1
00:00:02,720 --> 00:00:15,880
Speaker 1: Bloomberg Audio Studios, Podcasts, radio News.

2
00:00:18,600 --> 00:00:21,760
Speaker 2: Hello and welcome to another episode of the All Thoughts podcast.

3
00:00:21,960 --> 00:00:23,439
Speaker 3: I'm Tracy Allaway.

4
00:00:23,079 --> 00:00:24,360
Speaker 4: And I'm Jill Wisenthal.

5
00:00:24,840 --> 00:00:25,480
Speaker 3: So Joe.

6
00:00:25,560 --> 00:00:28,800
Speaker 2: We are here in Hong Kong still and we're learning

7
00:00:28,800 --> 00:00:31,240
Speaker 2: a lot of different things, a lot of interesting things.

8
00:00:31,400 --> 00:00:33,920
Speaker 2: One of the most interesting things we heard actually came

9
00:00:34,280 --> 00:00:37,360
Speaker 2: yesterday from the bay Do CFO where we were just

10
00:00:37,440 --> 00:00:40,640
Speaker 2: talking casually ahead of our interview, but he was saying

11
00:00:40,680 --> 00:00:45,879
Speaker 2: that the word token has now officially been added to

12
00:00:46,440 --> 00:00:50,879
Speaker 2: the Chinese Mandarin dictionary and that the characters that actually

13
00:00:50,920 --> 00:00:54,680
Speaker 2: make up the Chinese word for token are something like

14
00:00:54,960 --> 00:00:56,000
Speaker 2: word currency.

15
00:00:56,280 --> 00:00:59,240
Speaker 4: I think that's so fascinating, you know, I'm fascinated by

16
00:00:59,280 --> 00:01:02,720
Speaker 4: the etymology the word token specifically, so I was thrilled

17
00:01:03,200 --> 00:01:04,920
Speaker 4: to hear that. But it is, you know, it's like,

18
00:01:04,959 --> 00:01:07,720
Speaker 4: we know the word token in the monetary context, we

19
00:01:07,800 --> 00:01:11,200
Speaker 4: no chuck e cheese tokens, no crypto tookens. But we

20
00:01:11,280 --> 00:01:14,280
Speaker 4: also know that since the middle of the twentieth century,

21
00:01:14,560 --> 00:01:19,039
Speaker 4: linguists have been using token to describe more or less

22
00:01:19,040 --> 00:01:22,160
Speaker 4: a word. And then obviously with LLLMS, we talk about

23
00:01:22,160 --> 00:01:25,080
Speaker 4: these linguistic tokens a lot so to see that in

24
00:01:25,200 --> 00:01:28,840
Speaker 4: Chinese the formal term is a merger of these two concepts.

25
00:01:29,040 --> 00:01:34,080
Speaker 4: Is I find a very intellectually satisfying thing to learn about.

26
00:01:34,319 --> 00:01:36,920
Speaker 2: Bringing up chuck e cheese tokens is really a way

27
00:01:36,959 --> 00:01:39,640
Speaker 2: to make sure everyone knows that you're a millennial.

28
00:01:39,720 --> 00:01:42,000
Speaker 4: Tem Yeah, that's right, that's right, arcade, I should just

29
00:01:42,000 --> 00:01:43,440
Speaker 4: say arcade token, that's right.

30
00:01:43,560 --> 00:01:44,200
Speaker 3: Yeah, okay.

31
00:01:44,520 --> 00:01:47,880
Speaker 2: And I think what's really interesting about word currency itself

32
00:01:48,040 --> 00:01:52,600
Speaker 2: is it implies that it's connected to spending, right of course,

33
00:01:52,840 --> 00:01:54,840
Speaker 2: And so when you think about the big companies that

34
00:01:54,880 --> 00:01:57,520
Speaker 2: are spending all this money on tokens, and I guess

35
00:01:57,520 --> 00:02:01,360
Speaker 2: the AI infrastructure build out more broadly, CFO has become

36
00:02:01,440 --> 00:02:03,080
Speaker 2: really important, right totally.

37
00:02:03,120 --> 00:02:06,920
Speaker 4: So we you know, the headlines by and large are

38
00:02:06,960 --> 00:02:10,520
Speaker 4: about Capex, right, and that's going to continue to be

39
00:02:10,520 --> 00:02:12,440
Speaker 4: because of the data center builder. We're going to talk

40
00:02:12,440 --> 00:02:14,680
Speaker 4: about that, but a lot of it is also going

41
00:02:14,720 --> 00:02:16,960
Speaker 4: to be opex and figuring out how to within a

42
00:02:17,040 --> 00:02:21,640
Speaker 4: company allocating token permissions and caps and so forth. And

43
00:02:21,720 --> 00:02:25,160
Speaker 4: I doubt anyone has figured out the final answer. But

44
00:02:25,280 --> 00:02:29,040
Speaker 4: if two different people can get different value out of

45
00:02:29,600 --> 00:02:32,880
Speaker 4: using AI models, then there is no way that it

46
00:02:32,919 --> 00:02:35,000
Speaker 4: makes sense for them to have the same token budgets.

47
00:02:35,080 --> 00:02:36,959
Speaker 4: By the way, Tracy, can I ask you a personal

48
00:02:37,080 --> 00:02:38,480
Speaker 4: question that I've never asked you before?

49
00:02:38,840 --> 00:02:39,120
Speaker 3: Oh?

50
00:02:39,240 --> 00:02:42,840
Speaker 4: Okay, outside of the work context, are you a macro

51
00:02:42,960 --> 00:02:43,720
Speaker 4: PC person?

52
00:02:44,320 --> 00:02:47,959
Speaker 2: Oh? I, I only have I only have work computers

53
00:02:48,000 --> 00:02:48,480
Speaker 2: at the moment.

54
00:02:48,520 --> 00:02:50,160
Speaker 4: Okay, so all right, before that, but.

55
00:02:51,520 --> 00:02:53,640
Speaker 3: Before that, definitely PC?

56
00:02:53,800 --> 00:02:54,239
Speaker 4: Okay, good?

57
00:02:54,639 --> 00:02:57,519
Speaker 3: Yeah, And in fact, it's not it's not a choice,

58
00:02:57,680 --> 00:02:57,840
Speaker 3: is it.

59
00:02:58,080 --> 00:02:58,200
Speaker 1: Like?

60
00:02:58,320 --> 00:03:00,560
Speaker 3: I sorry, I really don't like Max.

61
00:03:00,680 --> 00:03:04,760
Speaker 4: I don't either, And in particular, I am long been

62
00:03:04,800 --> 00:03:08,040
Speaker 4: a fan of the what used to be the IBM

63
00:03:08,160 --> 00:03:11,440
Speaker 4: think Pad laptop with the famous red button, which is

64
00:03:11,480 --> 00:03:12,480
Speaker 4: now owned by Lenovo.

65
00:03:13,320 --> 00:03:16,440
Speaker 2: That's right, you've talked about that computer before. I can

66
00:03:16,680 --> 00:03:20,120
Speaker 2: say that this is categorically true. Joe likes that computer.

67
00:03:20,680 --> 00:03:23,840
Speaker 2: But what's interesting about Lenovo is like, okay, it's famous

68
00:03:23,840 --> 00:03:26,680
Speaker 2: for the computer with a little red button in the middle,

69
00:03:26,760 --> 00:03:27,760
Speaker 2: but it's now.

70
00:03:27,480 --> 00:03:28,760
Speaker 3: Making an AI play.

71
00:03:28,919 --> 00:03:32,200
Speaker 2: Sure, I mean everyone's making an AI play, but it's

72
00:03:32,320 --> 00:03:36,720
Speaker 2: doing it from a different perspective. So AI integrating into

73
00:03:37,000 --> 00:03:40,960
Speaker 2: the actual computer, the hardware, but it's also doing cloud right,

74
00:03:41,200 --> 00:03:44,240
Speaker 2: So this is a really good opportunity to I guess

75
00:03:44,440 --> 00:03:48,080
Speaker 2: take the temperature on AI, the AI build out, the

76
00:03:48,120 --> 00:03:53,160
Speaker 2: AI spend from a bunch of different perspectives. So we do,

77
00:03:53,240 --> 00:03:55,520
Speaker 2: in fact have the perfect guest. We're going to be

78
00:03:55,520 --> 00:04:00,000
Speaker 2: speaking with, Winston Chang. He is the CFO of Lenovo's

79
00:04:00,200 --> 00:04:02,720
Speaker 2: Winston thanks so much for coming on O lots.

80
00:04:02,400 --> 00:04:04,240
Speaker 5: Hey, thank you Tracy, and thank you Joe.

81
00:04:04,400 --> 00:04:07,280
Speaker 2: Why don't you go ahead and describe what Lenovo is

82
00:04:07,480 --> 00:04:10,280
Speaker 2: and what the AI play is and how it actually

83
00:04:10,280 --> 00:04:11,960
Speaker 2: fits into the existing business.

84
00:04:12,360 --> 00:04:16,760
Speaker 5: Yeah, Lenovo is a global AI infrastructure company that provides

85
00:04:16,880 --> 00:04:22,040
Speaker 5: pocket to cloud AI infrastructure for the consumer and the enterprise.

86
00:04:22,279 --> 00:04:25,239
Speaker 5: I think that's really in a nutshell, and so today

87
00:04:25,560 --> 00:04:29,480
Speaker 5: we're able to provide this to hyperscalers which are doing

88
00:04:29,640 --> 00:04:33,960
Speaker 5: a lot of the spending are driven by training demand today.

89
00:04:34,240 --> 00:04:38,120
Speaker 5: And then given our IBM x eighty six heritage, which

90
00:04:38,120 --> 00:04:41,320
Speaker 5: we also acquire the server from the IBM actually, so

91
00:04:41,720 --> 00:04:45,000
Speaker 5: we actually are very strong in CPU compute as well,

92
00:04:45,560 --> 00:04:47,840
Speaker 5: given that IBM was a dominant player in the x

93
00:04:47,920 --> 00:04:51,680
Speaker 5: eighty six architecture. So from that perspective, we're well positioned

94
00:04:51,680 --> 00:04:56,000
Speaker 5: for inferencing needs of the enterprise and also the hyper

95
00:04:56,000 --> 00:04:59,120
Speaker 5: scalers in terms of in the cloud as well. So

96
00:04:59,160 --> 00:05:01,919
Speaker 5: I think from the person active of then you talked

97
00:05:01,920 --> 00:05:05,840
Speaker 5: about tokens, and I think from a token perspective, really

98
00:05:06,120 --> 00:05:08,480
Speaker 5: today people are on the early stage of how much

99
00:05:08,520 --> 00:05:12,400
Speaker 5: I'm really paying for I heard someone saying that there

100
00:05:12,480 --> 00:05:14,920
Speaker 5: was an engineer at a Portuit company which I will

101
00:05:14,920 --> 00:05:17,720
Speaker 5: not mention, that apparently spend one hundred million dollars a

102
00:05:17,760 --> 00:05:21,840
Speaker 5: month on tokens. And so as a CFO, I would

103
00:05:21,839 --> 00:05:23,920
Speaker 5: have concerns because that was clearly not in the budget,

104
00:05:24,000 --> 00:05:26,480
Speaker 5: right not saying that I was from the Novo So

105
00:05:26,520 --> 00:05:29,080
Speaker 5: we cannot have that happen. And I think we need

106
00:05:29,120 --> 00:05:33,480
Speaker 5: to be able to drive the productivity or efficiencies as

107
00:05:33,480 --> 00:05:35,919
Speaker 5: it relates to that budgeting of the token generation. And

108
00:05:35,960 --> 00:05:38,080
Speaker 5: I think a lot of that would happen on device

109
00:05:38,520 --> 00:05:41,520
Speaker 5: where you may just pay a higher price for a device,

110
00:05:41,960 --> 00:05:43,680
Speaker 5: but you know what you're going to be able to

111
00:05:43,760 --> 00:05:47,320
Speaker 5: do on compute for the security of the data and

112
00:05:47,400 --> 00:05:49,760
Speaker 5: for the privacy that you want to interact with your

113
00:05:49,760 --> 00:05:50,279
Speaker 5: AI agent.

114
00:05:50,600 --> 00:05:54,200
Speaker 2: So just to be clear, for the computers themselves, they

115
00:05:54,240 --> 00:05:58,000
Speaker 2: can do some inference, right, but where it makes sense

116
00:05:58,160 --> 00:05:59,719
Speaker 2: you route it to the cloud.

117
00:06:00,080 --> 00:06:00,560
Speaker 3: Is that right?

118
00:06:01,000 --> 00:06:05,920
Speaker 5: That that is our goal. So the Lenovo agentic AI today,

119
00:06:06,240 --> 00:06:09,919
Speaker 5: what we are good at is really integrating and maximizing

120
00:06:09,960 --> 00:06:13,800
Speaker 5: the compute capabilities on a device. And from that perspective,

121
00:06:13,880 --> 00:06:18,400
Speaker 5: given the various needs in terms of various operating systems

122
00:06:19,279 --> 00:06:21,560
Speaker 5: agents that may now want to sit on top of

123
00:06:21,560 --> 00:06:25,120
Speaker 5: a device, I think our agent aims to orchestrate the

124
00:06:25,200 --> 00:06:29,159
Speaker 5: various LMS, will take the compressed versions of these lllms.

125
00:06:29,680 --> 00:06:33,560
Speaker 5: We will, depending on the partnership, be able to do

126
00:06:33,600 --> 00:06:38,080
Speaker 5: the on device compress LLMS versions and do that as

127
00:06:38,120 --> 00:06:41,760
Speaker 5: a local compute, but in certain queries allow it to

128
00:06:41,880 --> 00:06:45,520
Speaker 5: go on the cloud and therefore probably would allow the

129
00:06:45,600 --> 00:06:49,160
Speaker 5: user probably to spend in terms of the token generation.

