,880
Speaker 4: a buy the rumor sell the news. Obviously, this could

52
00:02:45,919 --> 00:02:48,280
Speaker 4: be a good place for investors to trim to sort

53
00:02:48,320 --> 00:02:51,520
Speaker 4: of take some profits, especially after such an incredible rally.

54
00:02:51,880 --> 00:02:53,560
Speaker 4: And the other thing I'd say is that so in

55
00:02:53,680 --> 00:02:56,040
Speaker 4: Vidia is still the most important stock in the S

56
00:02:56,080 --> 00:02:58,160
Speaker 4: and P five hundred, it's the largest company in the world,

57
00:02:58,160 --> 00:03:00,680
Speaker 4: it commands the most weight in the index, but it's

58
00:03:00,720 --> 00:03:03,960
Speaker 4: not immune to broader macro pressures. And we have seen

59
00:03:04,320 --> 00:03:08,919
Speaker 4: sort of the pull and of the AI trade continuing

60
00:03:09,000 --> 00:03:12,280
Speaker 4: over this year. So while it's very important to see

61
00:03:12,320 --> 00:03:15,640
Speaker 4: the future demand there for these stocks, it's also no

62
00:03:15,760 --> 00:03:18,680
Speaker 4: guarantee that this is going to continue to march forward

63
00:03:18,720 --> 00:03:21,040
Speaker 4: and up to the right, especially just in the days

64
00:03:21,080 --> 00:03:21,799
Speaker 4: after the report.

65
00:03:22,000 --> 00:03:25,280
Speaker 2: Hey common how often before the show and before big

66
00:03:25,280 --> 00:03:27,880
Speaker 2: moments are Ibu, and I say, calm and check me

67
00:03:27,919 --> 00:03:28,240
Speaker 2: on this.

68
00:03:28,840 --> 00:03:29,680
Speaker 3: Let's do it again.

69
00:03:30,160 --> 00:03:33,400
Speaker 2: In Vidia's twenty four times forward earnings twelve month forward

70
00:03:33,400 --> 00:03:36,440
Speaker 2: earnings that seems below.

71
00:03:36,520 --> 00:03:37,720
Speaker 3: It's historical average.

72
00:03:37,880 --> 00:03:40,560
Speaker 2: I'm basically saying, tell me about the valuation of this

73
00:03:40,640 --> 00:03:41,880
Speaker 2: company going into this print.

74
00:03:42,400 --> 00:03:46,320
Speaker 4: Yeah, So we've seen in Vidia's valuation sort of continuously

75
00:03:46,480 --> 00:03:49,920
Speaker 4: tick down even through the stocks rally, because its earnings

76
00:03:49,960 --> 00:03:54,440
Speaker 4: growth is so astronomical. So basically, what investors are paying

77
00:03:54,640 --> 00:03:57,080
Speaker 4: for that forward earnings growth is less than a lot

78
00:03:57,120 --> 00:03:59,760
Speaker 4: of other companies, even other companies in the space that

79
00:03:59,800 --> 00:04:01,440
Speaker 4: come manda much higher valuation.

80
00:04:01,720 --> 00:04:03,960
Speaker 5: So it's interesting that it's so cheap.

81
00:04:04,120 --> 00:04:07,560
Speaker 4: Usually that entices investors to jump back into the stock,

82
00:04:07,920 --> 00:04:11,119
Speaker 4: but that remains to be seen. We'll see what happens

83
00:04:11,160 --> 00:04:12,440
Speaker 4: in the after hours today.

84
00:04:13,360 --> 00:04:15,800
Speaker 2: So smart and given the impact at the index level,

85
00:04:15,880 --> 00:04:17,640
Speaker 2: that's why I wanted to start on the stock story.

86
00:04:17,640 --> 00:04:21,360
Speaker 2: Bloomberg's Carmen Ryinike, thank you very much. That's the broader picture.

87
00:04:21,480 --> 00:04:24,839
Speaker 2: Let's dive into Nvidia's specific concerns. We've got to Bloomberg's

88
00:04:24,839 --> 00:04:28,279
Speaker 2: Maggie Eastland, and this is about a story of AI momentum,

89
00:04:28,440 --> 00:04:32,039
Speaker 2: the infrastructure build out, the basics, What does the world

90
00:04:32,080 --> 00:04:34,920
Speaker 2: think Nvidia is going to say after the market closed.

91
00:04:34,680 --> 00:04:39,800
Speaker 6: Today, Well, look across the board, folks are expecting a

92
00:04:39,920 --> 00:04:43,640
Speaker 6: really strong performance from Nvidia, But like we discussed earlier,

93
00:04:43,680 --> 00:04:46,800
Speaker 6: that might not necessarily translate to stock market gains because

94
00:04:46,839 --> 00:04:48,480
Speaker 6: at this point they're expected to do well.

95
00:04:48,640 --> 00:04:49,840
Speaker 3: AI demand is strong.

96
00:04:50,120 --> 00:04:52,720
Speaker 6: Jensen Wong has said that repeatedly. The thing that they're

97
00:04:52,720 --> 00:04:56,200
Speaker 6: going to be looking through is trying to determine how

98
00:04:56,279 --> 00:05:00,000
Speaker 6: long can this AI boom really last? Is this sustainable

99
00:05:00,160 --> 00:05:03,000
Speaker 6: for demand to be this strong for this long? And

100
00:05:03,040 --> 00:05:05,760
Speaker 6: then more specifically, what are some of the supply constraints.

101
00:05:06,000 --> 00:05:07,880
Speaker 6: Obviously demand is skyrocketing.

102
00:05:08,080 --> 00:05:09,080
Speaker 5: There are things like.

103
00:05:09,279 --> 00:05:12,800
Speaker 6: Memory, like networking that in video could run into bottlenecks

104
00:05:12,960 --> 00:05:16,080
Speaker 6: that ultimately limit how much of their AI compute they're

105
00:05:16,080 --> 00:05:18,480
Speaker 6: able to supply to hyperscalers and other customers.

106
00:05:20,160 --> 00:05:24,120
Speaker 2: Maggie, we're in the unusual situation, aren't we, Where it's Wednesday,

107
00:05:24,240 --> 00:05:26,920
Speaker 2: we have earnings. I just got back from Las Vegas.

108
00:05:26,920 --> 00:05:27,359
Speaker 3: We spoke to.

109
00:05:27,400 --> 00:05:29,880
Speaker 2: Jensen Wong on Monday, and to your point on the

110
00:05:29,920 --> 00:05:32,520
Speaker 2: supply demand dynamic, this is what he had to say.

111
00:05:33,320 --> 00:05:35,159
Speaker 7: We have the largest supply chain in the world.

112
00:05:35,760 --> 00:05:39,920
Speaker 8: Our partners have done a great job securing supply for us,

113
00:05:40,279 --> 00:05:43,080
Speaker 8: and so all of the pieces go together. The silicon,

114
00:05:43,160 --> 00:05:46,440
Speaker 8: photonics is lined up, Everything is all lined up. It's

115
00:05:46,480 --> 00:05:49,839
Speaker 8: just that the demand is much greater than the overall

116
00:05:49,880 --> 00:05:51,000
Speaker 8: capacity of the world.

117
00:05:52,520 --> 00:05:55,039
Speaker 2: Then there's the overhang, and I think you and I

118
00:05:55,080 --> 00:05:57,640
Speaker 2: both know the overhang on in Video is probably China,

119
00:05:57,839 --> 00:06:01,159
Speaker 2: even if it's not material in some sense, where do

120
00:06:01,240 --> 00:06:02,200
Speaker 2: we stand on China?

121
00:06:04,440 --> 00:06:08,400
Speaker 6: Look, the outlook is still very cloudy. So back in March,

122
00:06:08,920 --> 00:06:12,520
Speaker 6: Jensen Wong told investors that he was firing up H

123
00:06:12,560 --> 00:06:15,520
Speaker 6: two hundred production and he had US permissions. Now the

124
00:06:15,600 --> 00:06:18,520
Speaker 6: US has confirmed that. But what we're seeing, even after

125
00:06:18,640 --> 00:06:21,760
Speaker 6: Jensen joined for the trip to China last week, is

126
00:06:21,800 --> 00:06:24,680
Speaker 6: that US officials are saying China is blocking its companies

127
00:06:24,720 --> 00:06:27,320
Speaker 6: from purchasing H two hundreds. So that's a question that

128
00:06:27,360 --> 00:06:29,880
Speaker 6: investors are going to be asking, and in Video is

129
00:06:30,080 --> 00:06:33,520
Speaker 6: likely to address some of that cloudiness during its earnings.

130
00:06:33,120 --> 00:06:33,920
Speaker 5: Call after hours.

131
00:06:34,960 --> 00:06:37,680
Speaker 2: It's been really interesting sitting next to you at some

132
00:06:37,720 --> 00:06:39,560
Speaker 2: of the events of the last twelve months. Right we

133
00:06:39,600 --> 00:06:43,080
Speaker 2: started with CS in January, then GtC. A lot of

134
00:06:43,120 --> 00:06:45,760
Speaker 2: this is about Jensen one trying to convince the world

135
00:06:46,240 --> 00:06:49,520
Speaker 2: that AI is doing things that are meaningful and useful.

136
00:06:49,960 --> 00:06:51,719
Speaker 2: How much do you expect him to kind of linger

137
00:06:51,760 --> 00:06:53,560
Speaker 2: on that on the call, not necessarily talk about the

138
00:06:53,600 --> 00:06:56,840
Speaker 2: data centers getting built within video GPUs, but he tends

139
00:06:56,880 --> 00:06:59,080
Speaker 2: to offer up, hey, this is what I'm doing with

140
00:06:59,120 --> 00:07:01,440
Speaker 2: this particular piece of software, or this is what I

141
00:07:01,440 --> 00:07:02,480
Speaker 2: see in the enterprise.

142
00:07:04,600 --> 00:07:08,159
Speaker 6: Absolutely, I'm fully expecting Judson to emphasize all of the

143
00:07:08,240 --> 00:07:11,680
Speaker 6: applications for AI. Right That's the question that hyperskillers are

144
00:07:11,720 --> 00:07:13,640
Speaker 6: also sort of facing, is what is the end use

145
00:07:13,680 --> 00:07:16,000
Speaker 6: of this? Where is the return on investment going to

146
00:07:16,040 --> 00:07:18,520
Speaker 6: finally arrive? So we're definitely going to see and focus

147
00:07:18,560 --> 00:07:21,040
Speaker 6: on that. You also may see a focus on imprinting

148
00:07:21,400 --> 00:07:24,520
Speaker 6: with that Grock acquisition. You know, there is this story

149
00:07:24,560 --> 00:07:28,280
Speaker 6: in AI land that essentially compute demand is moving from

150
00:07:28,320 --> 00:07:31,200
Speaker 6: training to imprincing, and the question for nvideo is are

151
00:07:31,200 --> 00:07:33,679
Speaker 6: they going to be able to maintain their competitive mode

152
00:07:33,920 --> 00:07:35,520
Speaker 6: as the industry makes that shift.

