1
00:00:00,680 --> 00:00:05,520
Speaker 1: Welcome to Zero I am Akshatrati. This week AI's climate impact.

2
00:00:17,480 --> 00:00:21,400
Speaker 1: We've been wanting to do episodes exploring artificial intelligence, energy

3
00:00:21,680 --> 00:00:24,200
Speaker 1: and the climate for some time now, and we have

4
00:00:24,280 --> 00:00:26,479
Speaker 1: some exciting things in the works for you in the

5
00:00:26,480 --> 00:00:30,800
Speaker 1: weeks ahead. There are basically two big questions here, what

6
00:00:30,840 --> 00:00:35,000
Speaker 1: can artificial intelligence do to help fight climate change? And

7
00:00:35,320 --> 00:00:41,000
Speaker 1: on the flip side, how might AI's resource consumption exacerbate emissions?

8
00:00:42,720 --> 00:00:46,280
Speaker 1: Today we are talking about the second question because we've

9
00:00:46,400 --> 00:00:49,640
Speaker 1: just found the first concrete example of how the pursuit

10
00:00:49,680 --> 00:00:53,199
Speaker 1: of AI is colliding with the efforts to cut emissions.

11
00:00:54,200 --> 00:00:57,280
Speaker 1: Last week, my Bloomber colleague Dina Bass and I wrote

12
00:00:57,320 --> 00:01:01,560
Speaker 1: a story about the impact of Microsoft's AI expansion, and

13
00:01:01,640 --> 00:01:05,640
Speaker 1: we interviewed Microsoft president Bradsmith to get his Reactionshi. Hello

14
00:01:06,520 --> 00:01:11,399
Speaker 1: audio test here, Hi les. In twenty twenty, Microsoft pledge

15
00:01:11,440 --> 00:01:14,000
Speaker 1: to remove more carbon than it emits by the end

16
00:01:14,000 --> 00:01:17,160
Speaker 1: of this decade. If the company was on track, then

17
00:01:17,200 --> 00:01:20,760
Speaker 1: its emissions should have been roughly thirty percent lower in

18
00:01:20,840 --> 00:01:25,080
Speaker 1: twenty twenty three, but as the company's AI ambitions have expanded,

19
00:01:25,480 --> 00:01:30,760
Speaker 1: its emissions have instead increased by thirty percent. Some of

20
00:01:30,800 --> 00:01:34,720
Speaker 1: the numbers are startling. Last year, the tech giant's power

21
00:01:34,800 --> 00:01:39,759
Speaker 1: consumption rivaled that of a small European country, Slovenia consumed

22
00:01:40,000 --> 00:01:45,080
Speaker 1: less electricity in twenty twenty three than Microsoft did, and

23
00:01:45,120 --> 00:01:48,720
Speaker 1: its development that's likely to be repeated by other competitors

24
00:01:48,760 --> 00:01:52,920
Speaker 1: that are also chasing their own AI goals. After all,

25
00:01:53,120 --> 00:01:56,800
Speaker 1: Microsoft's AI push has made it the world's most valuable company,

26
00:01:57,120 --> 00:02:02,600
Speaker 1: worth more than three trillion dollars. Companies want what Microsoft has,

27
00:02:03,680 --> 00:02:06,400
Speaker 1: but Microsoft is also the company that's been leading on

28
00:02:06,480 --> 00:02:09,440
Speaker 1: climate plans. It's set one of the most ambitious goals

29
00:02:09,440 --> 00:02:12,480
Speaker 1: for any large company, not just going carbon negative by

30
00:02:12,520 --> 00:02:16,680
Speaker 1: twenty thirty, but also promising to erase its legacy emissions,

31
00:02:17,280 --> 00:02:20,600
Speaker 1: and it came up with a comprehensive plan which is

32
00:02:20,720 --> 00:02:24,840
Speaker 1: throwing money and other resources at almost every possible climate

33
00:02:24,880 --> 00:02:29,560
Speaker 1: solution it could deploy. So when it comes to cutting emissions,

34
00:02:29,880 --> 00:02:33,880
Speaker 1: if Microsoft can't do it, can any other company. On

35
00:02:33,919 --> 00:02:38,000
Speaker 1: today's episode, zero's producer Mightily Rao asks me the questions.

36
00:02:38,200 --> 00:02:41,639
Speaker 2: Hi, that's right, We're going to discuss Microsoft's climate challenge.

37
00:02:41,760 --> 00:02:44,200
Speaker 2: What Brad Smith had to say about the company's rising

38
00:02:44,200 --> 00:02:47,400
Speaker 2: emissions and how to think about climate in the AI

39
00:02:47,520 --> 00:03:03,400
Speaker 2: age Actually let's start with some background in terms of

40
00:03:03,400 --> 00:03:07,600
Speaker 2: reducing emissions. What did Microsoft pledge to do and how

41
00:03:07,639 --> 00:03:09,440
Speaker 2: realistic did that promise seem at the time.