130
00:06:49,480 --> 00:06:53,080
Speaker 4: It's interesting this word orchestrate because so you have the

131
00:06:53,160 --> 00:06:56,599
Speaker 4: AI agent orchestrating maybe a bunch of different subagents to

132
00:06:56,640 --> 00:06:59,960
Speaker 4: complete some tasks, and that is something that perhaps is

133
00:07:00,360 --> 00:07:03,919
Speaker 4: done best on a CPU. A company that builds servers

134
00:07:04,080 --> 00:07:07,640
Speaker 4: is also an orchestrator of the supply chain and acquiring

135
00:07:07,680 --> 00:07:11,040
Speaker 4: the different components that go into a server, et cetera.

136
00:07:11,400 --> 00:07:13,080
Speaker 4: I want to go back to something you said in

137
00:07:13,160 --> 00:07:15,600
Speaker 4: your first answer, because I think this action will get

138
00:07:15,640 --> 00:07:18,360
Speaker 4: to the core of this new era. You said, Okay,

139
00:07:18,840 --> 00:07:21,280
Speaker 4: an engineer spends one hundred million dollars in a month

140
00:07:21,320 --> 00:07:23,880
Speaker 4: on tokens, And it's like that would not make you

141
00:07:23,920 --> 00:07:26,080
Speaker 4: happy as a CFO, but it could.

142
00:07:25,840 --> 00:07:26,440
Speaker 5: Make you happy.

143
00:07:26,480 --> 00:07:30,040
Speaker 4: Right. What if you had a multi year database migration

144
00:07:30,320 --> 00:07:32,760
Speaker 4: plan that you think, oh, this would be a five

145
00:07:32,840 --> 00:07:36,120
Speaker 4: hundred million dollar job and the engineer does it in

146
00:07:36,160 --> 00:07:40,280
Speaker 4: a month for one hundred million dollars via tokens. Don't

147
00:07:40,360 --> 00:07:43,360
Speaker 4: you have to at least be open to the possibility

148
00:07:43,800 --> 00:07:45,080
Speaker 4: that that was money well spent.

149
00:07:45,840 --> 00:07:50,360
Speaker 5: Absolutely, Joe. I think everything is about the return, right

150
00:07:50,400 --> 00:07:53,440
Speaker 5: and the planning. So we're not afraid to invest. As

151
00:07:53,440 --> 00:07:56,240
Speaker 5: a CFO. You are there to allocate capital. You're not

152
00:07:56,360 --> 00:08:00,160
Speaker 5: there to constrain capital. You have to allocate, but have

153
00:08:00,200 --> 00:08:02,880
Speaker 5: to be clear in terms of that return. And I

154
00:08:02,880 --> 00:08:06,520
Speaker 5: think in that case, it's probably one where they weren't

155
00:08:06,600 --> 00:08:10,480
Speaker 5: sure in terms of that what they were particularly doing

156
00:08:10,520 --> 00:08:13,080
Speaker 5: in terms of spending. It wasn't in budgeting, and that

157
00:08:13,120 --> 00:08:15,560
Speaker 5: goes to the point of what is happening today. I

158
00:08:15,560 --> 00:08:19,840
Speaker 5: think most enterprises were at the early stages of how

159
00:08:19,960 --> 00:08:23,480
Speaker 5: people were changing from a subscription based model to a

160
00:08:23,520 --> 00:08:27,560
Speaker 5: token usage model in terms of the compute capabilities, and

161
00:08:27,600 --> 00:08:31,040
Speaker 5: so that is at the beginning and enterprises are starting

162
00:08:31,040 --> 00:08:33,560
Speaker 5: to figure out how do I really track that spend

163
00:08:33,840 --> 00:08:35,839
Speaker 5: and therefore, what do I really want to get out

164
00:08:35,840 --> 00:08:38,280
Speaker 5: of that return? And so we're on that early stage.

165
00:08:38,360 --> 00:08:41,679
Speaker 5: So you're absolutely right in terms of that we would

166
00:08:41,720 --> 00:08:44,360
Speaker 5: want to spend if we can get a return out

167
00:08:44,360 --> 00:08:46,679
Speaker 5: of it. You're absolutely You're absolutely right. And I think

168
00:08:46,960 --> 00:08:50,319
Speaker 5: preliminary data and anecdotal evidence from a lot of people

169
00:08:50,320 --> 00:08:53,560
Speaker 5: in a chat recognized the power of AI and the

170
00:08:53,640 --> 00:08:59,760
Speaker 5: probability of generating increased productivity and potentially even cost savings.

171
00:09:00,200 --> 00:09:04,080
Speaker 2: More about measuring the actual return, And again I realized

172
00:09:04,120 --> 00:09:06,840
Speaker 2: it's early stages, But I'm curious how you think it

173
00:09:06,880 --> 00:09:09,720
Speaker 2: will actually be done because we've been asking a bunch

174
00:09:09,800 --> 00:09:14,920
Speaker 2: of executives how do you internally benchmark your AI projects

175
00:09:15,040 --> 00:09:19,640
Speaker 2: or your token spend? And everyone says productivity or cost

176
00:09:19,679 --> 00:09:23,640
Speaker 2: savings or whatever, But those are they sound very theoretical.

177
00:09:23,920 --> 00:09:28,120
Speaker 2: So is there more concrete? I guess KPIs that you

178
00:09:28,120 --> 00:09:29,040
Speaker 2: were looking at.

179
00:09:29,160 --> 00:09:33,240
Speaker 5: Yeah, I think it really depends on your specific enterprise.

180
00:09:33,880 --> 00:09:37,360
Speaker 5: In our case, I would really hong in on a

181
00:09:37,360 --> 00:09:40,520
Speaker 5: few areas for me in terms of pricing and that

182
00:09:40,720 --> 00:09:44,839
Speaker 5: visibility of demand, in terms of channel inventory and the

183
00:09:45,080 --> 00:09:49,040
Speaker 5: demand of one hundred and eighty markets and what I

184
00:09:49,080 --> 00:09:53,720
Speaker 5: would take to actually provide and work with my channel

185
00:09:53,760 --> 00:09:57,920
Speaker 5: partners in terms of managing that inventory supply. There's a

186
00:09:57,960 --> 00:10:01,199
Speaker 5: lot of economics involved in that. We can optimize that

187
00:10:01,600 --> 00:10:04,800
Speaker 5: to run by AI rather than spread out in one

188
00:10:04,920 --> 00:10:07,240
Speaker 5: hundred and a markets. I think there could be a

189
00:10:07,240 --> 00:10:09,880
Speaker 5: lot of productivity gains from that that could be seen

190
00:10:10,280 --> 00:10:13,960
Speaker 5: from a dollar perspective rather than just generally in terms

191
00:10:14,000 --> 00:10:18,200
Speaker 5: of making very blanket general statements. Other areas I would

192
00:10:18,240 --> 00:10:22,480
Speaker 5: say that you could also increase productivity is pretty clear

193
00:10:22,480 --> 00:10:25,120
Speaker 5: in terms of data spend that you have today, A

194
00:10:25,120 --> 00:10:28,439
Speaker 5: lot of hedge fund friends and investor friends really talk

195
00:10:28,480 --> 00:10:31,760
Speaker 5: about in terms of what their savings could be in

196
00:10:31,840 --> 00:10:35,360
Speaker 5: terms of how much they're paying for certain data subscriptions.

197
00:10:35,800 --> 00:10:39,559
Speaker 5: And then, of course, as it relates to IR functions

198
00:10:39,640 --> 00:10:41,440
Speaker 5: to M and A functions in terms of how you

199
00:10:41,840 --> 00:10:45,840
Speaker 5: increase productivity, those functions tend to be smaller in headcount,

200
00:10:45,920 --> 00:10:48,840
Speaker 5: so it's really less about the heckcount reduction, but probably

201
00:10:48,880 --> 00:10:52,320
Speaker 5: about the increased productivity. And therefore, are you making the

202
00:10:52,400 --> 00:10:57,440
Speaker 5: right decision because those are small headcount high impact areas, right,

203
00:10:57,559 --> 00:11:00,240
Speaker 5: you could have a higher stock price and therefore your

204
00:11:00,240 --> 00:11:02,760
Speaker 5: cost of capital becomes much more efficient, and you're financing

205
00:11:03,440 --> 00:11:06,120
Speaker 5: even M and A. If you have better analysis, you're

206
00:11:06,160 --> 00:11:08,600
Speaker 5: optimizing in terms of how you're returning in terms of

207
00:11:08,640 --> 00:11:11,720
Speaker 5: that capital deployment. So from those perspectives, I think it

208
00:11:11,760 --> 00:11:14,400
Speaker 5: works out. And then of course, areas like marketing where

209
00:11:14,400 --> 00:11:18,280
Speaker 5: we're spending a lot out there in terms of you know,

210
00:11:18,440 --> 00:11:21,840
Speaker 5: video and other production generation, those actually can be done

211
00:11:21,840 --> 00:11:24,040
Speaker 5: by AI today, so we really need to look at that.

212
00:11:24,080 --> 00:11:27,160
Speaker 5: So that's just a few examples, but I think there

213
00:11:27,240 --> 00:11:29,640
Speaker 5: must be hundreds and hundreds and we're in the process

214
00:11:29,640 --> 00:11:30,959
Speaker 5: of trying to sort through that.

215
00:11:46,679 --> 00:11:49,679
Speaker 4: Let's go back to the theoretical engineer who spends one

216
00:11:49,720 --> 00:11:54,679
Speaker 4: hundred million dollars on tokens in a month, because you know,

217
00:11:54,720 --> 00:11:57,560
Speaker 4: it seems to me like you're going to have someone

218
00:11:57,640 --> 00:12:02,280
Speaker 4: who actually can with their you know, intuitions and skills

219
00:12:02,320 --> 00:12:06,800
Speaker 4: with AI et cetera, actually deliver ROI, even if the

220
00:12:06,920 --> 00:12:09,679
Speaker 4: nominal amount they spend is nose bleeding. And then other

221
00:12:09,679 --> 00:12:11,560
Speaker 4: people are just gonna waste a lot of money. And

222
00:12:11,600 --> 00:12:14,960
Speaker 4: I guarantee I know a lot about wasting money on AI,

223
00:12:15,520 --> 00:12:18,320
Speaker 4: and I guarantee that it's very easy to develop that

224
00:12:18,960 --> 00:12:22,400
Speaker 4: psychosis whereby you believe you're doing some incredible things that

225
00:12:22,480 --> 00:12:25,720
Speaker 4: feel like magic and productivity, and it turns out that

226
00:12:25,760 --> 00:12:28,839
Speaker 4: you really haven't built anything, et cetera. Like, it's very

227
00:12:28,840 --> 00:12:33,080
Speaker 4: easy to delude yourself into thinking you're being productive with

228
00:12:33,240 --> 00:12:37,760
Speaker 4: one's AI use. What I'm curious about, specifically is how

229
00:12:37,800 --> 00:12:43,000
Speaker 4: do you even go about, say, identifying the employees or

230
00:12:43,040 --> 00:12:47,840
Speaker 4: the teams who might really be people who you really

231
00:12:47,880 --> 00:12:51,000
Speaker 4: should give a long leash to with their spend How

232
00:12:51,040 --> 00:12:54,840
Speaker 4: do you even like identify who those are that should Yes,

233
00:12:54,960 --> 00:12:58,800
Speaker 4: let's not be too aggressive in their consumption, in capping their.

234
00:12:58,640 --> 00:13:01,960
Speaker 2: Consumption, don't give your tokens to a joe with AI

235
00:13:02,040 --> 00:13:02,600
Speaker 2: side vis.

236
00:13:02,920 --> 00:13:05,760
Speaker 4: How do you avoid giving all of the more, giving

237
00:13:05,840 --> 00:13:09,200
Speaker 4: me the uncut budget and finding that person who actually

238
00:13:09,200 --> 00:13:11,520
Speaker 4: deserves a very liberal spending budget.

239
00:13:11,679 --> 00:13:14,319
Speaker 5: Yeah, and you know, these are great questions that are

240
00:13:14,320 --> 00:13:17,120
Speaker 5: so relevant, particularly in a large enterprise like us, in

241
00:13:17,160 --> 00:13:20,280
Speaker 5: a global for a global two hundred company, right in

242
00:13:20,320 --> 00:13:23,560
Speaker 5: one hundred and eighty markets, eighty plus billion in terms

243
00:13:23,600 --> 00:13:26,800
Speaker 5: of revenues. We have massive scales and so many tens

244
00:13:26,800 --> 00:13:30,200
Speaker 5: and thousands of employees. So it's really difficult from a

245
00:13:30,240 --> 00:13:35,400
Speaker 5: cfo's perspective to micromanage down to the person's allocation. So

246
00:13:35,440 --> 00:13:38,080
Speaker 5: you really have to be able to say, first of all,

247
00:13:38,480 --> 00:13:43,200
Speaker 5: what are the agents in each domain capable of doing today? Right?

248
00:13:43,520 --> 00:13:47,400
Speaker 5: And how what is that productivity and also capability, and

249
00:13:47,440 --> 00:13:50,840
Speaker 5: then you can say, well, what specifically can it do today?