153
00:07:36,440 --> 00:07:40,440
Speaker 2: Kloomberg's Maggie Eastland top Top Reporting. Thank you very much.

154
00:07:40,760 --> 00:07:42,800
Speaker 2: There's another big one out there today and come in up.

155
00:07:43,280 --> 00:07:46,880
Speaker 2: SpaceX moves closer to what could be the biggest IPO

156
00:07:46,960 --> 00:07:49,680
Speaker 2: in market history. We're waiting on a pretty key document

157
00:07:49,920 --> 00:07:52,280
Speaker 2: and we're going to break down what to expect. It

158
00:07:52,320 --> 00:07:56,000
Speaker 2: is primed for list off. This is what financial markets

159
00:07:56,040 --> 00:07:58,600
Speaker 2: look like right now. Then a's that one hundred rebounding

160
00:07:58,680 --> 00:08:03,320
Speaker 2: up more than a percentage point. Semiconductors rebounding again. Friday, Monday, Tuesday,

161
00:08:03,320 --> 00:08:05,800
Speaker 2: we were headed for correction territory. We're up almost four

162
00:08:05,840 --> 00:08:07,760
Speaker 2: percent on the socks in videos at the heart of

163
00:08:07,760 --> 00:08:10,360
Speaker 2: the story, and we're going to keep that story throughout

164
00:08:10,400 --> 00:08:10,800
Speaker 2: the show.

165
00:08:11,040 --> 00:08:12,720
Speaker 3: We'll be right back. This is Bloomberg Tech.

166
00:08:27,440 --> 00:08:32,600
Speaker 2: Investors are watching waiting for SPACEXIPO. The s one could

167
00:08:32,640 --> 00:08:35,200
Speaker 2: flip public as soon as today. That's the reporting, and

168
00:08:35,240 --> 00:08:39,080
Speaker 2: the company reportedly targeting a seventy five billion dollar raise

169
00:08:39,320 --> 00:08:42,760
Speaker 2: in a listing that could value SpaceX up to more

170
00:08:42,840 --> 00:08:45,360
Speaker 2: than two trillion. We also learn more about the banks.

171
00:08:45,440 --> 00:08:49,520
Speaker 2: Let's talk to Bloomberg's Anthony Hughes. It looks like Goldman

172
00:08:49,640 --> 00:08:54,240
Speaker 2: is flush lead left. Surprising because Morgan Stanley is also

173
00:08:54,480 --> 00:08:55,720
Speaker 2: a lead bank.

174
00:08:56,559 --> 00:09:00,480
Speaker 9: Ah Hi ed, Yes, well, I think that's not that's prizing.

175
00:09:01,840 --> 00:09:04,800
Speaker 9: Goldman has got the coveted lead left role here, But

176
00:09:04,840 --> 00:09:07,400
Speaker 9: whether that really means they're leading the IPO. I mean,

177
00:09:07,440 --> 00:09:09,960
Speaker 9: they are leading the IPO, but to what extent they

178
00:09:09,960 --> 00:09:12,720
Speaker 9: are leading it in a joint capacity with Morgan Stanley

179
00:09:12,760 --> 00:09:15,320
Speaker 9: will be sort of you know, a big copic of

180
00:09:15,360 --> 00:09:19,040
Speaker 9: debate in banking circles. But as we you know, we

181
00:09:19,080 --> 00:09:22,280
Speaker 9: are expecting the prospectives to show that Goldman is named

182
00:09:22,320 --> 00:09:26,400
Speaker 9: first and Morgan Stanley second, and then three or four

183
00:09:26,720 --> 00:09:29,120
Speaker 9: of the other big banks will be you know, on

184
00:09:29,160 --> 00:09:32,079
Speaker 9: the top top line. But you know it's a it's

185
00:09:32,120 --> 00:09:35,200
Speaker 9: a it's for Golden Sacks. There's there's the bragging rights

186
00:09:35,200 --> 00:09:37,200
Speaker 9: associated with that. But we know that Morgan Stanley's very

187
00:09:37,200 --> 00:09:40,400
Speaker 9: heavily involved in the IPO as well, and you know,

188
00:09:40,440 --> 00:09:43,320
Speaker 9: it may be that those those names are in alphabetical order.

189
00:09:43,360 --> 00:09:44,280
Speaker 3: It's possible.

190
00:09:44,640 --> 00:09:47,319
Speaker 9: And also you know, we know the Golden Sacks has

191
00:09:47,520 --> 00:09:49,440
Speaker 9: obviously done a lot of work for Tesla in the

192
00:09:49,440 --> 00:09:52,240
Speaker 9: past as well, so you know, we knew that Golden

193
00:09:52,280 --> 00:09:53,760
Speaker 9: Sacks was going to have a pretty big role on

194
00:09:53,800 --> 00:09:55,400
Speaker 9: this as well.

195
00:09:56,080 --> 00:09:58,800
Speaker 2: We've must is anything as straightforward as things being in

196
00:09:58,800 --> 00:10:02,280
Speaker 2: alphabetical order. I mean, Morgan Stanley too, Wright has done

197
00:10:02,280 --> 00:10:04,640
Speaker 2: a lot with XAI, did a lot with Twitter at

198
00:10:04,679 --> 00:10:07,040
Speaker 2: the time that must bought it. The thing about you,

199
00:10:07,080 --> 00:10:09,599
Speaker 2: Anthony is you're a student of history. You've covered a

200
00:10:09,640 --> 00:10:11,560
Speaker 2: lot of these IPOs tech IPOs, Like if you look

201
00:10:11,559 --> 00:10:14,440
Speaker 2: at Ali Barber, there was a flush lead left then,

202
00:10:14,640 --> 00:10:17,480
Speaker 2: but after the fact, when you pour through it, it

203
00:10:17,559 --> 00:10:19,040
Speaker 2: was pretty much equal economics.

204
00:10:19,360 --> 00:10:20,920
Speaker 9: Yeah, if we go back to Ali Barber, which was

205
00:10:20,960 --> 00:10:23,680
Speaker 9: in twenty fourteen, and you know, that was the biggest

206
00:10:23,760 --> 00:10:26,800
Speaker 9: IPO of the biggest us IPO, it is still the

207
00:10:26,800 --> 00:10:29,480
Speaker 9: biggest us IPO of all time. But that IPO actually

208
00:10:29,480 --> 00:10:32,720
Speaker 9: listed Credit Swiss in that lead left spot, and Credit

209
00:10:32,800 --> 00:10:36,600
Speaker 9: Swiss doesn't exist in the form that it did anymore.

210
00:10:36,640 --> 00:10:39,440
Speaker 9: But you know, at that time, I do remember that,

211
00:10:40,320 --> 00:10:41,400
Speaker 9: you know, there was a lot of debate in the

212
00:10:41,400 --> 00:10:44,640
Speaker 9: bank and banking circles about which bank was really, you know,

213
00:10:44,880 --> 00:10:47,320
Speaker 9: really leading the IPO. And I think if you talk

214
00:10:47,320 --> 00:10:49,760
Speaker 9: to people who are involved in that transaction, Credit Swiss

215
00:10:49,760 --> 00:10:53,000
Speaker 9: wasn't necessarily the bank that did the bulk of the work.

216
00:10:53,080 --> 00:10:56,920
Speaker 9: But you know, what we'll actually see. What we'll see

217
00:10:56,920 --> 00:11:00,959
Speaker 9: here is that what we're expecting. I think what is

218
00:11:01,760 --> 00:11:05,480
Speaker 9: rumored is that the top five banks would actually get

219
00:11:05,760 --> 00:11:09,079
Speaker 9: ecoeconomics or similar fees. So you know, there'll be some

220
00:11:09,120 --> 00:11:11,560
Speaker 9: differentiation between the roles that the banks play, but they

221
00:11:11,559 --> 00:11:14,760
Speaker 9: are those top five banks or so we'll likely be

222
00:11:14,760 --> 00:11:16,960
Speaker 9: getting similar fees. And if you go back to Ali

223
00:11:17,000 --> 00:11:19,679
Speaker 9: Barbara in twenty fourteen, the top five or six banks

224
00:11:19,679 --> 00:11:22,079
Speaker 9: did get the same fees, so you know there is

225
00:11:22,160 --> 00:11:23,520
Speaker 9: some parallels with that transaction.

226
00:11:24,640 --> 00:11:27,600
Speaker 2: Again, we are expecting the biggest IPO of all time,

227
00:11:27,840 --> 00:11:29,960
Speaker 2: and those banks are going to get big business from that.

228
00:11:29,960 --> 00:11:33,800
Speaker 2: Bloomberg is Anthony Hughes, thank you very much. Indeed, staying

229
00:11:33,840 --> 00:11:37,880
Speaker 2: with SpaceX. The company is reportedly planning to acquire coding

230
00:11:37,920 --> 00:11:42,359
Speaker 2: startup Cursor just thirty days or within thirty days.

231
00:11:42,240 --> 00:11:43,960
Speaker 3: After its IPO. That's according to sources.

232
00:11:43,960 --> 00:11:46,640
Speaker 2: In April, SpaceX said it reached an agreement giving it

233
00:11:46,679 --> 00:11:50,120
Speaker 2: the right to acquire Cursor for sixty billion dollars later

234
00:11:50,160 --> 00:11:53,560
Speaker 2: this year, alternatively pay a ten billion dollar fee tied

235
00:11:53,720 --> 00:11:57,640
Speaker 2: to the company's partnership. Bloomberg's Rachel Metz has been across this,

236
00:11:58,040 --> 00:12:01,360
Speaker 2: the chaos of it, every twist in turn. I mean,

237
00:12:01,400 --> 00:12:04,000
Speaker 2: it's so interesting, right. We were asking ourselves at the

238
00:12:04,040 --> 00:12:07,240
Speaker 2: time in April, like why is it structured this way?

239
00:12:07,480 --> 00:12:10,680
Speaker 2: And then you break the news last night that actually

240
00:12:11,000 --> 00:12:13,400
Speaker 2: they go public and then there's a race to get

241
00:12:13,400 --> 00:12:14,800
Speaker 2: this transaction done.

242
00:12:16,080 --> 00:12:16,440
Speaker 10: Yeah.