42
00:03:10,160 --> 00:03:13,120
Speaker 1: Let me take one more step back, actually and start

43
00:03:13,160 --> 00:03:16,960
Speaker 1: in twenty eighteen, because before that we didn't even have

44
00:03:17,240 --> 00:03:20,560
Speaker 1: ned zero. It was a report from the Intergovernmental Panel

45
00:03:20,560 --> 00:03:23,680
Speaker 1: on Climate Change that said that if the world wants

46
00:03:23,680 --> 00:03:26,720
Speaker 1: to keep warming below one point five degrees celsius, then

47
00:03:26,960 --> 00:03:30,200
Speaker 1: the entire world needs to reach net zero carbon dioxide

48
00:03:30,200 --> 00:03:34,040
Speaker 1: emissions by twenty fifty. Soon we saw a wave of

49
00:03:34,120 --> 00:03:37,320
Speaker 1: companies coming up with climate targets. Tech companies, which of

50
00:03:37,320 --> 00:03:39,880
Speaker 1: course have a lot of money and mostly have their

51
00:03:39,920 --> 00:03:44,040
Speaker 1: emissions tied to consuming electricity, were ahead in the game.

52
00:03:44,560 --> 00:03:47,320
Speaker 1: We got a goal in twenty twenty from Microsoft saying

53
00:03:47,560 --> 00:03:49,840
Speaker 1: we're going to be carbon negative by the end of

54
00:03:49,880 --> 00:03:53,760
Speaker 1: this decade. Very ambitious, yes, and it wasn't clear then

55
00:03:54,320 --> 00:03:56,760
Speaker 1: how it would reach that. But because it was a

56
00:03:56,800 --> 00:04:00,760
Speaker 1: tech company whose emissions were down to electricity, and electricity

57
00:04:01,240 --> 00:04:03,920
Speaker 1: was getting cleaner, and of course solar and wind were

58
00:04:03,920 --> 00:04:07,760
Speaker 1: getting cheaper, it felt like it is something a tech

59
00:04:07,760 --> 00:04:08,520
Speaker 1: company could do.

60
00:04:09,480 --> 00:04:13,120
Speaker 2: So Microsoft set this climate goal in twenty twenty, but

61
00:04:13,560 --> 00:04:16,720
Speaker 2: around the same time, AI started to become a bigger

62
00:04:16,760 --> 00:04:20,680
Speaker 2: priority for the company. Let's talk about that parallel timeline.

63
00:04:21,080 --> 00:04:24,320
Speaker 1: So a little known story now, but in twenty nineteen,

64
00:04:24,680 --> 00:04:26,640
Speaker 1: just a year before it put out its climate plan,

65
00:04:27,720 --> 00:04:31,560
Speaker 1: Microsoft invested in open Ai, which was then an up

66
00:04:31,640 --> 00:04:35,400
Speaker 1: and coming startup that was starting to build products around

67
00:04:35,440 --> 00:04:38,400
Speaker 1: this thing called generative AI. Of course we know that

68
00:04:38,440 --> 00:04:41,839
Speaker 1: company today as the maker of chad GPT, and everybody

69
00:04:41,880 --> 00:04:45,360
Speaker 1: has their own feelings about it. But what it meant

70
00:04:45,600 --> 00:04:49,080
Speaker 1: was AI is going to start to consume a lot

71
00:04:49,160 --> 00:04:53,719
Speaker 1: of computing resources, a lot of data processing. And Microsoft,

72
00:04:53,800 --> 00:04:57,720
Speaker 1: as a provider of cloud services of computing services, both

73
00:04:57,760 --> 00:05:00,480
Speaker 1: could be a customer to an AI produce, sir, but

74
00:05:00,600 --> 00:05:03,960
Speaker 1: also could be a seller of those AI products. And

75
00:05:04,600 --> 00:05:08,640
Speaker 1: we've seen Microsoft's bet on OpenAI is starting to pay off.

76
00:05:08,760 --> 00:05:12,960
Speaker 1: There is clearly demand for these products. Maybe it's hype

77
00:05:13,040 --> 00:05:16,720
Speaker 1: right now, but what is clear is that other competitors

78
00:05:16,880 --> 00:05:21,040
Speaker 1: want what Microsoft has. So Google is investing in AI, Meta,

79
00:05:21,160 --> 00:05:23,960
Speaker 1: the company behind Facebook, is investing in AI. Amazon is

80
00:05:24,000 --> 00:05:26,280
Speaker 1: doing it. There a bunch of Chinese companies doing it.

81
00:05:26,480 --> 00:05:28,200
Speaker 1: Everybody is in this race now.