250
00:13:51,200 --> 00:13:53,599
Speaker 5: And then let's get the dollars at least from the

251
00:13:53,640 --> 00:13:58,360
Speaker 5: CFO perspective, let's then eliminate the dollar spend in those

252
00:13:58,400 --> 00:14:01,240
Speaker 5: areas and really force the AI. I have come to

253
00:14:01,280 --> 00:14:05,200
Speaker 5: the probably the conclusion that to really effectively drive things,

254
00:14:05,600 --> 00:14:09,240
Speaker 5: I need to then force discipline or starve certain budgets

255
00:14:09,480 --> 00:14:13,520
Speaker 5: to then allocate, because that really changes behavior because if

256
00:14:13,559 --> 00:14:16,840
Speaker 5: you continue to allocate that budget, they will continue to

257
00:14:16,960 --> 00:14:19,600
Speaker 5: use it and go on the old behavior. But I

258
00:14:19,600 --> 00:14:23,200
Speaker 5: think human beings are really adaptable, and I think we

259
00:14:23,360 --> 00:14:26,760
Speaker 5: tend to be good at survival, right, And so if

260
00:14:26,800 --> 00:14:29,880
Speaker 5: you force a certain situation, I think people, particularly with

261
00:14:29,920 --> 00:14:32,320
Speaker 5: the functions of AI today, people will be able to

262
00:14:32,680 --> 00:14:35,040
Speaker 5: then really try to use it because they don't have

263
00:14:35,080 --> 00:14:37,200
Speaker 5: the budget for the alternative. I think that's really from

264
00:14:37,200 --> 00:14:37,960
Speaker 5: a cfo's place.

265
00:14:38,000 --> 00:14:39,840
Speaker 4: Yeah, yeah, well, let me press on it. I find

266
00:14:39,880 --> 00:14:42,720
Speaker 4: this to be very interesting and maybe it even gets

267
00:14:42,720 --> 00:14:48,040
Speaker 4: to a certain philosophical split perhaps among CFOs, because I think, okay,

268
00:14:48,040 --> 00:14:50,920
Speaker 4: there's one school of thought would say, okay, you keep

269
00:14:50,920 --> 00:14:54,920
Speaker 4: these very tight constraints and then you see who can

270
00:14:55,440 --> 00:14:57,960
Speaker 4: do the most with the constraints, and then you derive

271
00:14:58,080 --> 00:15:00,920
Speaker 4: valuable information that comes back into your office that that

272
00:15:01,160 --> 00:15:03,080
Speaker 4: is valuable data there. And then there seems to be

273
00:15:03,120 --> 00:15:07,520
Speaker 4: another school, which is, let everyone have really high caps.

274
00:15:07,560 --> 00:15:10,520
Speaker 4: And we read these stories of American tech companies about

275
00:15:10,560 --> 00:15:14,480
Speaker 4: the internal quasi tooken maxing competition, and I think that

276
00:15:14,560 --> 00:15:17,240
Speaker 4: school of thought would say, look, let the employees go

277
00:15:17,360 --> 00:15:20,360
Speaker 4: to town and that is the way that we discover

278
00:15:20,920 --> 00:15:24,680
Speaker 4: who can actually use these tools productively, et cetera. But

279
00:15:24,720 --> 00:15:27,120
Speaker 4: would you say that that is actually sort of a

280
00:15:27,200 --> 00:15:30,920
Speaker 4: philosophical split in CFOs of tech companies in the early

281
00:15:31,000 --> 00:15:33,280
Speaker 4: years figuring out this optimal path.

282
00:15:33,640 --> 00:15:38,400
Speaker 5: And I think that that decision exists within the same company.

283
00:15:38,720 --> 00:15:40,720
Speaker 5: But it could be two paths that you just described,

284
00:15:41,120 --> 00:15:44,360
Speaker 5: because for example, where we really it's a face of

285
00:15:44,400 --> 00:15:47,840
Speaker 5: innovation today in terms of where the market's going, and

286
00:15:47,880 --> 00:15:50,440
Speaker 5: so we really need to put more dollars in terms

287
00:15:50,480 --> 00:15:52,520
Speaker 5: of innovation. And you can see that in terms of

288
00:15:52,520 --> 00:15:55,920
Speaker 5: our spend so where we really and that's the emphasis

289
00:15:55,920 --> 00:15:59,440
Speaker 5: from our chairman and CEO as well, it's really around

290
00:15:59,840 --> 00:16:03,560
Speaker 5: making sure that we take advantage of this period. We're

291
00:16:03,640 --> 00:16:06,040
Speaker 5: calling this the AI decade within the NOVO, and we're

292
00:16:06,040 --> 00:16:08,280
Speaker 5: in the first year of that journey. And I think

293
00:16:08,360 --> 00:16:11,480
Speaker 5: this is an opportune time for us because the way

294
00:16:11,480 --> 00:16:14,960
Speaker 5: we're positioned today, where we have global manufacturing in a

295
00:16:15,160 --> 00:16:19,000
Speaker 5: fairly fragmented world, where we have a lot of data

296
00:16:19,040 --> 00:16:21,760
Speaker 5: and production and supply chain concerns, where we are able

297
00:16:21,840 --> 00:16:26,000
Speaker 5: to be local and be able to supply that on

298
00:16:26,040 --> 00:16:30,240
Speaker 5: the local basis and comply to each region's needs, sustainability

299
00:16:30,280 --> 00:16:34,040
Speaker 5: needs for certain regions, particularly like EU. One hundred percent

300
00:16:34,040 --> 00:16:37,200
Speaker 5: of our laptops that you like is actually in recyclable

301
00:16:37,400 --> 00:16:40,640
Speaker 5: box paper, and also within the laptops a lot of

302
00:16:40,640 --> 00:16:44,280
Speaker 5: them are also recyclable materials as well, and we also

303
00:16:44,320 --> 00:16:47,880
Speaker 5: have a refurbished effort given the components constraints today. So

304
00:16:47,920 --> 00:16:50,520
Speaker 5: I think from that perspective, we really want to spend

305
00:16:50,600 --> 00:16:54,360
Speaker 5: dollar on innovation now. As you say, then there's another pocket.

306
00:16:54,720 --> 00:16:58,320
Speaker 5: Should the likes of even within finance, be able to

307
00:16:58,800 --> 00:17:02,040
Speaker 5: have an unlimited budget on AI, I think you could

308
00:17:02,040 --> 00:17:05,119
Speaker 5: probably be sent more sensible in that and say, well,

309
00:17:05,160 --> 00:17:09,520
Speaker 5: specifically within certain analysis and FP and A or accounting,

310
00:17:09,960 --> 00:17:11,959
Speaker 5: what was it that you actually used to spend? So,

311
00:17:12,000 --> 00:17:15,600
Speaker 5: for example, without naming a party, we outsource certain back

312
00:17:15,680 --> 00:17:18,680
Speaker 5: office function to a service provider and that's a pretty

313
00:17:18,720 --> 00:17:21,760
Speaker 5: costly one for us annually. Now in terms of that

314
00:17:21,800 --> 00:17:24,840
Speaker 5: capabilities that it's really a process flow. So should we

315
00:17:25,280 --> 00:17:29,879
Speaker 5: ultimately be able to automate everything and be able to

316
00:17:30,200 --> 00:17:33,320
Speaker 5: achieve savings and that now we are okay to continue

317
00:17:33,320 --> 00:17:37,359
Speaker 5: to outsource as long as that partner also innovates and

318
00:17:37,440 --> 00:17:39,480
Speaker 5: be able to share that savings with us as well.

319
00:17:39,880 --> 00:17:42,400
Speaker 5: So I think it really depends on where you want

320
00:17:42,440 --> 00:17:46,399
Speaker 5: to go, very specific in each area within your function

321
00:17:46,600 --> 00:17:49,360
Speaker 5: and your company. So in my function, there's so many

322
00:17:49,400 --> 00:17:55,600
Speaker 5: areas planning, tax, FP and a accounting, treasury M and

323
00:17:55,680 --> 00:17:58,960
Speaker 5: A right you name it, corporate finance. So within all

324
00:17:59,000 --> 00:18:02,040
Speaker 5: of these areas, where are we going to optimize be

325
00:18:02,119 --> 00:18:05,560
Speaker 5: more efficient productive? And I think that cost all or

326
00:18:05,680 --> 00:18:09,560
Speaker 5: sometimes to me is the return is significant. I was

327
00:18:09,600 --> 00:18:13,159
Speaker 5: just with the head of my tax this morning. I

328
00:18:13,200 --> 00:18:15,840
Speaker 5: was said, if we just pull together and drive the

329
00:18:15,880 --> 00:18:20,040
Speaker 5: next year's efforts, the tax savings or optimization right that

330
00:18:20,119 --> 00:18:22,359
Speaker 5: we do, we'll more than pay for a little small

331
00:18:22,400 --> 00:18:25,360
Speaker 5: trip of twenty something people together. That's very very much

332
00:18:25,400 --> 00:18:27,879
Speaker 5: well spent. But I think that's just a philosophy. It's

333
00:18:27,960 --> 00:18:31,080
Speaker 5: less AI driven, but it's really the philosophy of where

334
00:18:31,160 --> 00:18:33,520
Speaker 5: that return is. And I think you allocated earlier, you

335
00:18:33,640 --> 00:18:36,480
Speaker 5: spend based on the ROI generation, and I think today

336
00:18:36,520 --> 00:18:40,480
Speaker 5: you have to recognize first the AI capabilities that you

337
00:18:40,560 --> 00:18:42,520
Speaker 5: can have. I think the other thing we haven't touched

338
00:18:42,560 --> 00:18:45,840
Speaker 5: upon is really the security within the enterprise of using AI, right.

339
00:18:45,880 --> 00:18:47,800
Speaker 5: I think that's a broad topic that we really should

340
00:18:47,840 --> 00:18:52,080
Speaker 5: touch on because today I think there is really concerns

341
00:18:52,240 --> 00:18:56,840
Speaker 5: about what we don't know about using external AI within

342
00:18:56,880 --> 00:18:59,000
Speaker 5: our enterprise. So I think that's also a concern in

343
00:18:59,080 --> 00:19:00,800
Speaker 5: terms of how effective you can use AI.

344
00:19:01,080 --> 00:19:03,439
Speaker 2: We should definitely talk more about that. But you just

345
00:19:03,560 --> 00:19:07,080
Speaker 2: mentioned the word unlimited in the context of AI spend

346
00:19:07,160 --> 00:19:10,680
Speaker 2: and coming from the US, this is something that we're

347
00:19:10,720 --> 00:19:14,359
Speaker 2: really seeing and I guess living through at the moment.

348
00:19:14,480 --> 00:19:17,600
Speaker 2: Like people talk about the total addressable market being like

349
00:19:17,800 --> 00:19:20,000
Speaker 2: basically infinite at this point, if you look at the

350
00:19:20,040 --> 00:19:24,000
Speaker 2: SpaceX IPO, like the total addressable market is now the

351
00:19:24,160 --> 00:19:27,399
Speaker 2: entire universe, right. I guess my question is how do

352
00:19:27,440 --> 00:19:30,679
Speaker 2: you compete against those types of companies who seem to

353
00:19:30,680 --> 00:19:33,600
Speaker 2: have investors throwing money at them. I know your stock

354
00:19:33,640 --> 00:19:37,000
Speaker 2: price has gone up quite a lot, so that obviously helps.

355
00:19:37,280 --> 00:19:38,040
Speaker 3: And then a more.

356
00:19:37,880 --> 00:19:43,600
Speaker 2: Broader question, what are the key differences between how Chinese

357
00:19:43,640 --> 00:19:48,080
Speaker 2: companies are approaching AI versus how US companies are approaching AI?

358
00:19:48,640 --> 00:19:51,880
Speaker 5: Sounds good. I think there's two very distinct questions there.

359
00:19:52,400 --> 00:19:56,080
Speaker 5: The first one really with respect to Lenovo's position in

360
00:19:56,119 --> 00:19:59,600
Speaker 5: the entire tech ecosystem. I was in Davos earlier this year,

361
00:19:59,600 --> 00:20:02,680
Speaker 5: and when I can bag was particularly affected by Mark

362
00:20:02,720 --> 00:20:06,879
Speaker 5: Kartney's speech and really talking about the superpowers in the

363
00:20:06,880 --> 00:20:09,520
Speaker 5: middle powers, and of course within the middle power their superpowers.

364
00:20:09,720 --> 00:20:12,919
Speaker 5: I actually use that within my own management committee in

365
00:20:13,000 --> 00:20:17,080
Speaker 5: terms of referring our own position within the tech stack.