243
00:12:16,520 --> 00:12:20,840
Speaker 11: So this transaction, according to our reporting, is set to

244
00:12:20,920 --> 00:12:27,560
Speaker 11: proceed thirty days after the IPO. So what that doesn't

245
00:12:27,600 --> 00:12:30,280
Speaker 11: mean that it will close necessarily right then. Sometimes there's

246
00:12:31,160 --> 00:12:34,200
Speaker 11: regulatory looking into that happens, and that could take up

247
00:12:34,240 --> 00:12:36,440
Speaker 11: to a few months, but it is set to go

248
00:12:36,520 --> 00:12:40,319
Speaker 11: ahead or expected, I should say, to go ahead around

249
00:12:40,400 --> 00:12:43,560
Speaker 11: that time. And the other thing that we found that

250
00:12:43,640 --> 00:12:47,360
Speaker 11: was interesting is that this ten billion dollar fee we

251
00:12:47,440 --> 00:12:50,200
Speaker 11: knew in the past from past reporting that it was

252
00:12:50,280 --> 00:12:53,800
Speaker 11: essentially a breakup fee if the transaction didn't happen. But

253
00:12:53,880 --> 00:12:56,520
Speaker 11: we also found that it's going to be in all cash,

254
00:12:56,559 --> 00:12:59,079
Speaker 11: it's not going to be in credits for computer or

255
00:12:59,120 --> 00:12:59,840
Speaker 11: something like that.

256
00:13:00,840 --> 00:13:02,360
Speaker 2: There's going to be a section of the blombog tech

257
00:13:02,400 --> 00:13:04,800
Speaker 2: audience with respect that they're going to say, hold on,

258
00:13:05,520 --> 00:13:11,080
Speaker 2: SpaceX is acquiring a coding startup, Cursor, and let's fill

259
00:13:11,120 --> 00:13:14,600
Speaker 2: the gaps for them. SpaceX acquired Xai just before it

260
00:13:14,640 --> 00:13:18,760
Speaker 2: plans to go public. Spacexai is like this bigger entity.

261
00:13:18,840 --> 00:13:20,480
Speaker 2: Where does Cursor fit in.

262
00:13:21,679 --> 00:13:24,360
Speaker 11: Yeah, that's a really good question, and I think part

263
00:13:24,400 --> 00:13:27,439
Speaker 11: of it, as you mentioned, is this combination between SpaceX

264
00:13:27,480 --> 00:13:28,120
Speaker 11: and Xai.

265
00:13:28,760 --> 00:13:31,359
Speaker 5: Xai has a ton of compute.

266
00:13:31,440 --> 00:13:35,000
Speaker 11: Compute is something that Cursor really really needs more of,

267
00:13:35,800 --> 00:13:38,480
Speaker 11: as we saw just the other day, they announced their

268
00:13:38,559 --> 00:13:41,920
Speaker 11: latest model, which is meant to help people with coding,

269
00:13:42,120 --> 00:13:44,240
Speaker 11: as some of its past have done, and it was

270
00:13:44,280 --> 00:13:48,640
Speaker 11: partially trained on Colossus two, which is a new data

271
00:13:48,679 --> 00:13:52,720
Speaker 11: center from Xai. So you're starting to see the companies

272
00:13:52,760 --> 00:13:56,840
Speaker 11: already working together and this sense that they can be

273
00:13:56,880 --> 00:13:57,960
Speaker 11: really helpful to each other.

274
00:14:01,120 --> 00:14:05,960
Speaker 2: What I understand is about people. Something that happened post

275
00:14:06,120 --> 00:14:09,079
Speaker 2: Xai SpaceX integration, a lot of talent left.

276
00:14:09,679 --> 00:14:12,040
Speaker 3: Do you have any sense of how Cursor feels about that? Right?

277
00:14:12,080 --> 00:14:14,120
Speaker 2: There is the wow, we get to work with Elon Musk,

278
00:14:14,160 --> 00:14:16,280
Speaker 2: and there's also the ut, oh we have to work

279
00:14:16,320 --> 00:14:17,319
Speaker 2: for Elon Musk.

280
00:14:18,240 --> 00:14:19,280
Speaker 10: That's a really good question.

281
00:14:19,440 --> 00:14:21,760
Speaker 11: I mean, I think these two companies have been familiar

282
00:14:21,760 --> 00:14:24,040
Speaker 11: with each other for a while that there have been

283
00:14:24,080 --> 00:14:26,800
Speaker 11: a few Cursor people. I believe that I've gone over

284
00:14:26,840 --> 00:14:29,800
Speaker 11: to Xai. So it's going to be interesting to see

285
00:14:29,800 --> 00:14:33,080
Speaker 11: what happens, right. I mean, as we know, as you mentioned,

286
00:14:33,080 --> 00:14:36,600
Speaker 11: a lot of people had left Xai over time, but

287
00:14:36,720 --> 00:14:39,120
Speaker 11: I mean Cursor they feels a really strong team, and

288
00:14:40,280 --> 00:14:41,440
Speaker 11: we'll just have to wait and see.

289
00:14:42,480 --> 00:14:46,160
Speaker 2: Bloomboa's Rachel Matt's top Reporting. Thank you very much. Coming

290
00:14:46,160 --> 00:14:48,479
Speaker 2: out from the program, we're going to speak to AMCA

291
00:14:48,560 --> 00:14:52,320
Speaker 2: CEO Joi Malik after raising a three hundred million dollars

292
00:14:52,360 --> 00:14:58,200
Speaker 2: Series B Unicorn status manufacturing in critical Defense and Aerospace.

293
00:14:58,240 --> 00:15:16,520
Speaker 2: That's next. This is Bloomberg Tech. Let's take a look

294
00:15:16,520 --> 00:15:20,240
Speaker 2: at today's big number. Eight thousand. That's how many employees

295
00:15:20,360 --> 00:15:23,320
Speaker 2: Meta is laying off starting today as part of a

296
00:15:23,360 --> 00:15:28,480
Speaker 2: previously announced restructuring aimed at reducing costs while investing in AI.

297
00:15:28,560 --> 00:15:31,720
Speaker 2: The company began notifying workers around the world this morning.

298
00:15:32,240 --> 00:15:34,800
Speaker 2: Metas engineering and product teams are expected to be the

299
00:15:34,840 --> 00:15:38,480
Speaker 2: most impacted. Meanwhile, remember we told you yesterday about Standard

300
00:15:38,560 --> 00:15:41,520
Speaker 2: Chartered CEO Bill Winter saying his company plans to cut

301
00:15:41,560 --> 00:15:46,280
Speaker 2: staff too, to replace quote lower value human capital with AI.

302
00:15:46,720 --> 00:15:46,880
Speaker 3: Well.

303
00:15:46,920 --> 00:15:49,440
Speaker 2: Those comments received a bit of a backlash online, including

304
00:15:49,680 --> 00:15:53,960
Speaker 2: from former Singapore President Halimi Yakob, calling it disturbing to

305
00:15:54,040 --> 00:15:57,280
Speaker 2: describe workers in such clinical terms. Singapore and Hong Kong

306
00:15:57,520 --> 00:16:00,960
Speaker 2: serve as the primary hubs for Standard CHARTSOBAL operations. That's

307
00:16:01,000 --> 00:16:04,120
Speaker 2: all prompted the CEO to reassure staff in a memo

308
00:16:04,200 --> 00:16:07,600
Speaker 2: seen by Bloomberg, Winters is striking a more empathetic tone

309
00:16:07,640 --> 00:16:13,960
Speaker 2: and emphasizing the bank's commitment to transitioning its workforce really

310
00:16:14,000 --> 00:16:17,440
Speaker 2: well read Today on Bloomberg, AMKA has announced a three

311
00:16:17,520 --> 00:16:21,400
Speaker 2: hundred million Series B, valuing the company the more than

312
00:16:21,440 --> 00:16:24,520
Speaker 2: a billion dollars. The company offers an integrated platform to

313
00:16:24,640 --> 00:16:29,720
Speaker 2: rapidly develop and manufacture critical aerospace and defense of components,

314
00:16:29,800 --> 00:16:33,640
Speaker 2: joining us as AMCA CEO J Malik in the private markets,

315
00:16:33,720 --> 00:16:36,600
Speaker 2: this space is alive and well, right, we've talked so

316
00:16:36,760 --> 00:16:42,240
Speaker 2: much about reindustrializing America. Help me understand what AMKA is.

317
00:16:42,320 --> 00:16:45,640
Speaker 2: You know, it sounds like contract manufacturing for the industries

318
00:16:45,960 --> 00:16:50,040
Speaker 2: defense technology, aerospace that are a forefront of what's happening

319
00:16:50,080 --> 00:16:50,480
Speaker 2: right now.

320
00:16:51,160 --> 00:16:54,680
Speaker 12: Absolutely, So what AMCA is actually doing is we're developing

321
00:16:55,520 --> 00:17:01,000
Speaker 12: and then manufacturing critical components, not just contract manufacturing, for

322
00:17:01,280 --> 00:17:05,560
Speaker 12: critical readiness and production gaps that our customers have. And

323
00:17:05,640 --> 00:17:08,600
Speaker 12: so typically these components there's only a single source for

324
00:17:08,680 --> 00:17:13,040
Speaker 12: in the country, unlike other contract manufacturing categories, and our

325
00:17:13,080 --> 00:17:17,840
Speaker 12: business designs these components, qualifies them rapidly and then manufactures them.

326
00:17:17,840 --> 00:17:19,160
Speaker 10: So it's a little bit different.

327
00:17:19,000 --> 00:17:21,879
Speaker 12: Than your typical contract manufacturing business, which is, you know,

328
00:17:21,960 --> 00:17:27,160
Speaker 12: just a machine shop you're looking to produce at scale.

329
00:17:26,720 --> 00:17:30,600
Speaker 2: Jay case study Why harness right, That's something we talked

330
00:17:30,640 --> 00:17:33,879
Speaker 2: about a lot in this program. Heavy dependence in that

331
00:17:34,040 --> 00:17:38,399
Speaker 2: case study on China. Give me a similar product, a

332
00:17:38,480 --> 00:17:42,280
Speaker 2: similar sort of component that you are solving for where

333
00:17:42,320 --> 00:17:45,600
Speaker 2: there might be a reliance on an external country to

334
00:17:45,920 --> 00:17:46,280
Speaker 2: get it.

335
00:17:47,280 --> 00:17:49,679
Speaker 10: Yeah, you know. I'll give you two examples.

336
00:17:50,119 --> 00:17:52,360
Speaker 12: One example are the sensors that we put onto our

337
00:17:52,480 --> 00:17:54,440
Speaker 12: you know, put on put onto almost all of our industrial

338
00:17:54,440 --> 00:17:57,800
Speaker 12: platforms today. A lot of those sensors are sourced to

339
00:17:57,880 --> 00:18:01,480
Speaker 12: sensing elements are sourced from off or you know, offshore

340
00:18:01,600 --> 00:18:04,000
Speaker 12: companies and nations allied with them.

341
00:18:04,800 --> 00:18:04,960
Speaker 5: You know.

342
00:18:05,119 --> 00:18:05,840
Speaker 10: That's one example.

343
00:18:05,840 --> 00:18:08,800
Speaker 12: Another example would be, you know, capacitors in other passive

344
00:18:08,800 --> 00:18:12,679
Speaker 12: electronic electrical components. These are often components that have been

345
00:18:12,680 --> 00:18:16,400
Speaker 12: offshore you know, to Asian countries over the past three decades.

346
00:18:16,600 --> 00:18:18,320
Speaker 10: In America today has.