82
00:05:29,360 --> 00:05:32,600
Speaker 2: So in the meantime, Microsoft has fallen short on the

83
00:05:32,680 --> 00:05:36,160
Speaker 2: emissions goals that it's set. You asked Microsoft president Brad

84
00:05:36,200 --> 00:05:39,120
Speaker 2: Smith about that, and we can play a little bit

85
00:05:39,160 --> 00:05:39,960
Speaker 2: of his answer.

86
00:05:40,440 --> 00:05:43,599
Speaker 3: You know, in twenty twenty, we unveiled what we called

87
00:05:43,600 --> 00:05:47,520
Speaker 3: our carbon Moonshot, our goal of being carbon negative by

88
00:05:47,680 --> 00:05:53,920
Speaker 3: twenty thirty. That was before the explosion and artificial intelligence.

89
00:05:54,800 --> 00:05:58,400
Speaker 3: So in many ways, as I say across Microsoft, the

90
00:05:58,400 --> 00:06:01,720
Speaker 3: moon has moved. It's it's more than five times as

91
00:06:01,760 --> 00:06:05,120
Speaker 3: far away as it was in twenty twenty. If you

92
00:06:05,279 --> 00:06:09,279
Speaker 3: just think about our own forecast for the expansion of

93
00:06:09,320 --> 00:06:11,960
Speaker 3: AI and its electrical.

94
00:06:11,520 --> 00:06:14,320
Speaker 2: Needs, and again let's go back to the story you wrote,

95
00:06:14,320 --> 00:06:18,520
Speaker 2: because the figures are pretty astonishing. Microsoft's total planet warming

96
00:06:18,600 --> 00:06:21,760
Speaker 2: impact is about thirty percent higher today than it was

97
00:06:21,800 --> 00:06:25,320
Speaker 2: in twenty twenty. So is that all down to AI.

98
00:06:26,520 --> 00:06:29,560
Speaker 1: Well, that seems to be the case. Now. AI has

99
00:06:29,600 --> 00:06:32,760
Speaker 1: two aspects to it. One is that it consumes more

100
00:06:32,800 --> 00:06:36,760
Speaker 1: and more compute, as they say, which requires electricity, And

101
00:06:36,880 --> 00:06:40,839
Speaker 1: of course more compute means it requires more computing power,

102
00:06:40,960 --> 00:06:43,880
Speaker 1: which comes from building data centers that are built out

103
00:06:43,920 --> 00:06:49,680
Speaker 1: of steel, cement and microchips. So because Microsoft says it's

104
00:06:49,680 --> 00:06:52,279
Speaker 1: one hundred percent powered by renewables, it's not really the

105
00:06:52,320 --> 00:06:56,960
Speaker 1: electricity consumption that is contributing towards this emissions increase, which

106
00:06:57,000 --> 00:07:00,320
Speaker 1: means most of those emissions increase is coming from the

107
00:07:00,360 --> 00:07:04,839
Speaker 1: construction of data centers because there is no green cement

108
00:07:05,000 --> 00:07:08,560
Speaker 1: or steel or microchips which are zero carbon that you

109
00:07:08,560 --> 00:07:12,280
Speaker 1: can deploy right away, and that's a big problem for

110
00:07:12,360 --> 00:07:13,360
Speaker 1: Microsoft right now.

111
00:07:13,760 --> 00:07:16,920
Speaker 2: On what basis is Microsoft saying that they're entirely powered

112
00:07:16,960 --> 00:07:18,080
Speaker 2: by renewables.

113
00:07:18,560 --> 00:07:21,960
Speaker 1: It's a good question, and more people should ask this

114
00:07:22,120 --> 00:07:24,840
Speaker 1: question because a lot of companies say they are one

115
00:07:24,920 --> 00:07:29,320
Speaker 1: hundred percent powered by renewables, which is not technically true,

116
00:07:29,560 --> 00:07:33,440
Speaker 1: because what is happening is that Microsoft and other tech

117
00:07:33,480 --> 00:07:38,440
Speaker 1: companies sign something called corporate power purchase agreements PPAs, and

118
00:07:38,520 --> 00:07:41,400
Speaker 1: these are where they actually go out pay a developer

119
00:07:41,440 --> 00:07:44,880
Speaker 1: to deploy a solar plant and consume all the electricity

120
00:07:44,920 --> 00:07:48,200
Speaker 1: that solar plant is producing. So that is renewable energy,

121
00:07:48,760 --> 00:07:52,040
Speaker 1: but there's a lot of electricity there is also consumed

122
00:07:52,080 --> 00:07:54,920
Speaker 1: at night when the sun is not shining that might

123
00:07:55,000 --> 00:07:58,960
Speaker 1: come from a fossil fuel grid, coal or gas. Many