366
00:20:17,760 --> 00:20:19,879
Speaker 5: And I think as big as we are as a

367
00:20:19,920 --> 00:20:23,240
Speaker 5: middle power, I think we're definitely not a superpower. We

368
00:20:23,320 --> 00:20:26,320
Speaker 5: have to recognize that. But I think from our perspective

369
00:20:26,320 --> 00:20:29,719
Speaker 5: being the number one device PC manufacturer in the world,

370
00:20:30,200 --> 00:20:34,840
Speaker 5: having a strong ecosystem and the best broadest portfolio set

371
00:20:35,040 --> 00:20:40,200
Speaker 5: right across PC, tablets and mobile phones, and great partners

372
00:20:40,280 --> 00:20:44,000
Speaker 5: to Microsoft and Googles of the world in terms of

373
00:20:44,040 --> 00:20:46,880
Speaker 5: that type of operating system, and of course the chip

374
00:20:46,880 --> 00:20:51,000
Speaker 5: suppliers like Nvidia, Intel, AMD and Qualcom and of course

375
00:20:51,040 --> 00:20:54,160
Speaker 5: now aren't based chips as well, and also the Chinese

376
00:20:54,040 --> 00:20:57,639
Speaker 5: tech stack that's also coming up. I think from that perspective,

377
00:20:58,200 --> 00:21:01,160
Speaker 5: we are there to really enable and I think prior

378
00:21:01,280 --> 00:21:03,919
Speaker 5: to this stock run, we're really not recognized by the

379
00:21:03,920 --> 00:21:07,240
Speaker 5: market in terms of our ability to be partners to

380
00:21:07,480 --> 00:21:12,880
Speaker 5: two thousand suppliers in the tech ecosystem driving almost two

381
00:21:12,920 --> 00:21:15,520
Speaker 5: trillion dollars plus of spend and we're right in the

382
00:21:15,520 --> 00:21:18,920
Speaker 5: middle of it and distributing this across the board, even

383
00:21:19,520 --> 00:21:21,720
Speaker 5: in one hundred and eighty markets. So I think from

384
00:21:21,720 --> 00:21:24,479
Speaker 5: that perspective we really needed to be recognized. But at

385
00:21:24,520 --> 00:21:26,639
Speaker 5: the same time, we're very humble, and I think our

386
00:21:26,680 --> 00:21:28,840
Speaker 5: chairman really used that word in terms of staying humble,

387
00:21:29,280 --> 00:21:31,520
Speaker 5: and I think we are because we're also micro. From

388
00:21:31,560 --> 00:21:34,960
Speaker 5: that perspective, most of the profits have been towards the

389
00:21:35,200 --> 00:21:37,800
Speaker 5: IC companies, the OS companies. If you look at the

390
00:21:37,800 --> 00:21:41,199
Speaker 5: PC revolution and the tech stack, the OS company is

391
00:21:41,200 --> 00:21:45,600
Speaker 5: now trillion dollars. The chip companies are now crossing trillion dollars,

392
00:21:45,600 --> 00:21:48,600
Speaker 5: but have before been hundreds of billions of dollars, and

393
00:21:48,680 --> 00:21:52,560
Speaker 5: we only most recently became around I think thirty to

394
00:21:52,600 --> 00:21:56,520
Speaker 5: forty billion dollars, but before that sub twenty billion. So

395
00:21:56,560 --> 00:21:58,720
Speaker 5: I think that's the value chain that you're in. But

396
00:21:59,320 --> 00:22:02,040
Speaker 5: I think we probably did not optimize it as well,

397
00:22:02,280 --> 00:22:04,880
Speaker 5: and I think there's an opportunity time for the device

398
00:22:04,920 --> 00:22:09,800
Speaker 5: makers today to optimize. And of course, distribution is very critical.

399
00:22:10,200 --> 00:22:14,439
Speaker 5: Supply chain is critical. Ability to aggregate supply chain, to

400
00:22:14,520 --> 00:22:19,000
Speaker 5: manufacture at scale, at cost efficiently and deliver it to

401
00:22:19,080 --> 00:22:22,359
Speaker 5: your customers, to be able to service it, to provide

402
00:22:22,400 --> 00:22:25,800
Speaker 5: a security and trust and the aftermarket service support. That

403
00:22:25,920 --> 00:22:28,840
Speaker 5: must have a value beyond what we're being recognized by

404
00:22:28,840 --> 00:22:31,760
Speaker 5: the market. I mean, I absolutely believe in that, and

405
00:22:31,800 --> 00:22:33,520
Speaker 5: I think we need to drive that in terms of

406
00:22:33,520 --> 00:22:38,600
Speaker 5: financial returns from that perspective, and AI wave definitely creates

407
00:22:38,600 --> 00:22:41,240
Speaker 5: the opportunity for the likes of Lenovo, and I think

408
00:22:41,240 --> 00:22:43,920
Speaker 5: we're the best position in this because there's no one

409
00:22:44,080 --> 00:22:47,000
Speaker 5: company like us from the end to end portfolio and

410
00:22:47,040 --> 00:22:50,240
Speaker 5: the global manufacturing, So you know, I think that's a

411
00:22:50,560 --> 00:22:53,280
Speaker 5: very important point. And in terms of how the US

412
00:22:53,320 --> 00:22:57,400
Speaker 5: and Chinese tech stact today are differentiated, or AI companies

413
00:22:57,440 --> 00:23:00,840
Speaker 5: are differentiated today, obviously because of the restrictions. I know

414
00:23:00,920 --> 00:23:04,160
Speaker 5: that probably the most efficient compute today is probably from

415
00:23:04,200 --> 00:23:08,000
Speaker 5: the US chip company. So obviously the chip availability is

416
00:23:08,040 --> 00:23:11,239
Speaker 5: something that the Chinese AI companies probably have to work with,

417
00:23:11,600 --> 00:23:15,159
Speaker 5: but in that way they probably managed to produce at

418
00:23:15,280 --> 00:23:18,560
Speaker 5: much lower costs and very efficiently as well. So you

419
00:23:18,680 --> 00:23:21,760
Speaker 5: see today a lot of chatter about the cost per

420
00:23:21,800 --> 00:23:24,960
Speaker 5: token generation from the like of a deep seak being,

421
00:23:25,320 --> 00:23:27,199
Speaker 5: and I don't have official numbers, so this is just

422
00:23:27,280 --> 00:23:30,679
Speaker 5: quoting what I've heard, like one fiftieth of those in

423
00:23:30,680 --> 00:23:33,720
Speaker 5: the US. So I think from that perspective, right as

424
00:23:33,760 --> 00:23:37,400
Speaker 5: I said, human beings or companies. And if you really

425
00:23:37,400 --> 00:23:42,120
Speaker 5: think about the Darwinism in economic theories, the free markets

426
00:23:42,119 --> 00:23:45,520
Speaker 5: create the hot, tapest competitors and the most healthy competitors.

427
00:23:45,800 --> 00:23:48,679
Speaker 5: So I think in China, for those that compete in

428
00:23:48,760 --> 00:23:52,320
Speaker 5: such a there's a word called involution in China. Involution

429
00:23:53,280 --> 00:23:56,600
Speaker 5: it only happens in China market. So if they can

430
00:23:56,800 --> 00:24:01,240
Speaker 5: survive within their cutthrow market under the constraints of the

431
00:24:01,320 --> 00:24:05,399
Speaker 5: chip supply, they are pretty anywhere, they are pretty strong.

432
00:24:05,640 --> 00:24:08,200
Speaker 5: So I think that would be one difference. I would say, Yeah,

433
00:24:08,480 --> 00:24:08,960
Speaker 5: I want.

434
00:24:08,880 --> 00:24:12,840
Speaker 4: To talk about competition. You mentioned the deep and broad

435
00:24:12,920 --> 00:24:16,359
Speaker 4: relationship you have with all these suppliers across the supply chain,

436
00:24:16,480 --> 00:24:18,560
Speaker 4: and when I think about like when I think about

437
00:24:18,640 --> 00:24:23,320
Speaker 4: competition within the server space specifically, I feel like there

438
00:24:23,320 --> 00:24:27,280
Speaker 4: are two dimensions through which you could win. And one

439
00:24:27,320 --> 00:24:30,159
Speaker 4: of them is, of course, like, Okay, you have total

440
00:24:30,200 --> 00:24:35,600
Speaker 4: supply chain mastery. You can keep inventories low because you're

441
00:24:35,640 --> 00:24:37,920
Speaker 4: just in time delivery of everything, and you just have

442
00:24:38,000 --> 00:24:41,760
Speaker 4: this beautifully efficient supply chain with the world's best parts

443
00:24:41,760 --> 00:24:44,120
Speaker 4: makers all around the world. And then there's another way

444
00:24:44,119 --> 00:24:47,720
Speaker 4: you could win, which is your server is just more

445
00:24:47,800 --> 00:24:51,440
Speaker 4: performance than a competitors server et cetera. And we know

446
00:24:51,520 --> 00:24:55,600
Speaker 4: other competitors like like Adele or super Micro, et cetera.

447
00:24:56,240 --> 00:24:59,919
Speaker 4: Those are both strike me as dimensions upon which a

448
00:25:00,040 --> 00:25:04,639
Speaker 4: company could win the server market. Which to you is

449
00:25:04,720 --> 00:25:09,040
Speaker 4: more important? Is it the sort of supply chain excellence

450
00:25:09,520 --> 00:25:12,240
Speaker 4: or is it the quality of the server itself.

451
00:25:13,240 --> 00:25:15,440
Speaker 5: I think it's not an either or question. It's fully

452
00:25:15,440 --> 00:25:20,960
Speaker 5: integrated tax STAC today the value to a customer or

453
00:25:21,000 --> 00:25:23,600
Speaker 5: a partner, and I think there are less customers today

454
00:25:23,600 --> 00:25:26,280
Speaker 5: than there are really a partner because their architecture is

455
00:25:26,320 --> 00:25:29,120
Speaker 5: not driven and the spend is a multi yeer spent,

456
00:25:29,720 --> 00:25:32,480
Speaker 5: so they're coming to you for the multi yeer plan,

457
00:25:32,680 --> 00:25:35,480
Speaker 5: not just a single purchase. So we take that attitude

458
00:25:35,560 --> 00:25:37,520
Speaker 5: and we need to have that attitude with our suppliers

459
00:25:37,560 --> 00:25:40,440
Speaker 5: as well. We have invested well over ten years in

460
00:25:40,480 --> 00:25:42,720
Speaker 5: these factories. We have thirty factories of World twelve and

461
00:25:42,920 --> 00:25:45,919
Speaker 5: each region of the world. We can produce in Europe

462
00:25:45,960 --> 00:25:49,679
Speaker 5: and Hungary for the European market or the US market.

463
00:25:50,240 --> 00:25:52,800
Speaker 5: We can produce in Mexico in the US for the

464
00:25:52,880 --> 00:25:56,480
Speaker 5: US market, we can produce in Asia for the Asian market,

465
00:25:56,520 --> 00:25:58,520
Speaker 5: and now we're building that factory in the Middle in

466
00:25:58,640 --> 00:26:02,439
Speaker 5: Riad servers also for the Middle East market, and of

467
00:26:02,480 --> 00:26:05,200
Speaker 5: course China market. We always have factories there as well

468
00:26:05,200 --> 00:26:07,720
Speaker 5: as for the China market and other markets. So I

469
00:26:07,720 --> 00:26:10,400
Speaker 5: think from that perspective we are really truly global. From

470
00:26:10,440 --> 00:26:13,639
Speaker 5: that perspective today, the bottlenecks are not only in getting supply.

471
00:26:14,160 --> 00:26:17,639
Speaker 5: After you supply, can you then put it together? Do

472
00:26:17,720 --> 00:26:21,720
Speaker 5: you have the production capability to be able to produce?

473
00:26:22,240 --> 00:26:25,800
Speaker 5: Then do you have the capability to test what is

474
00:26:25,840 --> 00:26:29,119
Speaker 5: being produced? So we at Lenovo provide that end to

475
00:26:29,240 --> 00:26:33,159
Speaker 5: end capability to a customer and that ensures them to

476
00:26:33,359 --> 00:26:36,920
Speaker 5: have the safety and one stop shop. In fact, our

477
00:26:37,080 --> 00:26:40,840
Speaker 5: SSG can actually also so the Solutions and Services Group

478
00:26:40,920 --> 00:26:45,160
Speaker 5: at Lenovo can also build data centers, so different from

479
00:26:45,240 --> 00:26:49,000
Speaker 5: our OEM like competitors, they can actually build data centers.

480
00:26:49,040 --> 00:26:52,720
Speaker 5: So today the customers can hyperscallet can actually come to

481
00:26:52,800 --> 00:26:56,280
Speaker 5: Lenovo with a plan to say I have a plan

482
00:26:56,359 --> 00:27:00,000
Speaker 5: to have this site. So people who are not even hyperscalus,

483
00:27:00,080 --> 00:27:03,439
Speaker 5: people who are land owners and they have power in

484
00:27:03,520 --> 00:27:07,720
Speaker 5: specific areas, they come to Lenovo and they said, help

485
00:27:07,800 --> 00:27:10,040
Speaker 5: me build the data center. Because we also have a

486
00:27:10,080 --> 00:27:13,120
Speaker 5: modular solution that can build the data service within nine

487
00:27:13,160 --> 00:27:17,679
Speaker 5: months in a depending on what you already have available

488
00:27:17,800 --> 00:27:20,080
Speaker 5: in terms of infrastructure, we can do it as fast

489
00:27:20,080 --> 00:27:23,040
Speaker 5: as six months. So that is very very fast and

490
00:27:23,080 --> 00:27:26,920
Speaker 5: that gets the time to market and enable revenues immediately

491
00:27:26,960 --> 00:27:31,240
Speaker 5: for our partners. And then in terms of that total infrastructure,

492
00:27:31,800 --> 00:27:35,240
Speaker 5: plus if they can have GPU compute in the local market,

493
00:27:35,720 --> 00:27:38,840
Speaker 5: then of course Lenovo can also provide that, particularly with

494
00:27:38,880 --> 00:27:42,880
Speaker 5: our eleven thousand rack liquid cooling capability to be able

495
00:27:42,920 --> 00:27:46,320
Speaker 5: to service a GPU compute today, right and we're expanding

496
00:27:46,359 --> 00:27:50,000
Speaker 5: on that capability. So today we're the most end to

497
00:27:50,240 --> 00:27:52,800
Speaker 5: end and relevant partner really from that domain. So it's

498
00:27:52,840 --> 00:27:56,479
Speaker 5: really not from a perspective of really having the supply

499
00:27:57,640 --> 00:28:01,200
Speaker 5: and having the technology. Technology always import but that end

500
00:28:01,200 --> 00:28:03,960
Speaker 5: to end. Look, if you think about how people are

501
00:28:04,119 --> 00:28:08,080
Speaker 5: affecting and enabling AI compute today, they need everything and

502
00:28:08,119 --> 00:28:10,680
Speaker 5: of course, as Tracy said earlier, I think capital is

503
00:28:10,680 --> 00:28:13,480
Speaker 5: also important part of that. And you know, it does

504
00:28:13,520 --> 00:28:16,960
Speaker 5: help that that market is starting to recognize Renoble in

505
00:28:17,000 --> 00:28:17,760
Speaker 5: terms of our value.