347
00:18:18,200 --> 00:18:20,840
Speaker 12: A very very lowd windling supply base for these types

348
00:18:20,840 --> 00:18:23,639
Speaker 12: of entergy components. And so that those are two examples

349
00:18:24,040 --> 00:18:26,679
Speaker 12: of areas where we are designing in the country and

350
00:18:26,680 --> 00:18:27,879
Speaker 12: then manufacturing into the country.

351
00:18:27,880 --> 00:18:30,080
Speaker 10: Those components for the warfighter.

352
00:18:31,400 --> 00:18:33,600
Speaker 2: Jay, just have a few rapid fire questions if I may,

353
00:18:33,840 --> 00:18:35,520
Speaker 2: what years have you found this company?

354
00:18:36,960 --> 00:18:40,280
Speaker 10: Twenty twenty four in November twenty.

355
00:18:40,000 --> 00:18:42,320
Speaker 2: Twenty four in November? And is this one billion dollar

356
00:18:42,400 --> 00:18:44,240
Speaker 2: valuation post money or pre money?

357
00:18:44,800 --> 00:18:46,800
Speaker 10: That is post money?

358
00:18:46,880 --> 00:18:51,560
Speaker 2: So in about eighteen months you've founded, launched, and scaled

359
00:18:51,640 --> 00:18:54,400
Speaker 2: up this company and got to a billion dollar valuation.

360
00:18:55,520 --> 00:18:57,880
Speaker 3: What should I infer from that? The pace at which

361
00:18:57,920 --> 00:18:58,480
Speaker 3: you've done.

362
00:18:58,280 --> 00:19:00,920
Speaker 12: That, you should infer that you know, we're living at

363
00:19:01,280 --> 00:19:05,159
Speaker 12: I would say the largest gap between what America needs

364
00:19:05,480 --> 00:19:06,639
Speaker 12: and what America.

365
00:19:06,320 --> 00:19:08,680
Speaker 10: Is able to produce in generations.

366
00:19:08,720 --> 00:19:11,879
Speaker 12: It's a generational opportunity for folks who are looking to

367
00:19:11,920 --> 00:19:14,160
Speaker 12: build in this country. I think over the next five

368
00:19:14,160 --> 00:19:16,240
Speaker 12: to ten years you will see a lot more manufacturing

369
00:19:16,240 --> 00:19:20,240
Speaker 12: companies in America building for America. And we're just, you know,

370
00:19:20,280 --> 00:19:23,320
Speaker 12: sitting at the forefront of this. I think immense growth

371
00:19:23,400 --> 00:19:25,879
Speaker 12: growth curve that's that's coming and will and it is

372
00:19:25,920 --> 00:19:29,280
Speaker 12: already happening in the nation, both for defense and aerospace

373
00:19:29,320 --> 00:19:33,320
Speaker 12: customers as well as in other categories like AI, infrastructure, energy, robotics,

374
00:19:33,320 --> 00:19:33,720
Speaker 12: et cetera.

375
00:19:34,560 --> 00:19:35,720
Speaker 3: Jay, who are your customers?

376
00:19:37,359 --> 00:19:41,159
Speaker 12: You know, do the sensitive nature of you know, of

377
00:19:41,200 --> 00:19:43,040
Speaker 12: the work that we do. I can you know, disclose

378
00:19:43,280 --> 00:19:45,680
Speaker 12: any specific names, you know, but some of the names

379
00:19:45,680 --> 00:19:47,720
Speaker 12: that that we are we have been able to disclose

380
00:19:47,720 --> 00:19:51,560
Speaker 12: are you know, Boeing, Airbus, Honeywell, Lockheed Martin and then

381
00:19:51,800 --> 00:19:55,560
Speaker 12: you know certain branches of the military. But you know,

382
00:19:55,640 --> 00:19:57,840
Speaker 12: all of our components today that go on to platforms.

383
00:19:57,880 --> 00:20:00,960
Speaker 12: We all know about seven thirty seven, MAS, Triple seven,

384
00:20:01,160 --> 00:20:05,239
Speaker 12: F thirty five's, F fifteen's, F sixteen's, you know our

385
00:20:05,840 --> 00:20:09,320
Speaker 12: tanks and m K one abrams. So there's there's plenty

386
00:20:09,320 --> 00:20:12,520
Speaker 12: of plenty of your different platforms that we're on today

387
00:20:12,960 --> 00:20:14,320
Speaker 12: that the country actively uses.

388
00:20:15,520 --> 00:20:18,800
Speaker 2: You know, it's interesting we've we've probably spent a lot

389
00:20:18,800 --> 00:20:22,000
Speaker 2: of time understanding some of the problems that those customers

390
00:20:22,040 --> 00:20:25,240
Speaker 2: that you've just outlined have, but probably of ancor it's

391
00:20:25,240 --> 00:20:27,680
Speaker 2: better to talk about what your constraints are, where the

392
00:20:27,720 --> 00:20:29,960
Speaker 2: bottlenecks are for you and your own supply chain.

393
00:20:31,680 --> 00:20:34,600
Speaker 12: You know, in our own supply chain, there are certainly,

394
00:20:34,680 --> 00:20:37,320
Speaker 12: you know, big gaps right now at the sort of

395
00:20:37,400 --> 00:20:41,200
Speaker 12: like infrastructure level, like I mentioned, you know, for sensors,

396
00:20:41,240 --> 00:20:44,200
Speaker 12: for example, you know, we need sensing elements these are

397
00:20:44,320 --> 00:20:48,040
Speaker 12: highly precise, you know, elements that go into sensors that

398
00:20:48,400 --> 00:20:50,639
Speaker 12: the country doesn't have a lot of suppliers for. So

399
00:20:50,680 --> 00:20:52,800
Speaker 12: in those cases, you know, we're taking a new approach.

400
00:20:52,840 --> 00:20:56,800
Speaker 12: We're vertically integrating a lot of our you know, our supply,

401
00:20:57,280 --> 00:20:59,840
Speaker 12: designing that in house and then manufacturing it in house

402
00:21:00,160 --> 00:21:02,879
Speaker 12: to solve gas where the domestic supply chain is not

403
00:21:03,000 --> 00:21:04,080
Speaker 12: able to suppliers.

404
00:21:04,119 --> 00:21:05,680
Speaker 10: Complim J.

405
00:21:05,880 --> 00:21:09,080
Speaker 2: Malik of AMKA raising three hundred million dollars one billion

406
00:21:09,119 --> 00:21:12,120
Speaker 2: dollar valuation, Thank you very much. Plenty of other news

407
00:21:12,119 --> 00:21:15,280
Speaker 2: headlines for the show's time for Talking Tech. First up,

408
00:21:15,400 --> 00:21:18,160
Speaker 2: Open ai is planning its first international flag, the chat

409
00:21:18,240 --> 00:21:21,280
Speaker 2: GBT maker, committing two hundred and thirty four million dollars

410
00:21:21,400 --> 00:21:24,080
Speaker 2: to launch a new facility in Singapore, its first so

411
00:21:24,200 --> 00:21:27,960
Speaker 2: called applied AI lab outside the US. Partnering with local authorities.

412
00:21:28,000 --> 00:21:30,800
Speaker 2: To Hub will hire over two hundred engineers to embed

413
00:21:30,840 --> 00:21:34,640
Speaker 2: advanced models straight into healthcare, finance, and public infrastructure. Plus

414
00:21:34,720 --> 00:21:37,800
Speaker 2: Ali Baba's matching in video's breakneck pace its chip units.

415
00:21:37,840 --> 00:21:41,800
Speaker 2: Teahead just unveiled a new AI accelerator packaging one hundred

416
00:21:41,840 --> 00:21:45,840
Speaker 2: and forty four gigabytes of memory for complex autonomous agent tasks.

417
00:21:46,040 --> 00:21:49,080
Speaker 2: Ali Baba plans to publicly list the business to tap

418
00:21:49,080 --> 00:21:55,080
Speaker 2: investor appetite for Chinese silicon alternatives. And Google is redesigning

419
00:21:55,160 --> 00:21:58,560
Speaker 2: its iconic search box and adding new AI coding tools,

420
00:21:58,560 --> 00:22:00,840
Speaker 2: the biggest update to the product in more than a

421
00:22:00,920 --> 00:22:04,040
Speaker 2: quarter of a century. A Google Io event on Tuesday,

422
00:22:04,080 --> 00:22:07,040
Speaker 2: the company also rolled out several new tools for developers

423
00:22:07,200 --> 00:22:12,600
Speaker 2: to write code using AI and to manage agents. Coming up,

424
00:22:13,119 --> 00:22:17,560
Speaker 2: we're counting down to in Video earnings later today after

425
00:22:17,600 --> 00:22:20,200
Speaker 2: the closing bell, and up next we'll discuss its role

426
00:22:20,720 --> 00:22:25,280
Speaker 2: in the wider chip ecosystem with Bailey Giffard's Paulina o'padden.

427
00:22:25,760 --> 00:22:28,320
Speaker 2: This is what markets look like right now, and in

428
00:22:28,440 --> 00:22:31,360
Speaker 2: Vidia is central to that story. And that's that one

429
00:22:31,440 --> 00:22:35,840
Speaker 2: hundred rebounding up one point four percent semiconductors after three

430
00:22:35,920 --> 00:22:40,160
Speaker 2: straight danes of heavy declines, almost in correction, territory rebounding

431
00:22:40,440 --> 00:22:43,280
Speaker 2: up four percent. In Video is now near session highs,

432
00:22:43,520 --> 00:22:47,520
Speaker 2: up more than two percent halftime in the program, Stay

433
00:22:47,560 --> 00:22:59,120
Speaker 2: with us, This is Bloomberg Tech. Welcome back to Bloomberg Tech.

434
00:22:59,160 --> 00:23:03,120
Speaker 2: In Video Earnings on deck, company reporting its earnings results

435
00:23:03,200 --> 00:23:06,800
Speaker 2: after the market close, when near session highs up more

436
00:23:06,840 --> 00:23:10,800
Speaker 2: than two percent, and for many watchers, China is top

437
00:23:10,840 --> 00:23:11,240
Speaker 2: of mind.

438
00:23:13,520 --> 00:23:16,720
Speaker 8: The Chinese government has to decide how much of their

439
00:23:16,800 --> 00:23:19,120
Speaker 8: local market do they want to protect and how much

440
00:23:19,160 --> 00:23:20,639
Speaker 8: of their local markets do they want.

441
00:23:20,480 --> 00:23:23,240
Speaker 7: To expand with a more AI capacity.

442
00:23:23,800 --> 00:23:26,680
Speaker 8: My sense is that the demand in China is so incredible,

443
00:23:27,080 --> 00:23:30,600
Speaker 8: just like it is here. Agentic AI is also making

444
00:23:30,800 --> 00:23:31,920
Speaker 8: enormous progress there.

445
00:23:32,440 --> 00:23:35,920
Speaker 7: My sense is that over time the market will open.