124
00:07:58,960 --> 00:08:03,200
Speaker 1: companies then something called a renewable energy credit, which simply

125
00:08:03,280 --> 00:08:06,920
Speaker 1: turns a coal electron into a solar electron and that

126
00:08:07,120 --> 00:08:10,120
Speaker 1: was fine in the initial days when there wasn't enough

127
00:08:10,160 --> 00:08:14,040
Speaker 1: renewables on the grid, because those extra credits made renewables

128
00:08:14,080 --> 00:08:17,080
Speaker 1: a little more profitable because at the time they were expensive,

129
00:08:17,680 --> 00:08:22,720
Speaker 1: and that meant it incentivized the deployment of renewables. Today,

130
00:08:23,360 --> 00:08:28,160
Speaker 1: those credits do not really incentivize building out more renewables,

131
00:08:28,520 --> 00:08:31,440
Speaker 1: and so work that my colleague Ben Elgin has done

132
00:08:31,680 --> 00:08:35,840
Speaker 1: has found that in fact, using these renewable energy credits

133
00:08:36,040 --> 00:08:41,280
Speaker 1: can mask a company's emissions. Microsoft told us that in

134
00:08:41,320 --> 00:08:43,640
Speaker 1: the long term they want to stop buying these renewable

135
00:08:43,679 --> 00:08:46,240
Speaker 1: energy credits and move to just buying it through power

136
00:08:46,320 --> 00:08:49,479
Speaker 1: purchase agreements, but that still weighs away.

137
00:08:50,200 --> 00:08:53,960
Speaker 2: Right, But because of the data centers needed for AI,

138
00:08:54,400 --> 00:08:59,320
Speaker 2: Microsoft's emissions are still rising. Brad Smith seemed to suggest

139
00:08:59,559 --> 00:09:02,800
Speaker 2: that we shouldn't lose sight of a good AI can do.

140
00:09:03,520 --> 00:09:05,720
Speaker 3: So I think the right way to think about it

141
00:09:05,720 --> 00:09:09,680
Speaker 3: in the first instance is the call it benefits and costs,

142
00:09:09,760 --> 00:09:12,720
Speaker 3: or the good that we think AI will do for

143
00:09:12,800 --> 00:09:18,040
Speaker 3: the world versus the impact on the environment. And I

144
00:09:18,080 --> 00:09:21,440
Speaker 3: think we would clearly say it's going to do far

145
00:09:21,520 --> 00:09:24,680
Speaker 3: more good for the world, especially in terms of the

146
00:09:24,720 --> 00:09:26,800
Speaker 3: need and ability to bring it to everyone in the

147
00:09:26,840 --> 00:09:29,520
Speaker 3: world rather than have a group of AI haves and

148
00:09:29,559 --> 00:09:32,120
Speaker 3: AI have nots in different parts of the world.

149
00:09:33,080 --> 00:09:35,960
Speaker 2: So Microsoft is developing AI and using it for a

150
00:09:36,000 --> 00:09:38,520
Speaker 2: lot of different things. Can you talk me through what

151
00:09:38,559 --> 00:09:41,880
Speaker 2: we might think of as Microsoft's good AI, because from

152
00:09:41,880 --> 00:09:44,280
Speaker 2: what I've seen of chat GPT, I'd be happy to

153
00:09:44,280 --> 00:09:45,320
Speaker 2: be AI have not.

154
00:09:46,800 --> 00:09:49,240
Speaker 1: Yeah, this is the part of AI which is currently

155
00:09:49,440 --> 00:09:53,400
Speaker 1: sort of confusing because chat GPT is only one type

156
00:09:53,440 --> 00:09:56,480
Speaker 1: of AI, it's generative AI. There are other forms of

157
00:09:56,520 --> 00:10:00,280
Speaker 1: AI which can be used for doing climate goods, so

158
00:10:00,480 --> 00:10:02,960
Speaker 1: we know, for example, there are companies that are deploying

159
00:10:02,960 --> 00:10:05,600
Speaker 1: AI models to try and improve weather prediction.

160
00:10:06,000 --> 00:10:09,520
Speaker 3: Yeah, I would say AI is saving lives every day.

161
00:10:10,040 --> 00:10:15,400
Speaker 3: It's helping respond to natural disasters that are occurring more frequently,

162
00:10:15,440 --> 00:10:16,439
Speaker 3: so we shouldn't lose.