506
00:28:17,800 --> 00:28:36,040
Speaker 4: Add Tracy, I'm thinking about the difference between you know,

507
00:28:36,119 --> 00:28:38,959
Speaker 4: the server market today and say, like the PC market

508
00:28:39,040 --> 00:28:42,760
Speaker 4: of the nineties, and I think part of the reason

509
00:28:43,320 --> 00:28:46,240
Speaker 4: we talk about and remember the red button on the

510
00:28:46,320 --> 00:28:49,880
Speaker 4: laptop is simply because the companies were all unable to

511
00:28:49,880 --> 00:28:52,480
Speaker 4: differentiate themselves at all product wise, and it was like,

512
00:28:52,880 --> 00:28:56,440
Speaker 4: truly the PC market kind of killed itself because it

513
00:28:56,520 --> 00:28:59,440
Speaker 4: was just so unbelievably commodified. No one really knew the

514
00:28:59,440 --> 00:29:02,880
Speaker 4: difference or carred between the difference between an adult box

515
00:29:02,920 --> 00:29:06,640
Speaker 4: and a compact box and either the packard box and whatever.

516
00:29:06,880 --> 00:29:10,120
Speaker 4: But it does sound listening to a Wisden that like

517
00:29:10,440 --> 00:29:13,360
Speaker 4: you really can the partner, not the customer. You can

518
00:29:13,480 --> 00:29:15,520
Speaker 4: really differentiate your offering.

519
00:29:16,000 --> 00:29:19,160
Speaker 3: Yeah, I remember that environment. So I remember like.

520
00:29:19,680 --> 00:29:21,560
Speaker 2: You had to go to Best Buy, right, and you

521
00:29:21,600 --> 00:29:23,800
Speaker 2: had to basically get a salesman to tell you what

522
00:29:23,920 --> 00:29:26,120
Speaker 2: the difference was between all these different.

523
00:29:25,800 --> 00:29:26,760
Speaker 4: Things different Yeah.

524
00:29:26,840 --> 00:29:30,320
Speaker 2: Yeah, Okay, since we're on the topic of supply chains

525
00:29:30,480 --> 00:29:33,760
Speaker 2: and you mentioned data centers as well, give us some

526
00:29:33,840 --> 00:29:37,200
Speaker 2: color on what it's like to try to get I

527
00:29:37,200 --> 00:29:39,800
Speaker 2: guess when I think about constraints on the AI build out,

528
00:29:39,880 --> 00:29:46,920
Speaker 2: I think about memory, GPUs, CPUs, chips and transformers. Now

529
00:29:46,960 --> 00:29:49,280
Speaker 2: for data centers at least in the US.

530
00:29:49,520 --> 00:29:51,280
Speaker 4: How difficult until connectors?

531
00:29:51,320 --> 00:29:52,120
Speaker 3: Oh yes, thank you.

532
00:29:52,440 --> 00:29:55,560
Speaker 2: How difficult is that at the moment getting those key things?

533
00:29:56,080 --> 00:29:59,720
Speaker 5: Yeah, the market is very robust because the demands very robust,

534
00:29:59,720 --> 00:30:02,840
Speaker 5: and as we're sitting in Hong Kong today in the

535
00:30:02,880 --> 00:30:07,800
Speaker 5: Bloomberg offices. But really there are some really exciting IPOs

536
00:30:07,840 --> 00:30:11,120
Speaker 5: coming right and also capital races. We're seeing capital raises

537
00:30:11,560 --> 00:30:15,200
Speaker 5: at such significant levels, and that's really all going back

538
00:30:15,240 --> 00:30:19,520
Speaker 5: to AI infrastructure spent. So really, from the perspective of

539
00:30:19,640 --> 00:30:24,640
Speaker 5: that component shortage or demand driven challenge on the supply

540
00:30:24,720 --> 00:30:28,160
Speaker 5: chain will continue probably for at least two to three years,

541
00:30:28,840 --> 00:30:32,880
Speaker 5: but we know that the supply chain is starting to invest.

542
00:30:33,080 --> 00:30:37,120
Speaker 5: But in terms of some of the memory specific areas,

543
00:30:37,200 --> 00:30:39,080
Speaker 5: it does take almost two to three years for a

544
00:30:39,120 --> 00:30:41,920
Speaker 5: new fact to come online. So I think from that perspective,

545
00:30:42,000 --> 00:30:44,800
Speaker 5: and I think you're seeing bottlenecks across the board, right

546
00:30:44,880 --> 00:30:48,719
Speaker 5: And as I said earlier, land power, not just the transformers,

547
00:30:48,720 --> 00:30:51,280
Speaker 5: but actual power from the grid. And that's why we're

548
00:30:51,280 --> 00:30:54,040
Speaker 5: in the Middle East to work with the Saudi government

549
00:30:54,160 --> 00:30:59,400
Speaker 5: on sustainable but also cheap solar and much available energy

550
00:30:59,400 --> 00:31:02,040
Speaker 5: in the local mind so it's very efficient there and

551
00:31:02,080 --> 00:31:04,720
Speaker 5: they have very low cost energy there that could actually

552
00:31:04,800 --> 00:31:07,880
Speaker 5: support AI infrastructure. From that perspective, so we really need

553
00:31:07,920 --> 00:31:12,640
Speaker 5: to optimize where specifically the power is around the world,

554
00:31:13,120 --> 00:31:15,480
Speaker 5: rather than just saying, hey, everything has to be generated

555
00:31:15,520 --> 00:31:18,360
Speaker 5: within an area and that area doesn't even have power.

556
00:31:18,680 --> 00:31:20,520
Speaker 5: So I think we really need to be able to

557
00:31:20,520 --> 00:31:22,800
Speaker 5: do that put the global resources together.

558
00:31:23,120 --> 00:31:25,280
Speaker 4: One of the things that we all learned during the

559
00:31:25,360 --> 00:31:29,080
Speaker 4: pandemic is that you know, and we all learned about

560
00:31:29,480 --> 00:31:32,880
Speaker 4: how supply chain disruptions work and the bull whip effect,

561
00:31:33,000 --> 00:31:35,960
Speaker 4: or at least you know, us lay people learned about

562
00:31:35,960 --> 00:31:38,360
Speaker 4: these concepts for the first time. Everyone we talked to.

563
00:31:38,960 --> 00:31:40,800
Speaker 4: We recently did an episode with one of the co

564
00:31:40,840 --> 00:31:43,200
Speaker 4: founders of court Weave, one of the neo clouds. You know,

565
00:31:43,240 --> 00:31:46,280
Speaker 4: everyone we talked to, we'll talk about multi year long

566
00:31:46,520 --> 00:31:49,480
Speaker 4: order books and backlogs, et cetera. And when I hear

567
00:31:49,560 --> 00:31:51,800
Speaker 4: that and I started thinking back, I was like, yes,

568
00:31:51,920 --> 00:31:55,160
Speaker 4: I believe you when you say that. But it could

569
00:31:55,200 --> 00:31:59,640
Speaker 4: also be a bunch of players double ordering and triple

570
00:31:59,760 --> 00:32:03,840
Speaker 4: order and trying to get ahead of themselves just because

571
00:32:03,840 --> 00:32:07,200
Speaker 4: they're all aware of the shortages, et cetera. In which

572
00:32:07,280 --> 00:32:11,280
Speaker 4: case it does seem like what looks right now like

573
00:32:11,360 --> 00:32:15,760
Speaker 4: a major backlog and constraints and endless demand could not

574
00:32:15,960 --> 00:32:19,560
Speaker 4: be as sustainable as people thought. Should we perhaps be

575
00:32:19,720 --> 00:32:23,040
Speaker 4: worried that some of these endless order books are just

576
00:32:23,120 --> 00:32:27,200
Speaker 4: in fact companies overordering because they see everyone else over ordering.

577
00:32:27,520 --> 00:32:30,360
Speaker 5: I think these are very sophisticated companies in terms of

578
00:32:30,400 --> 00:32:33,800
Speaker 5: the final off takers and so, but I think I

579
00:32:33,840 --> 00:32:36,880
Speaker 5: would point out to the fact that the duplication could

580
00:32:36,920 --> 00:32:40,720
Speaker 5: happen in the pipeline, but not necessarily in a backlog,

581
00:32:40,800 --> 00:32:43,840
Speaker 5: because I think if people have standard definitions of the

582
00:32:43,880 --> 00:32:48,720
Speaker 5: backlog by that time, it's actually already a sign signed money,

583
00:32:48,800 --> 00:32:51,520
Speaker 5: so I think they wouldn't double signed with people. Yeah,

584
00:32:51,560 --> 00:32:54,040
Speaker 5: and then of course revenue recognition is a totally different thing.

585
00:32:54,080 --> 00:32:56,120
Speaker 5: I think you have already shipped the actual product and

586
00:32:56,200 --> 00:32:59,280
Speaker 5: you expect to collect on your accounts receivables. So I

587
00:32:59,280 --> 00:33:02,120
Speaker 5: think from that person aspect of really looking at the

588
00:33:02,160 --> 00:33:05,280
Speaker 5: pipeline that is massive in the market, there is likely

589
00:33:05,720 --> 00:33:08,720
Speaker 5: duplication there and people are waiting to see who has

590
00:33:08,760 --> 00:33:12,280
Speaker 5: the supply and who can deliver to you, right. But overall,

591
00:33:12,640 --> 00:33:15,480
Speaker 5: I think the commitment that we see in the market,

592
00:33:15,520 --> 00:33:20,240
Speaker 5: particularly for data center space, for equipment for other things,

593
00:33:20,640 --> 00:33:23,080
Speaker 5: is really a multi year commitment. And I think people

594
00:33:23,160 --> 00:33:25,880
Speaker 5: see this as a critical infrastructure. They learn from the

595
00:33:25,920 --> 00:33:28,920
Speaker 5: Internet error in terms of the cloud service providers how

596
00:33:29,000 --> 00:33:31,360
Speaker 5: much data can be generated, and this is just going

597
00:33:31,400 --> 00:33:33,160
Speaker 5: to be so much more. I think it's going to

598
00:33:33,160 --> 00:33:35,720
Speaker 5: be exponentially more then what it is because it will

599
00:33:35,760 --> 00:33:39,000
Speaker 5: take the existing data plus generate more data and then

600
00:33:39,040 --> 00:33:42,600
Speaker 5: you need to enable that data, right, and that enablement

601
00:33:42,640 --> 00:33:44,480
Speaker 5: I think these additional compute.

602
00:33:44,760 --> 00:33:47,920
Speaker 2: Lenovo has software as well, right, and you have the

603
00:33:47,960 --> 00:33:49,520
Speaker 2: gaming business, is that right?

604
00:33:49,760 --> 00:33:52,760
Speaker 5: We have a we're the leading online games sorry, the

605
00:33:52,800 --> 00:33:54,800
Speaker 5: game PC company in the world.

606
00:33:54,880 --> 00:33:58,719
Speaker 2: Yeah, So what do you think about the SaaS apocalypse idea?

607
00:33:58,960 --> 00:34:01,400
Speaker 2: So the idea that with AI, we're going to be

608
00:34:01,440 --> 00:34:05,320
Speaker 2: able to replicate whatever application we want, whether it's Microsoft

609
00:34:05,360 --> 00:34:08,640
Speaker 2: Word or Excel or something like that. Would that eat

610
00:34:08,680 --> 00:34:11,440
Speaker 2: into your margins or because you own the IP, can

611
00:34:11,480 --> 00:34:13,080
Speaker 2: you just do it more cheaply?

612
00:34:13,680 --> 00:34:16,879
Speaker 5: No. I think we are not a SaaS company by

613
00:34:16,920 --> 00:34:20,920
Speaker 5: any means. But I think from the perspective of Lenovo

614
00:34:21,040 --> 00:34:23,600
Speaker 5: or enterprises, I think really as it relates to what

615
00:34:23,680 --> 00:34:26,879
Speaker 5: we touched upon earlier, your IT department, and I think

616
00:34:26,920 --> 00:34:29,759
Speaker 5: it actually works for a company like Lenovo because we

617
00:34:29,840 --> 00:34:34,320
Speaker 5: are not a software centric company. These tools now lower

618
00:34:34,400 --> 00:34:37,040
Speaker 5: the bar for us. So I think our engineers or

619
00:34:37,160 --> 00:34:40,400
Speaker 5: hiring engineers should be able to create certain apps that

620
00:34:40,440 --> 00:34:43,719
Speaker 5: should be available or applications. So I think from that

621
00:34:43,800 --> 00:34:48,120
Speaker 5: perspective there are certain elements of applications that would be

622
00:34:48,160 --> 00:34:51,080
Speaker 5: able to produce in house and that would lower the

623
00:34:51,120 --> 00:34:55,080
Speaker 5: bar because these tools are very powerful. But in terms

624
00:34:55,120 --> 00:34:59,560
Speaker 5: of the SaaS cop popularist, I think it's hardware for

625
00:34:59,600 --> 00:35:02,239
Speaker 5: me to say, but I think from that perspective, I

626
00:35:02,320 --> 00:35:05,960
Speaker 5: believe there's certain data layers that are absolutely essential because

627
00:35:06,080 --> 00:35:09,319
Speaker 5: enterprises almost cannot get away from this, and not nor

628
00:35:09,360 --> 00:35:11,760
Speaker 5: did they really want to. I think you just need

629
00:35:12,200 --> 00:35:14,840
Speaker 5: the middle layer to be able to bring the data

630
00:35:14,880 --> 00:35:17,520
Speaker 5: together and be able to connect to AI, and that's

631
00:35:17,680 --> 00:35:20,360
Speaker 5: one of the biggest bottlenecks for enterprises today, and whoever

632
00:35:20,400 --> 00:35:22,560
Speaker 5: can do that right would be able to be a

633
00:35:22,680 --> 00:35:26,440
Speaker 5: helpful But I think rather than just purely a software solution,

634
00:35:26,760 --> 00:35:29,319
Speaker 5: there's probably a consultant element to that as well to

635
00:35:29,400 --> 00:35:33,000
Speaker 5: help enterprise enable this function. But I think we're a

636
00:35:33,080 --> 00:35:34,960
Speaker 5: long journey from that. You know.