446
00:23:37,760 --> 00:23:41,879
Speaker 2: Two themes, what's happening with China and demand outpacing supply

447
00:23:42,359 --> 00:23:44,800
Speaker 2: around the world. Let's discuss where video sits in the

448
00:23:44,840 --> 00:23:49,000
Speaker 2: global chip ecosystem with Paulina mcpadden, Investment Manager of International

449
00:23:49,280 --> 00:23:52,960
Speaker 2: Concentrated Growth Strategy at Bailey Gifford. I'm so excited to

450
00:23:53,000 --> 00:23:55,360
Speaker 2: talk to you because, yes, in video is in your strategy,

451
00:23:55,880 --> 00:23:58,919
Speaker 2: but so is everything else around the world. And what

452
00:23:59,080 --> 00:24:02,720
Speaker 2: Gensen outlined there on China's so interesting. The base case

453
00:24:02,720 --> 00:24:05,359
Speaker 2: assumption right now for Invidia is zero access to China.

454
00:24:05,520 --> 00:24:06,720
Speaker 3: Maybe we'll get some clarity.

455
00:24:07,720 --> 00:24:09,720
Speaker 2: There's everything else that comes with that, and I just

456
00:24:09,800 --> 00:24:13,040
Speaker 2: wonder how much attention you really pay to in Nvidia's

457
00:24:13,080 --> 00:24:15,520
Speaker 2: ability to service the China market or not.

458
00:24:16,800 --> 00:24:18,359
Speaker 5: I mean, it's a really good question.

459
00:24:18,560 --> 00:24:20,640
Speaker 13: And I think the honest answer is it's not really

460
00:24:20,680 --> 00:24:23,760
Speaker 13: factored into our upside cases for Nvidia for a while

461
00:24:23,880 --> 00:24:26,560
Speaker 13: because the problem there is you're having to try and

462
00:24:26,560 --> 00:24:29,680
Speaker 13: prejudge what politicians might want to do at some point

463
00:24:29,720 --> 00:24:32,359
Speaker 13: in the future. So each two hundreds have been approved

464
00:24:32,400 --> 00:24:34,800
Speaker 13: for export to China since last year, but it's the

465
00:24:34,880 --> 00:24:38,160
Speaker 13: Chinese government's decision not to approve purchases by Chinese companies

466
00:24:38,160 --> 00:24:42,040
Speaker 13: that has stopped that sale going through for for Nvidia. Now,

467
00:24:42,080 --> 00:24:44,520
Speaker 13: I think the really interesting long term question for US

468
00:24:44,600 --> 00:24:47,040
Speaker 13: is actually what that means for the broader supply chain

469
00:24:47,080 --> 00:24:49,439
Speaker 13: in the long term, because is there a case that

470
00:24:49,520 --> 00:24:52,600
Speaker 13: China could build up its own domestic chip supply chain.

471
00:24:53,160 --> 00:24:55,200
Speaker 5: Now long term, maybe that's.

472
00:24:55,000 --> 00:24:57,280
Speaker 13: Possible, maybe it does threaten in the video eventually, but

473
00:24:57,520 --> 00:24:59,520
Speaker 13: two things give me comfort on that or at least

474
00:24:59,560 --> 00:25:02,679
Speaker 13: the next five years. One of those is that the

475
00:25:02,720 --> 00:25:05,960
Speaker 13: existing high power chips in China are still a fraction

476
00:25:06,000 --> 00:25:08,360
Speaker 13: of the performance of Nvidia's leading chips, so I think

477
00:25:08,440 --> 00:25:11,480
Speaker 13: that gives non Chinese companies and AI.

478
00:25:11,359 --> 00:25:13,439
Speaker 5: Researchers a leg up for the long term.

479
00:25:14,160 --> 00:25:16,520
Speaker 13: And the second thing is it's actually really really hard

480
00:25:16,600 --> 00:25:20,760
Speaker 13: to build up a scaled and complex semiconductor supply chain.

481
00:25:20,800 --> 00:25:23,560
Speaker 5: It's not as easy as just putting capital into.

482
00:25:23,320 --> 00:25:25,240
Speaker 13: It, as China has done in the past with things

483
00:25:25,280 --> 00:25:28,520
Speaker 13: like EVS and emmersion into global dominance in the battery

484
00:25:28,560 --> 00:25:31,960
Speaker 13: supply chain, for example. These are extremely complex bits of

485
00:25:32,000 --> 00:25:35,359
Speaker 13: equipment that require cooperation around the world, and that's just

486
00:25:35,400 --> 00:25:37,440
Speaker 13: not something that I think you can do year on

487
00:25:37,600 --> 00:25:40,879
Speaker 13: year if you're only relying on the domestic demand that

488
00:25:40,920 --> 00:25:43,840
Speaker 13: you get from China and you're not able to sell internationally.

489
00:25:44,880 --> 00:25:47,399
Speaker 2: We just showed s k Heinex is a part of

490
00:25:47,440 --> 00:25:50,200
Speaker 2: your strategy, and what we played to you a moment

491
00:25:50,240 --> 00:25:52,359
Speaker 2: ago was a conversation I had with Jensen one on Monday,

492
00:25:53,200 --> 00:25:56,080
Speaker 2: and it seems memory is still the bottleneck. But the

493
00:25:56,080 --> 00:25:58,640
Speaker 2: way that Jensen outlined it to me is three years

494
00:25:58,640 --> 00:26:01,760
Speaker 2: ago he went to the Memory make is, including Skhinex,

495
00:26:02,160 --> 00:26:03,720
Speaker 2: and he said, this is what I think the world

496
00:26:03,760 --> 00:26:06,399
Speaker 2: will look like in a few years time, and he

497
00:26:06,480 --> 00:26:09,760
Speaker 2: tried to convince them to invest in permanent capacity. You

498
00:26:09,840 --> 00:26:12,639
Speaker 2: know better than I do, Paulina. Memory has been boom

499
00:26:12,640 --> 00:26:17,159
Speaker 2: and bust cyclical historically. In your strategy for SK is

500
00:26:17,200 --> 00:26:20,280
Speaker 2: that different now? Is there permanency to the capacity that's needed.

501
00:26:21,160 --> 00:26:25,000
Speaker 13: So I think bottlenecks is a really interesting framing because

502
00:26:25,280 --> 00:26:28,000
Speaker 13: people tend to think, oh, bottleneck is extremely investable and

503
00:26:28,040 --> 00:26:29,399
Speaker 13: that can lead to a lot of growth and a

504
00:26:29,400 --> 00:26:29,960
Speaker 13: lot of upside.

505
00:26:29,960 --> 00:26:31,359
Speaker 5: And that's true to a certain extent.

506
00:26:31,720 --> 00:26:34,080
Speaker 13: But I think Marcus are actually quite good in some

507
00:26:34,119 --> 00:26:37,280
Speaker 13: cases recognizing bottlenecks well ahead of time and investing in

508
00:26:37,320 --> 00:26:39,760
Speaker 13: them before the fundamentals.

509
00:26:39,040 --> 00:26:39,800
Speaker 5: Even come through.

510
00:26:40,080 --> 00:26:42,280
Speaker 13: So as long term investors, what we're trying to do

511
00:26:42,440 --> 00:26:46,359
Speaker 13: is identify these structural opportunities. Are companies that become even

512
00:26:46,400 --> 00:26:50,080
Speaker 13: better as they scale and deepen their competitive position in

513
00:26:50,119 --> 00:26:53,280
Speaker 13: the ecosystem. So i'd contrast sk Hinex here on the

514
00:26:53,280 --> 00:26:56,400
Speaker 13: one hand, with TSMC because to some extents of portfolio

515
00:26:56,400 --> 00:26:59,760
Speaker 13: construction question. So I think that there is a potential

516
00:26:59,760 --> 00:27:03,280
Speaker 13: that Skiheyinex becomes a substantially better company over the long term.

517
00:27:03,520 --> 00:27:06,120
Speaker 13: The fact that they are able to extract these long

518
00:27:06,200 --> 00:27:09,280
Speaker 13: term purchase agreements from customers now is really a substantial

519
00:27:09,280 --> 00:27:13,720
Speaker 13: difference over history. The industry has consolidated, it's not as

520
00:27:13,720 --> 00:27:15,919
Speaker 13: irrational as it used to be, and I think that

521
00:27:16,000 --> 00:27:17,879
Speaker 13: there is a case to be made that with the

522
00:27:17,960 --> 00:27:21,240
Speaker 13: rise of HBM and in particular custom HBM that creates

523
00:27:21,280 --> 00:27:23,560
Speaker 13: a degree of lock in with customers. That means that

524
00:27:23,600 --> 00:27:25,800
Speaker 13: these are no longer commoditized products as they have.

525
00:27:25,800 --> 00:27:26,520
Speaker 5: Been in the past.

526
00:27:26,960 --> 00:27:30,600
Speaker 13: But that's just a hypothesis, and I think we're going

527
00:27:30,600 --> 00:27:33,120
Speaker 13: to see how that develops over time. And i'd contrast

528
00:27:33,200 --> 00:27:35,639
Speaker 13: that with something like TSMC, which is actually the largest

529
00:27:35,680 --> 00:27:38,359
Speaker 13: position in our portfolio, where I think they have demonstrated

530
00:27:38,359 --> 00:27:41,920
Speaker 13: that extreme control and dominance over their entire supply chain

531
00:27:41,920 --> 00:27:45,080
Speaker 13: and the ability to coordinate multiple actors in an increasingly

532
00:27:45,119 --> 00:27:49,240
Speaker 13: complex ecosystem. And that means that as semiconductors become ever

533
00:27:49,320 --> 00:27:52,160
Speaker 13: more complicated to produce, TSMC just keeps getting better.

534
00:27:53,200 --> 00:27:56,600
Speaker 2: Sometimes you can't say what the story is from just

535
00:27:56,600 --> 00:28:00,239
Speaker 2: a few market sessions, but Friday through Tuesday's close, we

536
00:28:00,240 --> 00:28:03,919
Speaker 2: were almost in correction territory, particularly in the chip sector,

537
00:28:04,200 --> 00:28:08,040
Speaker 2: right and you're coming to us from sunny Scotland. I

538
00:28:08,080 --> 00:28:10,040
Speaker 2: was trying to make sense of what the net outcome

539
00:28:10,119 --> 00:28:12,560
Speaker 2: was of the President of the United States trip to China.

540
00:28:13,080 --> 00:28:14,879
Speaker 2: And without putting words in your mouth, I think that

541
00:28:14,920 --> 00:28:17,600
Speaker 2: you would say investors now start to look outside of

542
00:28:17,640 --> 00:28:19,360
Speaker 2: the US, look internationally.

543
00:28:20,680 --> 00:28:23,479
Speaker 13: I would certainly hope so there's such a rich hunting

544
00:28:23,520 --> 00:28:26,520
Speaker 13: ground for AI related names outside the US. I think

545
00:28:26,520 --> 00:28:29,480
Speaker 13: it's there's a really interesting dichotomy that chips and AI

546
00:28:29,640 --> 00:28:33,520
Speaker 13: is sort of designed in America, but it's manufactured internationally.