163
00:10:16,280 --> 00:10:19,599
Speaker 1: Sight of it that can really help. There are other examples

164
00:10:19,600 --> 00:10:23,040
Speaker 1: of using AI to find new materials that could help

165
00:10:23,120 --> 00:10:26,400
Speaker 1: us make better batteries, for example, or try and find

166
00:10:26,440 --> 00:10:30,760
Speaker 1: new materials to capture carbon. Those are examples where AI

167
00:10:31,360 --> 00:10:33,840
Speaker 1: has been put to use and will be increasingly put

168
00:10:33,880 --> 00:10:36,160
Speaker 1: to use, and it's something we want to explore on

169
00:10:36,200 --> 00:10:38,680
Speaker 1: the podcast. More so, we didn't get into all the

170
00:10:38,760 --> 00:10:41,440
Speaker 1: applications that AI can have for trying to solve some

171
00:10:41,480 --> 00:10:44,160
Speaker 1: of the climate problems we have, etc. But I think

172
00:10:44,360 --> 00:10:48,240
Speaker 1: he was telling us that just like computers and mobile

173
00:10:48,280 --> 00:10:51,320
Speaker 1: phones have done good in the world helping people connect

174
00:10:51,480 --> 00:10:55,160
Speaker 1: and make use of their own abilities to a better

175
00:10:55,400 --> 00:10:59,480
Speaker 1: extent to be more productive, that AI can do that too,

176
00:10:59,800 --> 00:11:01,560
Speaker 1: and that we should try and get this to as

177
00:11:01,600 --> 00:11:04,640
Speaker 1: many people in the world as possible, as quickly as possible.

178
00:11:05,120 --> 00:11:07,920
Speaker 1: But of course he would say that because he has

179
00:11:08,080 --> 00:11:10,240
Speaker 1: AI as a product and he would like to sell that.

180
00:11:10,920 --> 00:11:13,760
Speaker 2: And as we know, AI is also used to help

181
00:11:13,920 --> 00:11:17,200
Speaker 2: oil and gas companies find where to drill kind of

182
00:11:17,240 --> 00:11:23,000
Speaker 2: to expand fossil fuel production, and some people at Microsoft

183
00:11:23,040 --> 00:11:26,679
Speaker 2: aren't happy about the company's commitment to AI.

184
00:11:27,240 --> 00:11:29,920
Speaker 1: Yes, there is a group of ten thousand employees within

185
00:11:30,000 --> 00:11:34,280
Speaker 1: Microsoft who want to push Microsoft to deal with its

186
00:11:34,320 --> 00:11:37,800
Speaker 1: own emissions, but also to ensure that its AI is

187
00:11:37,800 --> 00:11:41,679
Speaker 1: not used to increase emissions. In fact, two former employees

188
00:11:41,800 --> 00:11:44,720
Speaker 1: wrote that they were deeply saddened to be so let

189
00:11:44,760 --> 00:11:47,960
Speaker 1: down by a company we love so much because they

190
00:11:47,960 --> 00:11:50,160
Speaker 1: didn't want Microsoft's AI to be used by oil and

191
00:11:50,200 --> 00:11:55,079
Speaker 1: gas companies for increasing extraction. So we asked Brad about that,

192
00:11:55,280 --> 00:11:58,160
Speaker 1: and he did have an answer. He basically said, we

193
00:11:58,520 --> 00:12:01,240
Speaker 1: have laid out our principle that we will only work

194
00:12:01,280 --> 00:12:03,800
Speaker 1: with energy companies that do have an net zero goal

195
00:12:04,000 --> 00:12:07,600
Speaker 1: and are on the climate pod, and that our AIUS

196
00:12:07,760 --> 00:12:09,319
Speaker 1: is towards the energy transition.

197
00:12:09,679 --> 00:12:11,480
Speaker 2: I want to ask you about something else Brad said

198
00:12:11,559 --> 00:12:15,280
Speaker 2: about Microsoft's long term impact. This goes back to the

199
00:12:15,320 --> 00:12:18,240
Speaker 2: point about the company's need for green cement and green

200
00:12:18,320 --> 00:12:22,440
Speaker 2: steel and green microchips. And what Brad basically said to

201
00:12:22,480 --> 00:12:25,360
Speaker 2: you and Dina was that when a gigantic tech company

202
00:12:25,400 --> 00:12:29,200
Speaker 2: like Microsoft throws its weight around, it can completely change

203
00:12:29,240 --> 00:12:34,760
Speaker 2: market realities and it can maybe even entirely jumpstart green industries.

204
00:12:36,000 --> 00:12:38,720
Speaker 2: Here's what Brad said on the question of green concrete.