637
00:35:35,040 --> 00:35:38,720
Speaker 4: I feel sorry for gamers because even prior to AI

638
00:35:39,200 --> 00:35:43,239
Speaker 4: gobbling up some of the GPUs, right before that, it

639
00:35:43,280 --> 00:35:45,520
Speaker 4: was crypto, and I remember they used to complain that

640
00:35:45,560 --> 00:35:48,480
Speaker 4: the ethereum miners were buying up all of their in

641
00:35:48,600 --> 00:35:52,080
Speaker 4: video GPUs, et cetera. But I'm curious, like you know,

642
00:35:52,160 --> 00:35:54,440
Speaker 4: this is one another one of the big meta themes

643
00:35:54,480 --> 00:35:58,080
Speaker 4: of the time, like in your own gaming business or

644
00:35:58,080 --> 00:36:02,600
Speaker 4: in the gaming industry in general. Obviously, GPUs seems like

645
00:36:02,680 --> 00:36:05,359
Speaker 4: the highest marginal value you can get from them is

646
00:36:05,680 --> 00:36:12,320
Speaker 4: with artificial intelligence. Do you see this continuing phenomenon where

647
00:36:12,440 --> 00:36:15,120
Speaker 4: raw inputs of any sort that might have at one

648
00:36:15,239 --> 00:36:18,960
Speaker 4: time go to gamers, either because just the gamers get

649
00:36:18,960 --> 00:36:21,520
Speaker 4: priced out, or they no one builds them for gamers

650
00:36:21,760 --> 00:36:24,120
Speaker 4: because you can make so much more money selling them

651
00:36:24,200 --> 00:36:26,040
Speaker 4: into the AI ecosystem.

652
00:36:26,520 --> 00:36:29,839
Speaker 5: Well, we got a fantastic franchise in terms of as

653
00:36:29,840 --> 00:36:33,560
Speaker 5: you alluded to, the think pad business. It's an iconic

654
00:36:33,840 --> 00:36:37,279
Speaker 5: franchise and we love it, our customers love it. I've

655
00:36:37,280 --> 00:36:41,400
Speaker 5: actually visited the R and D facility that originated actually

656
00:36:41,400 --> 00:36:44,759
Speaker 5: from Japan. It's from the Yama Moto lapse of IBM,

657
00:36:45,239 --> 00:36:48,319
Speaker 5: so we kept that intact also with a lot of

658
00:36:48,320 --> 00:36:51,520
Speaker 5: the people from it. But I think the original inventor

659
00:36:51,560 --> 00:36:54,200
Speaker 5: actually retired some years some years ago, but there's a

660
00:36:54,239 --> 00:36:56,840
Speaker 5: long history that we're very proud of and we've actually

661
00:36:57,320 --> 00:37:01,239
Speaker 5: maintained that in terms of our consumer and also as

662
00:37:01,239 --> 00:37:04,360
Speaker 5: it relates specifically to the online business online game business,

663
00:37:04,440 --> 00:37:07,480
Speaker 5: which is our legion business, it's actually also very famous, right,

664
00:37:07,680 --> 00:37:10,520
Speaker 5: so it's the largest by volume, and people really like

665
00:37:10,600 --> 00:37:12,880
Speaker 5: Lenovo for the price performance that we can deliver to

666
00:37:12,920 --> 00:37:17,280
Speaker 5: them and gamers really appreciate that. So these are amazing

667
00:37:17,280 --> 00:37:19,640
Speaker 5: franchises that we want to be able to keep and

668
00:37:19,680 --> 00:37:22,240
Speaker 5: they are loyal fan base and were the largest market

669
00:37:22,280 --> 00:37:24,319
Speaker 5: share in the world, so we want to be able

670
00:37:24,360 --> 00:37:26,319
Speaker 5: to protect that. We actually, rather than protect that, we

671
00:37:26,360 --> 00:37:28,920
Speaker 5: want to grow that market share. So I think from

672
00:37:29,000 --> 00:37:32,560
Speaker 5: that perspective, we will have to balance between the device

673
00:37:33,160 --> 00:37:36,200
Speaker 5: side of our business as well as the infrastructure side

674
00:37:36,200 --> 00:37:37,680
Speaker 5: of the business, which at least they have a lot

675
00:37:37,760 --> 00:37:39,880
Speaker 5: of demand today. I think it just means that we

676
00:37:39,960 --> 00:37:42,200
Speaker 5: have to get more share of our supply.

677
00:37:43,080 --> 00:37:46,040
Speaker 2: Talk to us about your decision to, I guess not

678
00:37:46,280 --> 00:37:49,920
Speaker 2: tie yourself to any one AI platform, because I'm sure

679
00:37:49,920 --> 00:37:52,200
Speaker 2: there are some very very big tech companies out there

680
00:37:52,239 --> 00:37:55,480
Speaker 2: who would love to have exclusivity with Lenovo, and there

681
00:37:55,520 --> 00:37:59,440
Speaker 2: are also broader arguments out there that eventually we're going

682
00:37:59,480 --> 00:38:04,080
Speaker 2: to have one AI platform that emerges triumphant in the

683
00:38:04,160 --> 00:38:07,400
Speaker 2: same way that we saw Google takeover search in the

684
00:38:07,440 --> 00:38:11,000
Speaker 2: early two thousands. Why did you decide to, I guess

685
00:38:11,600 --> 00:38:13,319
Speaker 2: work with a bunch of different people here.

686
00:38:13,960 --> 00:38:16,720
Speaker 5: I think today what we see is that we're still

687
00:38:16,760 --> 00:38:20,879
Speaker 5: in that journey of improvement, and you see improvements every

688
00:38:20,880 --> 00:38:24,239
Speaker 5: few months from the likes of Open Ai or Anthropic

689
00:38:24,520 --> 00:38:28,080
Speaker 5: or Gemini or whoever it might be. And so I

690
00:38:28,080 --> 00:38:30,000
Speaker 5: think from that perspective you see a lot of innovation

691
00:38:30,160 --> 00:38:33,880
Speaker 5: also from the China market. So from that perspective, you

692
00:38:33,960 --> 00:38:36,319
Speaker 5: really want to be able to be flexible, and we

693
00:38:36,400 --> 00:38:38,360
Speaker 5: think today we're not there to bet on who is

694
00:38:38,400 --> 00:38:41,600
Speaker 5: going to win. We are there to provide the AI compute.

695
00:38:41,760 --> 00:38:44,960
Speaker 5: So from now the orchestration that makes a lot of

696
00:38:44,960 --> 00:38:48,480
Speaker 5: sense for us to allow our customers to be able

697
00:38:48,520 --> 00:38:51,920
Speaker 5: to go through the Lenovo AI and reach what works

698
00:38:51,960 --> 00:38:54,799
Speaker 5: best for them, rather than them have to figure out

699
00:38:54,840 --> 00:38:58,640
Speaker 5: and download multiple apps to be able to do that. Right,

700
00:38:58,680 --> 00:39:01,719
Speaker 5: So I think from thatspec that the orchestrator would do

701
00:39:01,800 --> 00:39:03,640
Speaker 5: that for them, and I think that makes it much

702
00:39:03,640 --> 00:39:07,759
Speaker 5: more efficient both from a memory perspective and compute capability

703
00:39:07,760 --> 00:39:11,560
Speaker 5: perspective that is on your device. So with that architecture

704
00:39:11,560 --> 00:39:16,120
Speaker 5: should actually potentially optimize the performance on the device.

705
00:39:16,800 --> 00:39:20,759
Speaker 4: Should we believe these companies, let's say in Microsoft. So

706
00:39:20,840 --> 00:39:24,759
Speaker 4: Microsoft recently announced their flagship model I forget if it's

707
00:39:24,800 --> 00:39:27,920
Speaker 4: pronouncing MAYA or may or whatever or MAI. Haven't played

708
00:39:27,920 --> 00:39:30,040
Speaker 4: with it, and they're like, and it also runs best

709
00:39:30,040 --> 00:39:33,160
Speaker 4: in our own custom silicon. How should we read that?

710
00:39:33,440 --> 00:39:36,960
Speaker 4: Is there, in your view a lot of juice to

711
00:39:37,000 --> 00:39:42,600
Speaker 4: be squeezed from model silicon alignment? Or should we read

712
00:39:42,680 --> 00:39:48,480
Speaker 4: this as at least companies would like to begin having

713
00:39:48,520 --> 00:39:51,000
Speaker 4: a little bit of a wedge so that they're not

714
00:39:51,200 --> 00:39:56,440
Speaker 4: so dependent on in video for their hardware needs.

715
00:39:56,880 --> 00:39:58,239
Speaker 5: I think there's a lot of great things that in

716
00:39:58,360 --> 00:40:00,279
Speaker 5: video is doing today, and I think a lot of

717
00:40:00,320 --> 00:40:04,279
Speaker 5: people are recognizing. So their revenues are clearly growing significantly.

718
00:40:04,320 --> 00:40:07,239
Speaker 5: I think they're probably at the core of driving this

719
00:40:07,360 --> 00:40:10,560
Speaker 5: current and enabling this current AI wave. Of course, so

720
00:40:10,600 --> 00:40:12,920
Speaker 5: I think that continues. But of course that's the costs

721
00:40:13,680 --> 00:40:16,399
Speaker 5: become to a level where people have to find alternatives.

722
00:40:16,760 --> 00:40:20,400
Speaker 5: You're also seeing people from a cost perspective making selections,

723
00:40:20,600 --> 00:40:24,239
Speaker 5: but also probably from as you say, the flexibility of

724
00:40:24,480 --> 00:40:27,680
Speaker 5: having their own architecture. But I think that's really happening

725
00:40:27,800 --> 00:40:30,759
Speaker 5: with respect to the hyperscalers who have the scale and

726
00:40:30,840 --> 00:40:35,000
Speaker 5: also the technological capability and the capital to be able

727
00:40:35,040 --> 00:40:35,440
Speaker 5: to develop.

728
00:40:35,480 --> 00:40:39,120
Speaker 4: But is it about their models really will run better

729
00:40:39,280 --> 00:40:43,080
Speaker 4: on these ships that they're designing, or is it about

730
00:40:43,600 --> 00:40:47,839
Speaker 4: long term strategic at least you know, I wouldn't want

731
00:40:47,840 --> 00:40:51,440
Speaker 4: to say divorced from in video, but maybe you know,

732
00:40:52,000 --> 00:40:55,080
Speaker 4: some time apart they had a prenup or something like that,

733
00:40:55,440 --> 00:40:58,640
Speaker 4: just like having a little bit of like not so

734
00:40:58,880 --> 00:41:04,040
Speaker 4: dependent on one companies specific chip capacity and roadmap.

735
00:41:04,239 --> 00:41:07,200
Speaker 5: I can't speak for them, so there's really from an

736
00:41:07,239 --> 00:41:09,880
Speaker 5: external point of view, but I believe it's probably a

737
00:41:09,920 --> 00:41:12,200
Speaker 5: little bit of both, right, because if you were a

738
00:41:12,480 --> 00:41:16,080
Speaker 5: corporate really deciding on your strategy, I think that's probably

739
00:41:16,080 --> 00:41:18,520
Speaker 5: one where, especially given that a lot of them are

740
00:41:18,560 --> 00:41:22,080
Speaker 5: also tech stack focused companies as well, so they probably

741
00:41:22,080 --> 00:41:25,360
Speaker 5: want to have more control over their own tech stack.

742
00:41:25,600 --> 00:41:27,560
Speaker 5: But the other one really is around the cost and

743
00:41:27,600 --> 00:41:31,040
Speaker 5: also maybe future planning around what they're enabled to probably

744
00:41:31,040 --> 00:41:33,719
Speaker 5: give them more flexibility to be able to do more

745
00:41:33,960 --> 00:41:37,080
Speaker 5: if they have their own chips in other areas as well.

746
00:41:37,760 --> 00:41:41,799
Speaker 2: So speaking of diversification, if we were recording this episode

747
00:41:42,360 --> 00:41:44,880
Speaker 2: a year or so ago, we would probably still be

748
00:41:44,880 --> 00:41:47,800
Speaker 2: talking about AI, but I think we'd be talking about

749
00:41:47,800 --> 00:41:50,240
Speaker 2: trade and tariffs as well, right.

750
00:41:50,120 --> 00:41:51,000
Speaker 3: Remember that show?

751
00:41:51,120 --> 00:41:52,359
Speaker 4: I forget about that?

752
00:41:52,360 --> 00:41:52,960
Speaker 3: That's right.

753
00:41:53,560 --> 00:41:58,640
Speaker 2: So Lenovo obviously has a very large and complex supply

754
00:41:58,800 --> 00:42:02,840
Speaker 2: chain all across the world. But how does I guess

755
00:42:02,880 --> 00:42:08,120
Speaker 2: the general return of more trade restrictions actually impact your business?

756
00:42:09,640 --> 00:42:13,880
Speaker 5: Yeah, it actually happened very interestingly on April second, so

757
00:42:13,960 --> 00:42:16,480
Speaker 5: I took over on the rains our fiscal years April

758
00:42:16,520 --> 00:42:19,600
Speaker 5: first and took over on the rains on April first

759
00:42:19,719 --> 00:42:23,000
Speaker 5: or the second day. It was the most major event

760
00:42:23,520 --> 00:42:25,400
Speaker 5: in the history of a one hundred I think it

761
00:42:25,440 --> 00:42:28,080
Speaker 5: was two hundred plus countries that were enacted on this,

762
00:42:28,160 --> 00:42:29,919
Speaker 5: and we were probably in one hundred and eighty of those.