547
00:28:33,880 --> 00:28:37,160
Speaker 13: Whether it's TSMC actually making the chips themselves, whether it's

548
00:28:37,280 --> 00:28:41,440
Speaker 13: ke Heinex putting the memory inside the systems, or ASML

549
00:28:41,520 --> 00:28:46,640
Speaker 13: that builds the EUV machines the lithography machines for producing

550
00:28:46,640 --> 00:28:49,760
Speaker 13: the chips at TSMC FABS they do that in the Netherlands.

551
00:28:49,920 --> 00:28:52,680
Speaker 13: Like all of these companies are extremely interesting, again, extremely

552
00:28:52,680 --> 00:28:55,600
Speaker 13: dominant in their particular industries, and that makes them great

553
00:28:55,600 --> 00:28:57,840
Speaker 13: investments for a concentrated strategy.

554
00:28:58,320 --> 00:28:58,960
Speaker 5: Lake ICG.

555
00:28:59,520 --> 00:29:01,240
Speaker 13: I think the other thing that we can do, being

556
00:29:01,280 --> 00:29:05,560
Speaker 13: based in Sunny Scotland is take that step back and say, yes,

557
00:29:05,640 --> 00:29:07,920
Speaker 13: the market is doing its thing. There's a tendency, I

558
00:29:07,960 --> 00:29:10,720
Speaker 13: think for the market to price a lack of a

559
00:29:10,760 --> 00:29:11,520
Speaker 13: catalyst as.

560
00:29:11,440 --> 00:29:14,000
Speaker 5: A negative thing. The fact that there wasn't really any

561
00:29:14,040 --> 00:29:15,640
Speaker 5: news about at.

562
00:29:15,560 --> 00:29:18,440
Speaker 13: Chipex Sports coming out of the Chinese summit was seen

563
00:29:18,440 --> 00:29:20,360
Speaker 13: as a bad thing, But I think over the very

564
00:29:20,360 --> 00:29:22,800
Speaker 13: long term, none of that affects the fundamentals of any

565
00:29:22,840 --> 00:29:26,280
Speaker 13: of these companies. They are still extremely strong growing and

566
00:29:26,680 --> 00:29:30,360
Speaker 13: in fact, the fact that TSMC raised its guidance for

567
00:29:30,880 --> 00:29:34,320
Speaker 13: AI growth over the next five years to fifty six

568
00:29:34,360 --> 00:29:37,680
Speaker 13: percent is really encouraging because again they have this insight

569
00:29:37,720 --> 00:29:40,040
Speaker 13: into the entire supply chain, and so the fact that

570
00:29:40,280 --> 00:29:42,000
Speaker 13: they're able to do that with a great degree of

571
00:29:42,080 --> 00:29:44,200
Speaker 13: confidence gives us the conviction as well.

572
00:29:45,480 --> 00:29:48,720
Speaker 2: Paulina mcpadden of Bailey Gifford from Sunny Scotland putting a

573
00:29:48,760 --> 00:29:52,240
Speaker 2: pretty sunny disposition on this reporter's face, thank you very much. Indeed,

574
00:29:52,600 --> 00:29:55,760
Speaker 2: from public markets to the private markets again, today's Big

575
00:29:55,800 --> 00:29:59,560
Speaker 2: Take focuses on soft Banks Open Ai. Bet it's committed

576
00:29:59,560 --> 00:30:02,400
Speaker 2: more than sixty billion dollars to open Ai. We founder

577
00:30:02,440 --> 00:30:06,800
Speaker 2: Mashoosi's son reportedly convinced Sam Altman is leading the most

578
00:30:06,840 --> 00:30:10,520
Speaker 2: important technology shift of the century. But as rival Anthropic

579
00:30:10,600 --> 00:30:14,440
Speaker 2: gains ground, questions are growing even inside soft Bank about

580
00:30:14,440 --> 00:30:17,480
Speaker 2: whether the company may be too heavily tied to open

581
00:30:17,520 --> 00:30:21,640
Speaker 2: AI's success. Bloomberg's executive editor for Global Tech, Peter Elstrom,

582
00:30:21,680 --> 00:30:25,040
Speaker 2: has the story no surprise that around the world the

583
00:30:25,080 --> 00:30:28,360
Speaker 2: big tech take has clicked on right, take us inside it.

584
00:30:28,480 --> 00:30:30,640
Speaker 2: What are the details We're reporting, What do we learn

585
00:30:30,720 --> 00:30:35,080
Speaker 2: about this dynamic between Masosi's sun, his fixation on Sam Altman,

586
00:30:35,280 --> 00:30:37,600
Speaker 2: and what the rest of that organization is thinking.

587
00:30:39,360 --> 00:30:41,960
Speaker 14: Yeah, so this is a deep dive into the situation

588
00:30:42,120 --> 00:30:44,800
Speaker 14: at SoftBank. Masioshu Son, as you pointed out, has been

589
00:30:44,840 --> 00:30:48,240
Speaker 14: investing in starlups for a long time. He's had several

590
00:30:48,400 --> 00:30:52,200
Speaker 14: enormous hits, including Olive Babo was one of his biggest successes,

591
00:30:52,320 --> 00:30:55,040
Speaker 14: the Chinese Ai company, and then he went into the

592
00:30:55,120 --> 00:30:58,080
Speaker 14: Vision Fund and he made hundreds of birds on small AI,

593
00:30:58,520 --> 00:31:02,880
Speaker 14: small technology company, some successfully and some not so successfully.

594
00:31:03,080 --> 00:31:06,160
Speaker 14: And now with open Ai, he's committed more than sixty

595
00:31:06,280 --> 00:31:09,280
Speaker 14: billion dollars in capital, his biggest bet ever. It's a

596
00:31:09,400 --> 00:31:12,920
Speaker 14: very big concentration of the money that he's investing in

597
00:31:12,960 --> 00:31:15,000
Speaker 14: a single place, more money than he's ever put in

598
00:31:15,000 --> 00:31:18,640
Speaker 14: a single company before. He's not only selling assets, including

599
00:31:18,640 --> 00:31:21,640
Speaker 14: some Nvidia stock by the way, he's also borrowing some

600
00:31:21,720 --> 00:31:24,000
Speaker 14: money to be able to make this commitment to open

601
00:31:24,040 --> 00:31:26,600
Speaker 14: Ai and give Sam Altman the kind of capital that

602
00:31:26,640 --> 00:31:29,400
Speaker 14: he wants. Now, as you mentioned, there are concerns that

603
00:31:29,480 --> 00:31:33,280
Speaker 14: are being raised, including inside SoftBank, that maybe Monsieur Shesn

604
00:31:33,360 --> 00:31:33,960
Speaker 14: is a little.

605
00:31:33,720 --> 00:31:34,800
Speaker 7: Bit starstruck here.

606
00:31:34,960 --> 00:31:39,000
Speaker 14: Maybe he's being persuaded by this very charismatic founder to

607
00:31:39,040 --> 00:31:41,640
Speaker 14: put in more money than really should be at this point.

608
00:31:41,800 --> 00:31:45,040
Speaker 14: So it goes into some of the details of those concerns.

609
00:31:45,120 --> 00:31:46,640
Speaker 14: And as you point out, this is at a time

610
00:31:46,680 --> 00:31:51,240
Speaker 14: when open ai is facing some strategy challenge, some business challenges,

611
00:31:51,520 --> 00:31:54,640
Speaker 14: and also reputational challenges as we've seen with Sam Altman.

612
00:31:55,920 --> 00:31:59,800
Speaker 2: We're showing it sixty five billion dollars cumulatively through twenty twenty.

613
00:32:00,800 --> 00:32:04,480
Speaker 2: That's the reporting on the insider concern. What was soft

614
00:32:04,480 --> 00:32:05,840
Speaker 2: Bank's response to the big tape?

615
00:32:05,880 --> 00:32:10,000
Speaker 14: Peter Now soft Bank and open Ai, we should say,

616
00:32:10,040 --> 00:32:12,440
Speaker 14: pointed out that they have a great relationship. They feel

617
00:32:12,480 --> 00:32:15,480
Speaker 14: like it's a very strong partnership at this point. But

618
00:32:16,040 --> 00:32:17,920
Speaker 14: what some of the people have told us is that

619
00:32:17,920 --> 00:32:21,000
Speaker 14: there are a few areas of concern. First of all,

620
00:32:21,240 --> 00:32:24,160
Speaker 14: soft Bank and open Ai talked about this big stargate

621
00:32:24,240 --> 00:32:26,000
Speaker 14: venture that they were going to do in the United States.

622
00:32:26,080 --> 00:32:29,080
Speaker 14: To recall, it was last year when they got together

623
00:32:29,120 --> 00:32:31,080
Speaker 14: with President Trump and they talked about how they can

624
00:32:31,120 --> 00:32:33,800
Speaker 14: invest one hundred billion dollars, maybe even five hundred billion

625
00:32:33,840 --> 00:32:36,480
Speaker 14: dollars in the US. Now they have begun to make

626
00:32:36,520 --> 00:32:39,160
Speaker 14: some of those investments, but they've gone very very slowly,

627
00:32:39,440 --> 00:32:43,200
Speaker 14: and in the meantime, open Ai has struck some other

628
00:32:43,480 --> 00:32:46,920
Speaker 14: deals under this Stargate name. They've cut some deals with

629
00:32:46,960 --> 00:32:50,120
Speaker 14: other data center operators. So Massio she Son, who sort

630
00:32:50,120 --> 00:32:52,400
Speaker 14: of viewed himself as like an equal partner who's going

631
00:32:52,440 --> 00:32:54,120
Speaker 14: to be able to play a very key role at

632
00:32:54,160 --> 00:32:56,680
Speaker 14: this company, is not getting the kind of stature and

633
00:32:56,720 --> 00:32:58,680
Speaker 14: the attention that he would like. He doesn't have a

634
00:32:58,680 --> 00:33:00,680
Speaker 14: seat on the board. He doesn't even have an advisor

635
00:33:00,680 --> 00:33:02,520
Speaker 14: a sheeet on the board like many of the other

636
00:33:02,600 --> 00:33:03,400
Speaker 14: companies there.

637
00:33:03,440 --> 00:33:04,240
Speaker 3: But there are.

638
00:33:04,080 --> 00:33:08,120
Speaker 14: Constraints inside SoftBank in outside SoftBank about how this relationship

639
00:33:08,160 --> 00:33:08,960
Speaker 14: has going so far.

640
00:33:09,880 --> 00:33:12,920
Speaker 2: Bloomberg's Peter Elstrom with the big take, Thank you very much.