205
00:12:38,880 --> 00:12:42,800
Speaker 3: I'm probably especially excited about concrete and steel in a

206
00:12:42,800 --> 00:12:44,560
Speaker 3: way that I never thought I would be working at

207
00:12:44,600 --> 00:12:48,640
Speaker 3: a tech company, mostly because I think we have an

208
00:12:48,720 --> 00:12:54,600
Speaker 3: opportunity to punch way above our weight in starting to

209
00:12:54,720 --> 00:13:00,319
Speaker 3: build new markets for greener steel and greener concrete. One

210
00:13:00,360 --> 00:13:03,640
Speaker 3: of our goals for this next year is to do

211
00:13:04,200 --> 00:13:08,360
Speaker 3: for concrete what we've done, or at least help to

212
00:13:08,400 --> 00:13:12,320
Speaker 3: do for electricity. By that, I mean there are certain

213
00:13:12,360 --> 00:13:15,560
Speaker 3: places where we sense the potential for a market to

214
00:13:15,640 --> 00:13:19,640
Speaker 3: come together to buy green concrete. It may not be

215
00:13:19,960 --> 00:13:23,720
Speaker 3: next door to our data center. But what we're trying

216
00:13:23,720 --> 00:13:28,120
Speaker 3: to do is use our resources to then go buy

217
00:13:28,320 --> 00:13:32,560
Speaker 3: green concrete I'll just say in Finland, and then be

218
00:13:32,640 --> 00:13:39,360
Speaker 3: able to credit that for carbon accounting purposes against a

219
00:13:39,440 --> 00:13:43,360
Speaker 3: data center that might be in southern Europe. And if

220
00:13:43,360 --> 00:13:47,200
Speaker 3: we can do that, we will stimulate demand for green concrete.

221
00:13:47,360 --> 00:13:49,400
Speaker 3: It will help it grow to scale. As it grows

222
00:13:49,440 --> 00:13:52,480
Speaker 3: to scale, the price of green concrete starts to come down.

223
00:13:53,280 --> 00:13:55,440
Speaker 3: And if you then think about the role of the

224
00:13:55,480 --> 00:13:59,120
Speaker 3: tech sector as a whole, because the tech sector as

225
00:13:59,160 --> 00:14:02,000
Speaker 3: a whole has and so committed to meeting these goals,

226
00:14:02,320 --> 00:14:07,120
Speaker 3: and we have the resources to pursue this goal, we

227
00:14:07,160 --> 00:14:08,240
Speaker 3: can make a difference.

228
00:14:09,320 --> 00:14:11,360
Speaker 2: What do you make at this point? Does this kind

229
00:14:11,360 --> 00:14:14,120
Speaker 2: of influence make up for the company's rising emissions?

230
00:14:14,679 --> 00:14:18,280
Speaker 1: So let's look at the numbers. Microsoft's emissions at sixteen

231
00:14:18,320 --> 00:14:21,760
Speaker 1: or seventeen million tons per year are quite small relative

232
00:14:21,840 --> 00:14:24,480
Speaker 1: to say the global emissions which are thirty eight billion

233
00:14:24,480 --> 00:14:28,840
Speaker 1: tons last year, or even other large companies like Exomobil

234
00:14:28,920 --> 00:14:33,000
Speaker 1: which was five hundred million tons. And yes, it's true

235
00:14:33,400 --> 00:14:37,320
Speaker 1: that as Microsoft pursues green steel, green cement, and even

236
00:14:37,320 --> 00:14:41,800
Speaker 1: carbon removals, it is making an expensive product slightly cheaper

237
00:14:42,040 --> 00:14:44,640
Speaker 1: every time it buys it, and that will make it

238
00:14:44,960 --> 00:14:48,640
Speaker 1: easier for other companies to meet that goal. But that

239
00:14:48,640 --> 00:14:51,760
Speaker 1: does not mean Microsoft should get any credit for it

240
00:14:51,840 --> 00:14:53,280
Speaker 1: in its own carbon footprint.

241
00:14:54,240 --> 00:14:58,960
Speaker 2: So basically, the companies arguing for another kind of long

242
00:14:59,080 --> 00:15:03,160
Speaker 2: term thinking, don't just look at our emissions, don't just

243
00:15:03,200 --> 00:15:08,240
Speaker 2: look at our carbon footprint, look at this other intangible

244
00:15:08,280 --> 00:15:12,360
Speaker 2: effect we are having on industries adjacent to ask.

245
00:15:12,320 --> 00:15:14,960
Speaker 1: Right, and you should look at that intangible benefit, But

246
00:15:15,120 --> 00:15:18,280
Speaker 1: you cannot ignore the fact that the company's emissions right

247
00:15:18,360 --> 00:15:22,200
Speaker 1: now are rising. Microsoft has to both reduce its own

248
00:15:22,240 --> 00:15:26,160
Speaker 1: emissions while it may take some reputational benefit from making

249
00:15:26,240 --> 00:15:27,720
Speaker 1: green solutions cheaper.