763
00:42:30,320 --> 00:42:32,239
Speaker 5: So I think it's a major event and given the

764
00:42:32,280 --> 00:42:36,279
Speaker 5: significance of our business, But it's really about enabling and

765
00:42:36,400 --> 00:42:40,480
Speaker 5: providing our products at a very reasonable price to our

766
00:42:40,560 --> 00:42:43,480
Speaker 5: end customers, right, So I think the end customer suffers

767
00:42:43,920 --> 00:42:47,400
Speaker 5: if there is inefficiencies added. So I think that's the

768
00:42:47,440 --> 00:42:51,719
Speaker 5: most important aspect, which is and I think eventually right,

769
00:42:51,760 --> 00:42:54,759
Speaker 5: I think PC products were actually exempt, So I think

770
00:42:54,760 --> 00:42:58,000
Speaker 5: there was a conclusion that this is something that was needed,

771
00:42:58,120 --> 00:43:01,319
Speaker 5: right for productivity, for entertainment, for a lot of things

772
00:43:01,320 --> 00:43:03,680
Speaker 5: that people use devices for today, and we're one of

773
00:43:03,719 --> 00:43:06,239
Speaker 5: the largest providers across the board, whether it's in your

774
00:43:06,239 --> 00:43:08,759
Speaker 5: pocket every day or on your desktop or at home.

775
00:43:08,800 --> 00:43:11,440
Speaker 5: So I think from that perspective, this is probably a

776
00:43:11,600 --> 00:43:14,600
Speaker 5: very essential thing for consumers today, and I think that's

777
00:43:14,600 --> 00:43:18,440
Speaker 5: the most important in terms of giving them that product

778
00:43:18,440 --> 00:43:21,040
Speaker 5: at the lowest cost. You see that behavior really in

779
00:43:21,080 --> 00:43:23,640
Speaker 5: a lot of probably third world countries or second world

780
00:43:23,640 --> 00:43:27,359
Speaker 5: countries in terms of really emphasizing that, because it's really

781
00:43:27,360 --> 00:43:31,040
Speaker 5: about not leaving their citizens behind. It's really about making

782
00:43:31,040 --> 00:43:33,759
Speaker 5: sure that they have access to digital information, and the

783
00:43:33,840 --> 00:43:36,080
Speaker 5: device is the start of that journey, right, So that's

784
00:43:36,080 --> 00:43:40,160
Speaker 5: a very critical aspect of it. So lowering the tariff

785
00:43:40,160 --> 00:43:44,080
Speaker 5: barrier is actually essential to ensure that their citizens and

786
00:43:44,200 --> 00:43:48,920
Speaker 5: any citizens should have the lowest cost access to technology.

787
00:43:49,440 --> 00:43:52,319
Speaker 4: You mentioned something earlier, and you're like, okay, On a

788
00:43:52,360 --> 00:43:57,359
Speaker 4: per token basis, deep Seek is significantly cheaper than say,

789
00:43:57,680 --> 00:44:01,560
Speaker 4: you know, the most advanced models from say open AI Andthropic.

790
00:44:01,719 --> 00:44:05,720
Speaker 4: But strictly speaking, there is a wide agreement that still

791
00:44:05,760 --> 00:44:09,120
Speaker 4: the flagship models from the American labs are the best

792
00:44:09,160 --> 00:44:11,319
Speaker 4: models in the entire world. And I just don't think

793
00:44:11,320 --> 00:44:13,719
Speaker 4: there's much dispute about that in either the US or

794
00:44:13,960 --> 00:44:18,399
Speaker 4: Chinese communities, at least a few months behind, perhaps even

795
00:44:18,440 --> 00:44:22,799
Speaker 4: inside Lenovo. But do you see in general, and this

796
00:44:22,920 --> 00:44:24,320
Speaker 4: sort of gets back to the early part of the

797
00:44:24,400 --> 00:44:31,000
Speaker 4: question about CFO decisions, people not being sophisticated about recognizing

798
00:44:31,040 --> 00:44:33,719
Speaker 4: the queries don't always have to go to the most expensive,

799
00:44:33,719 --> 00:44:37,359
Speaker 4: most advanced models, and how much learning is there still

800
00:44:37,480 --> 00:44:41,799
Speaker 4: yet to do about how to optimize routing of the

801
00:44:41,920 --> 00:44:46,200
Speaker 4: query such that it goes to the most cost effective

802
00:44:46,360 --> 00:44:50,760
Speaker 4: model rather than just slamming the frontier model.

803
00:44:51,160 --> 00:44:54,960
Speaker 5: Yeah, I'm going to answer your question in how I

804
00:44:55,000 --> 00:44:58,200
Speaker 5: started my morning today, because Okay, I'm in the journey

805
00:44:58,480 --> 00:45:01,840
Speaker 5: of figuring this out and I've decided to spend some

806
00:45:01,960 --> 00:45:05,160
Speaker 5: money with external consultants, and so we're in that interview

807
00:45:05,239 --> 00:45:08,160
Speaker 5: process and as I said, you know, we spend an

808
00:45:08,160 --> 00:45:10,120
Speaker 5: hour and a half with the global team, and this

809
00:45:10,200 --> 00:45:12,359
Speaker 5: is one of the major firms. I'm starting to get

810
00:45:12,360 --> 00:45:16,640
Speaker 5: a sense from talking to various parties that everyone is

811
00:45:16,719 --> 00:45:19,279
Speaker 5: new to this game. So everyone, even the advisors who

812
00:45:19,320 --> 00:45:22,319
Speaker 5: are trying to earn a service fee from you, is

813
00:45:22,360 --> 00:45:24,799
Speaker 5: really also learning as part of this journey. And that's

814
00:45:24,800 --> 00:45:27,200
Speaker 5: what consultants do as well. Right, never asked what you're paying,

815
00:45:30,560 --> 00:45:38,120
Speaker 5: but I would not comments about another industry. Yeah, absolutely absolutely.

816
00:45:38,440 --> 00:45:40,800
Speaker 5: But you know, I think from the from that perspective

817
00:45:40,880 --> 00:45:43,400
Speaker 5: that goes to say that if they are the specialists

818
00:45:43,520 --> 00:45:45,879
Speaker 5: or those that see a lot in the market, then

819
00:45:45,920 --> 00:45:48,680
Speaker 5: what about specific departments or individuals in the company. So

820
00:45:48,719 --> 00:45:51,400
Speaker 5: I think from that perspective, everyone is in that journey,

821
00:45:51,400 --> 00:45:54,600
Speaker 5: and it's the people who can be most clear headed

822
00:45:54,640 --> 00:45:57,600
Speaker 5: to learn the fastest. I think it's probably the most essential, right,

823
00:45:57,640 --> 00:46:00,200
Speaker 5: So I think from that perspective, we need to be

824
00:46:00,280 --> 00:46:04,680
Speaker 5: able to take out the red tape, enable the processes

825
00:46:04,719 --> 00:46:09,080
Speaker 5: to work, accelerate. I think timing and speed is essential,

826
00:46:09,440 --> 00:46:11,759
Speaker 5: and they are in a new AI world, and then

827
00:46:11,800 --> 00:46:14,239
Speaker 5: we need to deploy that where it's sensible. And then,

828
00:46:14,280 --> 00:46:17,040
Speaker 5: as you say, where do you actually allow that spending

829
00:46:17,280 --> 00:46:21,360
Speaker 5: and the interaction to generate tokens to really happen? And

830
00:46:21,400 --> 00:46:24,680
Speaker 5: so we're in that journey. But I think we're not alone.

831
00:46:24,680 --> 00:46:26,640
Speaker 5: I think every company is the same. I'm actually going

832
00:46:26,640 --> 00:46:29,240
Speaker 5: to a conference next week hosted by a major consulting

833
00:46:29,280 --> 00:46:32,279
Speaker 5: firm that is a gathering with all of CFOs. But

834
00:46:32,640 --> 00:46:36,479
Speaker 5: in my regular diarog with CFOs, I understand why worry

835
00:46:36,520 --> 00:46:38,200
Speaker 5: a lot, But I think we're also not alone.

836
00:46:38,719 --> 00:46:41,399
Speaker 2: Joe, do you think in the future, when people meet

837
00:46:41,400 --> 00:46:43,480
Speaker 2: each other, they're going to say their names and their

838
00:46:43,560 --> 00:46:47,200
Speaker 2: nationality and then like declare Yeah, well, I was going

839
00:46:47,239 --> 00:46:50,600
Speaker 2: to say, declare the foundational model that like they use.

840
00:46:50,840 --> 00:46:52,960
Speaker 4: Oh, I was going to say, like, I wonder if

841
00:46:53,040 --> 00:46:57,200
Speaker 4: like at bars or like women when they go like,

842
00:46:57,280 --> 00:46:59,560
Speaker 4: I want a man who is like this much tooken

843
00:47:00,239 --> 00:47:04,480
Speaker 4: you know whatever. Sorry, it'll be part of their identities

844
00:47:04,520 --> 00:47:04,879
Speaker 4: for sure.

845
00:47:05,360 --> 00:47:05,839
Speaker 3: That's right.

846
00:47:05,880 --> 00:47:08,799
Speaker 2: Well, okay, so speaking of rich men, I have to

847
00:47:08,840 --> 00:47:12,480
Speaker 2: ask at least one cfo ish question, which is when

848
00:47:12,520 --> 00:47:15,720
Speaker 2: it comes to capital spend, I'm sure you could justify

849
00:47:15,840 --> 00:47:19,240
Speaker 2: pretty much anything right now. And there's so much money

850
00:47:19,280 --> 00:47:22,759
Speaker 2: that's actually going out on the AI build out. How

851
00:47:22,760 --> 00:47:27,400
Speaker 2: are you thinking about returning capital to shareholders because the

852
00:47:27,440 --> 00:47:29,920
Speaker 2: stock's gone up a lot, but you know, people can

853
00:47:29,960 --> 00:47:32,799
Speaker 2: always be wealthier. I'm sure they're into that.

854
00:47:33,200 --> 00:47:35,879
Speaker 5: Yeah, we paid the in the last fiscal year, which

855
00:47:35,960 --> 00:47:39,520
Speaker 5: just ended March thirty first, we paid the highest dividend

856
00:47:39,840 --> 00:47:42,920
Speaker 5: ever in the noble history, So our shareholders are very happy.

857
00:47:42,960 --> 00:47:46,800
Speaker 5: We'd share our success with our shareholders, but we also

858
00:47:46,920 --> 00:47:49,560
Speaker 5: see a lot of opportunities for growth, and as you

859
00:47:49,800 --> 00:47:53,439
Speaker 5: have mentioned many times on this discussion, is that there's

860
00:47:53,480 --> 00:47:55,600
Speaker 5: also a lot of capital needs to feel that growth.

861
00:47:55,760 --> 00:47:59,240
Speaker 5: So we really need to strike that balance between giving

862
00:47:59,280 --> 00:48:02,319
Speaker 5: our shareholders the media cash that they would like, but

863
00:48:02,440 --> 00:48:04,759
Speaker 5: also at the same time they want us to be

864
00:48:04,800 --> 00:48:08,080
Speaker 5: able to enable the capital appreciation on the stock. Because

865
00:48:08,120 --> 00:48:11,279
Speaker 5: the underlying fundamental The business is growing, so we need

866
00:48:11,320 --> 00:48:14,080
Speaker 5: to drive that growth. And if you there's there's going

867
00:48:14,120 --> 00:48:15,799
Speaker 5: to definitely going to be a period where you drive

868
00:48:15,800 --> 00:48:19,320
Speaker 5: that growth at the same time we're driving that margin expansion.

869
00:48:19,760 --> 00:48:22,720
Speaker 5: But there will be quarters here and there that could

870
00:48:22,719 --> 00:48:26,400
Speaker 5: potentially mismatch in terms of that growth versus the margins.

871
00:48:26,760 --> 00:48:30,000
Speaker 5: But overall, the long term trend for Lenovo is a

872
00:48:30,360 --> 00:48:33,960
Speaker 5: plan to drive growth with accelerated margin expansion.

873
00:48:34,480 --> 00:48:37,400
Speaker 4: Northern Virginia in the US is sort of understood to

874
00:48:37,480 --> 00:48:40,600
Speaker 4: be the data center capital of America, but you know,

875
00:48:40,680 --> 00:48:43,120
Speaker 4: it's expanding. There's a lot that's sort of like off

876
00:48:43,160 --> 00:48:46,040
Speaker 4: grid and Texas that's happening. And then there's also of

877
00:48:46,080 --> 00:48:49,319
Speaker 4: course anti data center politics in the US, and we

878
00:48:49,400 --> 00:48:51,799
Speaker 4: don't know how that will affect the map. What is

879
00:48:51,920 --> 00:48:56,120
Speaker 4: the Northern Virginia of East Asia right now? Where do

880
00:48:56,160 --> 00:48:58,840
Speaker 4: you see? Like which country is it? Because they have

881
00:48:58,920 --> 00:49:02,279
Speaker 4: the most whether it's regular access to energy, Where is

882
00:49:02,360 --> 00:49:05,920
Speaker 4: everyone trying to put up data centers? And I'm just curious,

883
00:49:05,960 --> 00:49:09,279
Speaker 4: like anywhere in Asia is there any sort of equivalent

884
00:49:09,400 --> 00:49:12,120
Speaker 4: anti data center politics that's emerging.

885
00:49:13,160 --> 00:49:16,040
Speaker 5: Well, I think Asia tends to be a little bit

886
00:49:16,040 --> 00:49:19,839
Speaker 5: more flexible from that perspective, but clearly the best infrastructure

887
00:49:19,880 --> 00:49:23,120
Speaker 5: of scale is in China. Yeah, I think from that perspective,

888
00:49:23,160 --> 00:49:26,120
Speaker 5: a lot of people, due to certain regulations and policies,

889
00:49:26,120 --> 00:49:28,120
Speaker 5: cannot have data center. So you have a lot of

890
00:49:28,200 --> 00:49:33,560
Speaker 5: idle resources today, whether it's power and data center therefore

891
00:49:33,960 --> 00:49:37,480
Speaker 5: accessibility and also supply chain. Right, that's not that the

892
00:49:37,480 --> 00:49:41,279
Speaker 5: world is spending additional money on in higher cost jurisdictions

893
00:49:41,640 --> 00:49:43,960
Speaker 5: because of regulations that you cannot put it in China.