641
00:33:13,360 --> 00:33:16,400
Speaker 2: An update on the situation with Samsung's workers' union, which

642
00:33:16,440 --> 00:33:20,680
Speaker 2: has decided to suspend its planned strike. South Korea's Yonhap

643
00:33:20,720 --> 00:33:23,400
Speaker 2: news agency says the union will put a tentative wage

644
00:33:23,440 --> 00:33:27,040
Speaker 2: agreement with Samsung management to a vote. This comes after

645
00:33:27,120 --> 00:33:30,400
Speaker 2: days of stop and start negotiations with the union threatening

646
00:33:30,440 --> 00:33:34,400
Speaker 2: an eighteen day walkout that was supposed to start tomorrow.

647
00:33:34,480 --> 00:33:36,920
Speaker 2: We'll keep tracking that story and now coming up Forum

648
00:33:37,040 --> 00:33:40,600
Speaker 2: AI CEO Campbell Brown will be joining us discuss how

649
00:33:40,600 --> 00:33:46,280
Speaker 2: the major AI models are fundamentally unready to handle news

650
00:33:46,560 --> 00:33:51,760
Speaker 2: and geopolitics and Nvidia earnings after the market close. That

651
00:33:51,920 --> 00:33:54,280
Speaker 2: is the big story and factor in markets right now.

652
00:33:54,440 --> 00:33:57,040
Speaker 2: This is what markets look like at the index level.

653
00:33:57,080 --> 00:33:59,680
Speaker 2: That's that one hundred up one point four percent, big

654
00:33:59,720 --> 00:34:02,200
Speaker 2: rebound after three days of declinentes in the socks of

655
00:34:02,400 --> 00:34:05,000
Speaker 2: four percent. Nvidio is off its session highs but up

656
00:34:05,000 --> 00:34:07,960
Speaker 2: two percent. There's a lot riding. There is a beaten

657
00:34:08,080 --> 00:34:10,719
Speaker 2: raised kind of expectation that Jensen one is going to

658
00:34:10,719 --> 00:34:12,960
Speaker 2: tell us something big later today.

659
00:34:13,120 --> 00:34:14,120
Speaker 3: This is Bloomberg Tech.

660
00:34:21,040 --> 00:34:23,800
Speaker 2: A new study by four am ais says quote AI

661
00:34:23,880 --> 00:34:27,319
Speaker 2: companies are grading their own homework. The report reveals are

662
00:34:27,320 --> 00:34:31,440
Speaker 2: staggering ninety percent failure of rate on election questions across

663
00:34:31,480 --> 00:34:35,520
Speaker 2: four major chatbots, and they are routinely serving up Russian

664
00:34:35,880 --> 00:34:40,239
Speaker 2: and Chinese state media as authoritative sources. Joining us now

665
00:34:40,280 --> 00:34:43,360
Speaker 2: as the CEO of FORUMAI, Campbell Brown, who believes AI

666
00:34:43,400 --> 00:34:48,440
Speaker 2: companies have to be different. The headline Campbell is that

667
00:34:48,560 --> 00:34:50,680
Speaker 2: chatbots struggle with news accuracy.

668
00:34:50,800 --> 00:34:51,919
Speaker 3: If we're going to make.

669
00:34:51,840 --> 00:34:54,480
Speaker 2: It as simple as we can, but I think the

670
00:34:54,480 --> 00:34:57,200
Speaker 2: best place to start is with the methodology. You know

671
00:34:57,239 --> 00:35:00,680
Speaker 2: what you went and tested in how you tested it.

672
00:35:00,520 --> 00:35:01,200
Speaker 3: Sure ed.

673
00:35:01,640 --> 00:35:06,040
Speaker 15: We've looked at essentially three dimensions, which is factual accuracy

674
00:35:06,080 --> 00:35:09,200
Speaker 15: biased generally, and then the quality of the sources that

675
00:35:09,239 --> 00:35:12,080
Speaker 15: they used. And the way we did it was training

676
00:35:12,400 --> 00:35:17,960
Speaker 15: judgment models with senior domain experts who architected benchmarks. This

677
00:35:18,000 --> 00:35:20,040
Speaker 15: is an area I think it's worth pointing out that

678
00:35:20,120 --> 00:35:23,000
Speaker 15: hasn't gotten a lot of measurement. Most of the measurement

679
00:35:23,080 --> 00:35:25,640
Speaker 15: and benchmarks that you see around the models is focused

680
00:35:25,760 --> 00:35:30,040
Speaker 15: on coding and math and model capability, which makes sense.

681
00:35:30,120 --> 00:35:31,839
Speaker 15: I mean, this is where the model companies are making

682
00:35:31,840 --> 00:35:34,560
Speaker 15: their money, but they're also marketing the chat bots as

683
00:35:34,560 --> 00:35:37,840
Speaker 15: consumer products. People are using them for all kinds of

684
00:35:37,920 --> 00:35:41,279
Speaker 15: questions and certainly in an election year, political information is

685
00:35:41,280 --> 00:35:43,279
Speaker 15: going to be really important and they're not right now

686
00:35:43,320 --> 00:35:44,040
Speaker 15: where they need to be.

687
00:35:45,080 --> 00:35:47,280
Speaker 2: I want to get to the midterms just very quickly.

688
00:35:47,320 --> 00:35:50,160
Speaker 2: The evail process is called news bench wide, right, yep?

689
00:35:50,280 --> 00:35:54,200
Speaker 2: Do we have the why on why chatbots struggle with

690
00:35:54,239 --> 00:35:55,040
Speaker 2: news accuracy?

691
00:35:55,239 --> 00:35:56,000
Speaker 3: What's the cause?

692
00:35:57,440 --> 00:36:00,719
Speaker 15: Honestly, I think it hasn't been a priority. As I said,

693
00:36:01,080 --> 00:36:04,080
Speaker 15: they're leaning into other areas in terms of the measurement.

694
00:36:05,040 --> 00:36:08,880
Speaker 15: I think that's going to change, not only because consumers

695
00:36:09,160 --> 00:36:12,520
Speaker 15: are looking for it and going to demand it eventually,

696
00:36:12,520 --> 00:36:14,800
Speaker 15: but I think enterprise is starting to demanding it, to

697
00:36:14,960 --> 00:36:18,240
Speaker 15: starting to demand it. And that's again where their focus

698
00:36:18,320 --> 00:36:21,719
Speaker 15: is from a business perspective, and being just okay on

699
00:36:21,840 --> 00:36:25,360
Speaker 15: accuracy overall on news and politics and geopolitics is not

700
00:36:25,400 --> 00:36:26,000
Speaker 15: going to cut it.

701
00:36:27,120 --> 00:36:30,160
Speaker 2: Let's talk about the midterms them. The Bloomberg report states

702
00:36:30,160 --> 00:36:33,799
Speaker 2: the chatbots answers about elections failed on accuracy, bias, or

703
00:36:33,840 --> 00:36:36,839
Speaker 2: source selection ninety percent of the time. What I found

704
00:36:36,880 --> 00:36:40,560
Speaker 2: interesting reading about that is there must be evidence.

705
00:36:40,160 --> 00:36:41,560
Speaker 3: Therefore that the electorate.

706
00:36:41,840 --> 00:36:46,000
Speaker 2: Everyday Americans go to chatbots for news in the first place.

707
00:36:46,480 --> 00:36:49,440
Speaker 15: You're seeing those numbers increase, I think more and more.

708
00:36:49,560 --> 00:36:51,840
Speaker 15: We've seen a couple of studies showing people are using

709
00:36:51,880 --> 00:36:54,839
Speaker 15: them for news more and more. And we found that

710
00:36:55,040 --> 00:36:58,040
Speaker 15: about third a third of the questions we asked or

711
00:36:58,040 --> 00:37:01,279
Speaker 15: on election related questions had faction errors in them, but

712
00:37:01,600 --> 00:37:06,080
Speaker 15: all of the chatbots failed on bias. Claude Gemini gave

713
00:37:06,239 --> 00:37:09,919
Speaker 15: left leaning answers on election related questions one hundred percent

714
00:37:10,000 --> 00:37:12,840
Speaker 15: of the time. Chat GBT ninety five percent of the time.

715
00:37:13,280 --> 00:37:17,359
Speaker 15: Grock was the only right leaning chatbot that gave right

716
00:37:17,440 --> 00:37:20,919
Speaker 15: leaning answers on these questions, I think about eighty five

717
00:37:20,960 --> 00:37:23,840
Speaker 15: percent of the time, So you were by no means

718
00:37:23,960 --> 00:37:27,440
Speaker 15: getting a straight up answer or a balanced perspective, and

719
00:37:27,520 --> 00:37:30,319
Speaker 15: people are asking, you know, who are my candidates, what

720
00:37:30,360 --> 00:37:32,240
Speaker 15: are their positions on the issues.

721
00:37:31,880 --> 00:37:33,120
Speaker 5: And who should I vote for?

722
00:37:33,280 --> 00:37:35,920
Speaker 15: So improving this between now and then, I mean, the

723
00:37:35,920 --> 00:37:40,200
Speaker 15: one thing I am positive hopeful about is there was

724
00:37:40,239 --> 00:37:44,480
Speaker 15: a broad difference. I think overall, Gemini handled a lot

725
00:37:44,480 --> 00:37:46,800
Speaker 15: of the questions better than some of the other models,

726
00:37:46,800 --> 00:37:49,520
Speaker 15: which shows there's area for improvement. But you can't fix

727
00:37:49,520 --> 00:37:50,640
Speaker 15: something if you haven't measured it.

728
00:37:51,920 --> 00:37:55,240
Speaker 2: Chimboll, I know you previously, of course a Meta vice

729
00:37:55,280 --> 00:37:58,239
Speaker 2: president of News and Global Media Partnerships, but also as

730
00:37:58,239 --> 00:38:01,400
Speaker 2: an industry colleague award win journalists and an k CNN,

731
00:38:01,520 --> 00:38:03,960
Speaker 2: NBC News both sides you see.

732
00:38:04,680 --> 00:38:07,840
Speaker 3: Do you see action on the news.

733
00:38:07,719 --> 00:38:11,560
Speaker 2: Side and the likes of Meta other frontier labs on

734
00:38:11,680 --> 00:38:14,239
Speaker 2: actively fixing this taking action on it?

735
00:38:14,520 --> 00:38:15,320
Speaker 3: Yeah, I do.

736
00:38:15,719 --> 00:38:17,759
Speaker 15: I mean I talk to people in the labs on

737
00:38:17,800 --> 00:38:20,120
Speaker 15: a regular basis that all of these companies and they

738
00:38:20,120 --> 00:38:22,799
Speaker 15: do care about this. And are beginning to approach it

739
00:38:23,880 --> 00:38:26,640
Speaker 15: with a sort of different way of thinking. I think

740
00:38:27,480 --> 00:38:29,640
Speaker 15: you know, one of the challenges we had, and you

741
00:38:29,640 --> 00:38:32,680
Speaker 15: said it in the opening, is today these model companies

742
00:38:32,719 --> 00:38:36,879
Speaker 15: are essentially grading their own homework. There's not independent evaluation.