250
00:15:34,080 --> 00:15:36,880
Speaker 2: After the break, does Microsoft need to choose between its

251
00:15:36,880 --> 00:15:40,680
Speaker 2: net zero commitments and its AI ambitions and has it

252
00:15:40,800 --> 00:15:44,040
Speaker 2: already made that decision? Aksha and Dina put the question

253
00:15:44,120 --> 00:15:47,360
Speaker 2: to Brad and we discuss his answer. And if you're

254
00:15:47,400 --> 00:15:50,400
Speaker 2: enjoying this episode, please do rate and review zero on

255
00:15:50,520 --> 00:16:02,560
Speaker 2: Apple or Spotify. Microsoft is clearly going to keep pushing

256
00:16:02,600 --> 00:16:06,000
Speaker 2: on AI and that means more resources, more electricity, and

257
00:16:06,040 --> 00:16:11,600
Speaker 2: thus probably more emissions. So what happens to its climate target?

258
00:16:11,920 --> 00:16:15,000
Speaker 2: You put that question to Brad Smith. He said, they're

259
00:16:15,040 --> 00:16:16,720
Speaker 2: not abandoning those pledges.

260
00:16:16,960 --> 00:16:20,240
Speaker 4: I don't want to give up now. Can I promise

261
00:16:20,360 --> 00:16:23,560
Speaker 4: that we'll get to this moon by twenty thirty? We

262
00:16:24,160 --> 00:16:27,800
Speaker 4: keep making more progress. The truth is we made so

263
00:16:27,960 --> 00:16:31,240
Speaker 4: much progress at the strategy level over the past year.

264
00:16:31,400 --> 00:16:34,280
Speaker 4: That's the real story to me inside Microsoft.

265
00:16:34,800 --> 00:16:35,960
Speaker 2: Can he have it both ways?

266
00:16:36,920 --> 00:16:40,160
Speaker 1: It's really tough. This is why the story we wrote

267
00:16:40,560 --> 00:16:43,200
Speaker 1: kind of en viral, because there is no good answer

268
00:16:43,480 --> 00:16:47,520
Speaker 1: right now. It's clear, even if your electricity was really clean,

269
00:16:47,760 --> 00:16:50,600
Speaker 1: fully clean one hundred percent renewables twenty four to seven,

270
00:16:51,040 --> 00:16:55,400
Speaker 1: that are still building data centers with steel and cement

271
00:16:55,640 --> 00:16:58,720
Speaker 1: and these chips. That is very common intensive and that's

272
00:16:58,720 --> 00:17:01,880
Speaker 1: going to increase emissions in the short term. So we

273
00:17:01,920 --> 00:17:05,000
Speaker 1: ask Bratsmith what exactly is the plan here? And the

274
00:17:05,040 --> 00:17:11,280
Speaker 1: answer is basically everything possible. Microsoft is investing in existing

275
00:17:11,280 --> 00:17:15,560
Speaker 1: technologies like nuclear power and geothermal, but also frontier technologies

276
00:17:15,600 --> 00:17:18,320
Speaker 1: like nuclear fusion, which we haven't even made work yet,

277
00:17:18,320 --> 00:17:21,960
Speaker 1: but maybe we will in the next decade, and of course,

278
00:17:22,359 --> 00:17:26,560
Speaker 1: if nothing else works, Microsoft is also investing in carbon removal.

279
00:17:26,600 --> 00:17:29,439
Speaker 1: It is the largest purchaser of carbon removal credits in

280
00:17:29,480 --> 00:17:33,560
Speaker 1: the world. And if say, by twenty twenty nine, emissions

281
00:17:33,640 --> 00:17:38,159
Speaker 1: are still high, Bratsmith said, carbon removal is always on

282
00:17:38,200 --> 00:17:41,360
Speaker 1: the table, But of course it is a very expensive

283
00:17:41,560 --> 00:17:44,080
Speaker 1: purchase to try and remove carbon dioxide from the air.

284
00:17:44,359 --> 00:17:47,280
Speaker 1: It would be much cheaper to not emit that carbon

285
00:17:47,280 --> 00:17:48,600
Speaker 1: dioxide in the first place.

286
00:17:50,480 --> 00:17:54,720
Speaker 2: You mentioned that last year Microsoft consume more energy than Slovenia,

287
00:17:55,760 --> 00:17:59,359
Speaker 2: but the same could be said for Google or Meta.

288
00:17:59,440 --> 00:18:02,920
Speaker 2: Probably right, This is not only a problem for Microsoft,

289
00:18:03,040 --> 00:18:06,040
Speaker 2: but is it disproportionately a problem for Microsoft.

290
00:18:06,560 --> 00:18:09,000
Speaker 1: No, it's a problem for all of the tech companies

291
00:18:09,080 --> 00:18:12,199
Speaker 1: trying to get after AI. And I'm sure as we

292
00:18:12,240 --> 00:18:15,320
Speaker 1: see the sustainability reports from Google and metacome for this

293
00:18:15,440 --> 00:18:18,040
Speaker 1: year and it's likely to show an increase in emissions

294
00:18:18,280 --> 00:18:22,439
Speaker 1: not very different from what Microsoft has. So it is

295
00:18:22,640 --> 00:18:25,560
Speaker 1: a problem that tech companies need to find an answer to.