894
00:49:44,000 --> 00:49:46,839
Speaker 5: So the world ends up having some of the inflationary

895
00:49:46,880 --> 00:49:50,560
Speaker 5: effects is because you're not actually using the most efficient

896
00:49:50,560 --> 00:49:52,480
Speaker 5: place in the world that you can actually do things.

897
00:49:52,840 --> 00:49:55,680
Speaker 5: But having said that, I think other countries are actually

898
00:49:55,719 --> 00:49:58,560
Speaker 5: catching up very fast as well. So we're seeing even

899
00:49:58,560 --> 00:50:01,360
Speaker 5: if the supply chain area slightly lower costs, but in

900
00:50:01,440 --> 00:50:03,920
Speaker 5: terms of the data center aspect and where the power

901
00:50:04,160 --> 00:50:06,319
Speaker 5: Southeast Asia has been very good, and I think what

902
00:50:06,440 --> 00:50:08,840
Speaker 5: we're seeing in this part of the world clearly in

903
00:50:08,880 --> 00:50:13,560
Speaker 5: places like Malaysia and Indonesia, I think those are natural

904
00:50:13,560 --> 00:50:16,920
Speaker 5: places where people have really been looking into. I mean

905
00:50:17,440 --> 00:50:20,399
Speaker 5: even Hong Kong these days. I think there's plans from

906
00:50:20,400 --> 00:50:22,680
Speaker 5: the government, you would think, and also Singapore, so some

907
00:50:22,719 --> 00:50:26,680
Speaker 5: of these smaller land mass areas, but I think also

908
00:50:26,719 --> 00:50:28,839
Speaker 5: have plans and I think right here in town there's

909
00:50:28,840 --> 00:50:31,960
Speaker 5: also a dedicated a few players as well, and they

910
00:50:32,000 --> 00:50:34,680
Speaker 5: have to be creative in terms of putting data centers

911
00:50:34,680 --> 00:50:39,480
Speaker 5: in these higher alt VTA structures. But there's all product,

912
00:50:39,920 --> 00:50:42,279
Speaker 5: all sort of creativity, and I think Hong Kong has

913
00:50:42,280 --> 00:50:44,680
Speaker 5: a mixed use in terms of that power generation. But

914
00:50:45,000 --> 00:50:47,920
Speaker 5: again that's a small area, but I think it's just

915
00:50:47,920 --> 00:50:50,120
Speaker 5: giving a data set of examples. People are also doing

916
00:50:50,120 --> 00:50:52,840
Speaker 5: it in Japan. Japan is also another one. From a

917
00:50:52,840 --> 00:50:57,239
Speaker 5: regulatory perspective, people do, but cost wise and also policy

918
00:50:57,239 --> 00:50:59,719
Speaker 5: wise in terms of getting necessary permits may not be

919
00:50:59,760 --> 00:51:03,640
Speaker 5: the fat a long time. There's so may have some

920
00:51:03,840 --> 00:51:06,680
Speaker 5: knowledge about this, but I think it really depends on

921
00:51:07,280 --> 00:51:11,360
Speaker 5: where you can have trust and partnership with the local government.

922
00:51:11,400 --> 00:51:15,080
Speaker 5: Because it talks, it involves land, it involves power, it

923
00:51:15,160 --> 00:51:19,360
Speaker 5: involves imports of specific goods, and I think that part

924
00:51:19,640 --> 00:51:23,080
Speaker 5: really depends on a lot of the local efficiencies.

925
00:51:23,360 --> 00:51:25,520
Speaker 3: Joe, we got to go visit a data center in

926
00:51:25,600 --> 00:51:26,040
Speaker 3: Hong Kong.

927
00:51:26,120 --> 00:51:27,920
Speaker 4: I want to well, I was going to say, Hong

928
00:51:28,000 --> 00:51:29,799
Speaker 4: Kong actually to me makes a lot of sense as

929
00:51:29,800 --> 00:51:31,680
Speaker 4: a data center location. Because if there's one thing I

930
00:51:31,719 --> 00:51:35,120
Speaker 4: know about Hong Kong's there's plenty of cheap real estate

931
00:51:35,320 --> 00:51:39,000
Speaker 4: here and you just can't and there's more than enough space,

932
00:51:39,120 --> 00:51:41,960
Speaker 4: and we know that people are very happy with the

933
00:51:42,000 --> 00:51:45,360
Speaker 4: price of how affordable the residence is here, et cetera.

934
00:51:45,760 --> 00:51:46,960
Speaker 4: So it makes sense.

935
00:51:46,800 --> 00:51:49,759
Speaker 2: That you know there is actually a lot of land,

936
00:51:49,800 --> 00:51:50,960
Speaker 2: you just can't build.

937
00:51:50,640 --> 00:51:53,000
Speaker 4: On it, you justild yeah right, yeah, so.

938
00:51:53,200 --> 00:51:56,040
Speaker 2: All right, interesting Winston Chang, thank you so much for

939
00:51:56,080 --> 00:51:56,960
Speaker 2: coming on all thoughts.

940
00:51:57,040 --> 00:51:57,960
Speaker 3: Really appreciate it.

941
00:51:58,160 --> 00:52:12,319
Speaker 5: Thank you, Tracy, thank you Joe Joe.

942
00:52:12,360 --> 00:52:13,200
Speaker 3: That was a lot of fun.

943
00:52:13,680 --> 00:52:17,320
Speaker 2: I feel like we haven't actually spoken about AI from

944
00:52:17,400 --> 00:52:20,040
Speaker 2: the perspective of a company like Lenovo.

945
00:52:20,160 --> 00:52:21,840
Speaker 3: No, we have never done that before.

946
00:52:21,960 --> 00:52:24,200
Speaker 4: Yeah, no, we haven't, and that was very nice. And look,

947
00:52:24,239 --> 00:52:26,680
Speaker 4: part of the I would say two things about AI,

948
00:52:27,040 --> 00:52:29,719
Speaker 4: which is we're doing a lot of AI episodes, and

949
00:52:29,719 --> 00:52:32,440
Speaker 4: I would say there's two reasons for that. One is

950
00:52:32,480 --> 00:52:35,120
Speaker 4: because I mean, it's just the biggest thing of our

951
00:52:35,160 --> 00:52:38,520
Speaker 4: lives probably in terms of the significance and trying to

952
00:52:38,560 --> 00:52:40,759
Speaker 4: everyone's trying to wrap their heads around it, so there's

953
00:52:40,800 --> 00:52:43,160
Speaker 4: a lot to learn. But it also is in our

954
00:52:43,200 --> 00:52:46,360
Speaker 4: wheelhouse because it's so physical and because it interacts with

955
00:52:46,440 --> 00:52:49,719
Speaker 4: supply chains R and so you think, like, Okay, a

956
00:52:49,800 --> 00:52:54,279
Speaker 4: company like Lenovo, a server company that also happens to

957
00:52:54,920 --> 00:52:57,600
Speaker 4: offer the ability to build complete data centers, not just

958
00:52:57,640 --> 00:53:00,879
Speaker 4: the servers, who therefore has been interact acting with all

959
00:53:00,960 --> 00:53:04,480
Speaker 4: types of players, is almost like the perfect sweet spot

960
00:53:04,600 --> 00:53:07,600
Speaker 4: for like an odd lots guest, even among all of

961
00:53:07,680 --> 00:53:10,919
Speaker 4: our perfect guests, someone who has that visibility on both

962
00:53:10,960 --> 00:53:14,120
Speaker 4: sides is really someone interesting to talk to.

963
00:53:14,360 --> 00:53:16,440
Speaker 2: Yeah, I think I've said before there's such a weird

964
00:53:16,520 --> 00:53:21,120
Speaker 2: tension between these sort of bodyless, faceless AI in interface

965
00:53:21,160 --> 00:53:24,520
Speaker 2: and then when you think about all the actual physical

966
00:53:24,560 --> 00:53:26,640
Speaker 2: infrastructure that supports it totally.

967
00:53:27,360 --> 00:53:30,080
Speaker 4: It's why I think it's so interesting. The other thing

968
00:53:30,120 --> 00:53:33,040
Speaker 4: to it that I think is interesting about hearing from

969
00:53:33,040 --> 00:53:37,240
Speaker 4: a Lenovo and again hearing from a CFO of Lenovo,

970
00:53:37,320 --> 00:53:39,640
Speaker 4: it's like, look, they're not a hyperscaler. I'm sure they're

971
00:53:39,680 --> 00:53:42,279
Speaker 4: investing a ton of money and building out their capabilities,

972
00:53:42,320 --> 00:53:45,239
Speaker 4: et cetera, but they're not one of these companies that

973
00:53:45,280 --> 00:53:47,959
Speaker 4: announces big things like we're going to be spending five

974
00:53:48,040 --> 00:53:51,360
Speaker 4: hundred billion dollars next year on building out data centers,

975
00:53:51,680 --> 00:53:55,400
Speaker 4: So they really do have that dimension where they have

976
00:53:55,520 --> 00:53:59,400
Speaker 4: to figure out the Googles and et cetera. They and

977
00:53:59,440 --> 00:54:02,719
Speaker 4: the open air they're talking about Capex and the Lenova's

978
00:54:02,719 --> 00:54:04,359
Speaker 4: of the world also have to think a lot more

979
00:54:04,360 --> 00:54:08,000
Speaker 4: about op X and the op X component of the conversation.

980
00:54:08,320 --> 00:54:11,839
Speaker 4: As he said, whether it's with fellow CFOs or the

981
00:54:11,880 --> 00:54:15,399
Speaker 4: consultant community, it doesn't sound like it feels like it's

982
00:54:15,400 --> 00:54:16,560
Speaker 4: in day one of figuring out.

983
00:54:16,600 --> 00:54:18,280
Speaker 3: Yeah, everything's up for grabs.

984
00:54:18,360 --> 00:54:18,760
Speaker 4: Yeah.

985
00:54:19,080 --> 00:54:23,960
Speaker 2: I also thought the involution point Yeah, was interesting when

986
00:54:24,000 --> 00:54:26,840
Speaker 2: we were talking about the key differences between China and

987
00:54:27,040 --> 00:54:31,160
Speaker 2: us AI and the idea that because competition is so

988
00:54:31,520 --> 00:54:35,240
Speaker 2: cut through in China that it just drives your costs

989
00:54:35,280 --> 00:54:38,279
Speaker 2: down and down and down, so that like, even if

990
00:54:38,280 --> 00:54:40,400
Speaker 2: you're not that competitive in China, you would still be

991
00:54:40,440 --> 00:54:41,600
Speaker 2: competitive competitive.

992
00:54:42,239 --> 00:54:42,800
Speaker 3: Yeah, exactly.

993
00:54:42,800 --> 00:54:45,120
Speaker 4: It's kind of funny to think about. Yeah. Absolutely, No,

994
00:54:45,200 --> 00:54:45,879
Speaker 4: I love that chat.

995
00:54:46,200 --> 00:54:47,640
Speaker 3: All right, shall we leave it there.

996
00:54:47,719 --> 00:54:48,439
Speaker 4: Let's leave it there.

997
00:54:48,520 --> 00:54:51,120
Speaker 2: This has been another episode of the All Thoughts podcast.

998
00:54:51,239 --> 00:54:54,440
Speaker 2: I'm Tracy Alloway. You can follow me at Tracy Alloway.

999
00:54:54,200 --> 00:54:56,560
Speaker 4: And I'm Joe Wisenthal. You can follow me at The

1000
00:54:56,640 --> 00:55:00,760
Speaker 4: Stalwart follow our producers Carmen Rodriguez at Carmen arm Dash,

1001
00:55:00,760 --> 00:55:04,640
Speaker 4: Ol Bennett at Dashbox, Calebrooks at Kilbrooks and Kevin Lozano

1002
00:55:04,719 --> 00:55:07,680
Speaker 4: at Kevin Lloyd Lozano and from our Odd Lots content.

1003
00:55:07,719 --> 00:55:10,040
Speaker 4: Go to Bloomberg dot com slash odd Lots or the

1004
00:55:10,120 --> 00:55:12,640
Speaker 4: daily newsletter and all of our episodes, and you can

1005
00:55:12,719 --> 00:55:14,840
Speaker 4: chat about all these topics twenty four to seven in

1006
00:55:14,960 --> 00:55:18,280
Speaker 4: our discord Discord dot gg slash Oddlines.

1007
00:55:18,840 --> 00:55:20,960
Speaker 2: And if you enjoy odd Lots, if you like it

1008
00:55:21,040 --> 00:55:24,440
Speaker 2: when you hear from the perspective of Chinese tech companies,

1009
00:55:24,440 --> 00:55:27,120
Speaker 2: then please leave us a positive review on your favorite

1010
00:55:27,120 --> 00:55:31,200
Speaker 2: podcast platform. And remember, if you are a Bloomberg subscriber,

1011
00:55:31,320 --> 00:55:32,640
Speaker 2: you can listen to all of our.

1012
00:55:32,560 --> 00:55:34,600
Speaker 3: Episodes absolutely ad free.

1013
00:55:34,800 --> 00:55:36,880
Speaker 2: All you need to do is find the Bloomberg channel

1014
00:55:36,920 --> 00:55:39,560
Speaker 2: on Apple Podcasts and follow the instructions there.

1015
00:55:40,160 --> 00:55:41,000
Speaker 3: Thanks for listening.

1016
00:56:06,120 --> 00:56:06,840
Speaker 5: In e