743
00:38:37,440 --> 00:38:40,160
Speaker 15: And I'm not calling for regulation. I have a private company,

744
00:38:40,160 --> 00:38:43,239
Speaker 15: but I do think we need an ecosystem of companies

745
00:38:43,360 --> 00:38:47,799
Speaker 15: and nonprofits that are doing independent evaluation. And I think

746
00:38:47,840 --> 00:38:52,399
Speaker 15: the companies, the model companies that lean into this are

747
00:38:52,440 --> 00:38:55,839
Speaker 15: going to make ultimately more trustworthy products. And I do

748
00:38:55,880 --> 00:38:57,840
Speaker 15: think you're going to see the demand move in that

749
00:38:57,920 --> 00:39:02,040
Speaker 15: direction again from enterprise. You're already seeing some states pass

750
00:39:02,160 --> 00:39:05,200
Speaker 15: laws where they're requiring independent evaluation, but I think it's

751
00:39:05,200 --> 00:39:07,960
Speaker 15: increasingly going to become more important.

752
00:39:08,920 --> 00:39:12,680
Speaker 2: Campbell Brown, CEO Forum AI. It's great to have you

753
00:39:12,680 --> 00:39:15,520
Speaker 2: on Blueberg Tech. Thank you very much. Indeed, coming up,

754
00:39:15,520 --> 00:39:16,920
Speaker 2: we're going to get back to the top story what

755
00:39:17,040 --> 00:39:20,760
Speaker 2: to expect from Nvidia's earnings the chip maker off session

756
00:39:20,840 --> 00:39:24,520
Speaker 2: highs still up just more than two percent, top line

757
00:39:24,520 --> 00:39:29,160
Speaker 2: growth eighty percent, EPs growth eighty five percent. The market

758
00:39:29,200 --> 00:39:32,359
Speaker 2: wants more than that. From Jensen one, this is Bloomberg Tech.

759
00:39:45,200 --> 00:39:50,120
Speaker 16: We've been licensed for many customers in China for Age

760
00:39:50,120 --> 00:39:55,320
Speaker 16: two hundred. We have received purchase orders from many customers

761
00:39:55,640 --> 00:39:59,200
Speaker 16: and we're in the process of restarting our manufacturing.

762
00:40:01,400 --> 00:40:04,160
Speaker 2: That was Jensen one in March, singing a pretty positive

763
00:40:04,200 --> 00:40:07,840
Speaker 2: tune about selling AI chips to China. This was the

764
00:40:07,920 --> 00:40:10,560
Speaker 2: Nvidia CEO striking a slightly different note when you spoke

765
00:40:10,600 --> 00:40:12,160
Speaker 2: to us on Monday, and.

766
00:40:12,239 --> 00:40:15,440
Speaker 8: The Chinese government has to decide how much of their

767
00:40:15,520 --> 00:40:17,839
Speaker 8: local market do they want to protect and how much

768
00:40:17,880 --> 00:40:19,360
Speaker 8: of their local market do they want to.

769
00:40:19,360 --> 00:40:22,520
Speaker 7: Expand with a more AI capacity.

770
00:40:23,480 --> 00:40:26,120
Speaker 2: That needs to be reconciled. Some analysts are hoping to

771
00:40:26,120 --> 00:40:29,040
Speaker 2: get more insight into Nvidia's future in China in its

772
00:40:29,080 --> 00:40:31,760
Speaker 2: earnings later today. There's also the battle in video faces

773
00:40:32,040 --> 00:40:35,960
Speaker 2: with custom Silicon Kungensabanni from Bloomberg Intelligence here to talk

774
00:40:36,040 --> 00:40:39,520
Speaker 2: us through those expectations, writing that Nvidia has extended its

775
00:40:39,520 --> 00:40:43,839
Speaker 2: AI lead even as custom ACIC competition grows, and so

776
00:40:44,000 --> 00:40:45,920
Speaker 2: I'm trying to understand this, right, We're talking about a

777
00:40:45,960 --> 00:40:49,439
Speaker 2: world where we go from training to inference and then

778
00:40:49,600 --> 00:40:53,280
Speaker 2: the custom acis also just as with the Nvidia GPUs,

779
00:40:53,360 --> 00:40:55,840
Speaker 2: or in a world where demand is outpacing supply. So

780
00:40:55,880 --> 00:40:58,640
Speaker 2: we're trying to work out the market dynamic. What do

781
00:40:58,640 --> 00:40:59,520
Speaker 2: you think we'll learn tonight?

782
00:41:00,880 --> 00:41:03,440
Speaker 17: Yeah, I mean, look tonight, the standard beaten raise numbers,

783
00:41:03,560 --> 00:41:05,480
Speaker 17: given where we are in the time and where the

784
00:41:05,480 --> 00:41:10,400
Speaker 17: stock is, don't matter unless they are able to pass

785
00:41:10,520 --> 00:41:13,640
Speaker 17: through a ninety billion dollar hurdle mark, which is the

786
00:41:13,719 --> 00:41:16,080
Speaker 17: high bar for the guide for the next quarter. Now,

787
00:41:16,120 --> 00:41:18,799
Speaker 17: that should really make the buyside balls happy. I think

788
00:41:18,840 --> 00:41:22,200
Speaker 17: tonight the key focus is going to be reassurance. There

789
00:41:22,200 --> 00:41:25,240
Speaker 17: has been some noise around supply chain issues or liquid

790
00:41:25,280 --> 00:41:28,000
Speaker 17: cooling with that Rubin ramp. We really need to hear

791
00:41:28,080 --> 00:41:31,240
Speaker 17: that the Rubin ramp for the second half remains intact,

792
00:41:31,400 --> 00:41:33,720
Speaker 17: and we would like to see that one trillion dollar

793
00:41:33,880 --> 00:41:36,680
Speaker 17: demand pipeline raise up the numbers from here.

794
00:41:37,520 --> 00:41:40,840
Speaker 2: So give me that ninety billion figure again, I just

795
00:41:40,840 --> 00:41:44,560
Speaker 2: put fa go on the Bloomberg back up, then consent,

796
00:41:45,080 --> 00:41:45,879
Speaker 2: go for it, go for it.

797
00:41:46,360 --> 00:41:48,279
Speaker 17: When the guide for the next quarter, the high bar,

798
00:41:48,520 --> 00:41:50,560
Speaker 17: it seems on the byside is ninety billion, So if

799
00:41:50,560 --> 00:41:53,000
Speaker 17: they're able to clear that, that should be a very

800
00:41:53,000 --> 00:41:53,800
Speaker 17: positive event.

801
00:41:54,560 --> 00:41:58,359
Speaker 2: Fiscal two Q twenty seven consensus is eighty seven point

802
00:41:58,440 --> 00:42:01,680
Speaker 2: three billion, the high side for the guide in the

803
00:42:01,680 --> 00:42:04,879
Speaker 2: current period being ninety billion dollars, so fascinating.

804
00:42:05,640 --> 00:42:07,880
Speaker 3: You know, I think Jensmongks talked about the.

805
00:42:09,640 --> 00:42:14,360
Speaker 2: Performance of it in Vidia based system against TPU against

806
00:42:14,360 --> 00:42:17,799
Speaker 2: other a six he may well get asked again, is

807
00:42:17,840 --> 00:42:20,120
Speaker 2: there any data set out there in the world, con Jan,

808
00:42:20,560 --> 00:42:24,080
Speaker 2: The evidence is your argument that the VERA Rubin platform

809
00:42:24,120 --> 00:42:27,120
Speaker 2: at least is extending in video's lead rather than seeing

810
00:42:27,120 --> 00:42:28,320
Speaker 2: its market share shrink.

811
00:42:29,840 --> 00:42:33,000
Speaker 17: There have been some third party you know, benchmarks, and

812
00:42:33,200 --> 00:42:35,480
Speaker 17: as we as the systems LU role out, which they

813
00:42:35,480 --> 00:42:38,600
Speaker 17: haven't yet, there will be more benchmarks. And we have

814
00:42:38,640 --> 00:42:41,760
Speaker 17: seen consistently in Vidia system the latest in video systems,

815
00:42:41,800 --> 00:42:45,200
Speaker 17: outperform what exists in the market today. But now it's

816
00:42:45,280 --> 00:42:49,680
Speaker 17: no longer just about your performance. You outperforming on the specs. Right,

817
00:42:49,920 --> 00:42:53,319
Speaker 17: there's a use case. There's a concept of dollars per

818
00:42:53,360 --> 00:42:55,839
Speaker 17: token per performance. So when we look at someone thing

819
00:42:55,920 --> 00:42:59,399
Speaker 17: like a TPU, where the entire architecture of the data

820
00:42:59,400 --> 00:43:03,120
Speaker 17: center is controlled and driven by the company using it itself,

821
00:43:03,600 --> 00:43:06,040
Speaker 17: they are able to achieve a much more a much

822
00:43:06,080 --> 00:43:08,520
Speaker 17: lower dollar per performance, which matters to them.

823
00:43:08,400 --> 00:43:10,360
Speaker 3: A lot real quick.

824
00:43:10,600 --> 00:43:15,560
Speaker 2: The Bloomberg intelligence thesis on China.

825
00:43:14,080 --> 00:43:17,799
Speaker 17: Yeah, we didn't hear anything positive coming out of the

826
00:43:17,840 --> 00:43:21,000
Speaker 17: trip with trumpet musicians. So we are not going to

827
00:43:21,040 --> 00:43:24,320
Speaker 17: increase or just any of our China revenue estimates anytime

828
00:43:25,239 --> 00:43:25,759
Speaker 17: in the New.

829
00:43:25,719 --> 00:43:30,520
Speaker 2: Town Keunjen Sabani, Bloomberg Intelligence, Thank you very much. That

830
00:43:30,640 --> 00:43:33,520
Speaker 2: does it for this edition of Bloomberg Tech. One last

831
00:43:33,560 --> 00:43:36,600
Speaker 2: look the countdown to in Video's earnings and A's that

832
00:43:36,719 --> 00:43:40,440
Speaker 2: one hundred session high one point five percent, socks rebounding

833
00:43:40,480 --> 00:43:42,480
Speaker 2: up four percent, and in Video is kind of the

834
00:43:42,480 --> 00:43:44,400
Speaker 2: heart of that story. There is a lot of optimism,

835
00:43:44,440 --> 00:43:47,839
Speaker 2: but Kanjensvanni just said, raised feet doesn't matter. It's all

836
00:43:47,840 --> 00:43:50,279
Speaker 2: about the outlook for the current period. We're just over

837
00:43:50,320 --> 00:43:53,560
Speaker 2: four hours away from A Video's fiscal first quarter twenty

838
00:43:53,640 --> 00:43:57,680
Speaker 2: twenty seven financial year earnings recap. On the podcast, you

839
00:43:57,680 --> 00:44:00,080
Speaker 2: know exactly where to find it and I would of

840
00:44:00,200 --> 00:44:02,920
Speaker 2: quality conversations throughout the hour. Thank you very much for

841
00:44:03,000 --> 00:44:04,560
Speaker 2: joining us. This is Bloomberg Tech.