296
00:18:25,920 --> 00:18:29,359
Speaker 1: Because we also ask Google, Meta and Amazon to respond

297
00:18:29,359 --> 00:18:32,639
Speaker 1: to our questions on their own emissions increase, and they

298
00:18:32,680 --> 00:18:35,960
Speaker 1: all said, we are not backing away from our climate targets.

299
00:18:35,960 --> 00:18:38,560
Speaker 1: Which in most cases are either carbon neutrality by twenty

300
00:18:38,640 --> 00:18:41,320
Speaker 1: thirty or in Amazon's case, by twenty forty.

301
00:18:42,160 --> 00:18:44,520
Speaker 2: Because I think most listeners know you've written a book

302
00:18:44,600 --> 00:18:48,280
Speaker 2: called Climate Capitalism, which is about the idea that rather

303
00:18:48,320 --> 00:18:52,560
Speaker 2: than trying to entirely remake our economic system overnight, it's

304
00:18:52,640 --> 00:18:56,080
Speaker 2: more fruitful to focus on getting businesses and governments to

305
00:18:56,160 --> 00:18:59,760
Speaker 2: respond to the reality of climate change within the capitalist

306
00:19:00,040 --> 00:19:03,040
Speaker 2: framework that we already live in. When you see a

307
00:19:03,080 --> 00:19:06,320
Speaker 2: big company like Microsoft move in this direction on its

308
00:19:06,320 --> 00:19:11,760
Speaker 2: own climate commitments, do you worry about capitalism's incentives just

309
00:19:11,800 --> 00:19:15,080
Speaker 2: being too much at odds with green goals.

310
00:19:16,200 --> 00:19:18,879
Speaker 1: Well, it's exactly the thesis of the book, which is

311
00:19:19,280 --> 00:19:24,320
Speaker 1: climate change has to modify how capitalism works. Microsoft isn't

312
00:19:24,320 --> 00:19:26,680
Speaker 1: going to back off from AI if it's got profits

313
00:19:26,720 --> 00:19:32,280
Speaker 1: in there. But as long as governments and people want

314
00:19:32,320 --> 00:19:35,600
Speaker 1: climate targets to be hit, Microsoft will have to find

315
00:19:35,760 --> 00:19:39,480
Speaker 1: a way to make those climate targets work. That may

316
00:19:39,600 --> 00:19:44,239
Speaker 1: mean a lot more investment in greener technologies, or if

317
00:19:44,320 --> 00:19:47,359
Speaker 1: it has to pay a really premium price and remove

318
00:19:47,400 --> 00:19:50,920
Speaker 1: all that carbon dioxide, so be it. But those targets

319
00:19:50,960 --> 00:19:52,920
Speaker 1: need to be set by governments so that we don't

320
00:19:52,960 --> 00:19:58,200
Speaker 1: rely on just voluntary commitments from companies like Microsoft doing

321
00:19:58,280 --> 00:20:01,680
Speaker 1: the right thing on climate. It is possible to hit

322
00:20:01,720 --> 00:20:04,639
Speaker 1: these goals with the technologies that we have and the

323
00:20:04,680 --> 00:20:08,960
Speaker 1: technologies that we do need to invent, because that modeling

324
00:20:09,000 --> 00:20:12,080
Speaker 1: has been done by very reputable organizations like the International

325
00:20:12,160 --> 00:20:16,840
Speaker 1: Energy Agency, Bloomberg and EF the Intergovernmental Panel on Climate Change.

326
00:20:17,359 --> 00:20:20,359
Speaker 1: What is needed is to actually do the work, and

327
00:20:20,440 --> 00:20:23,840
Speaker 1: Microsoft is showing that that work is not easy, but

328
00:20:24,160 --> 00:20:26,840
Speaker 1: it is something you need to be committed to because

329
00:20:26,880 --> 00:20:30,200
Speaker 1: climate is a problem everybody has to deal with. In fact,

330
00:20:30,400 --> 00:20:33,680
Speaker 1: Bradsmith ended our interview by saying exactly that.

331
00:20:34,200 --> 00:20:36,520
Speaker 3: You've got to be willing to invest and pay for it.

332
00:20:36,920 --> 00:20:40,159
Speaker 3: You've got to be willing to persist even when the

333
00:20:40,280 --> 00:20:45,800
Speaker 3: challenge seems so hard. It's all about incredibly sophisticated execution.

334
00:20:46,320 --> 00:20:49,160
Speaker 3: But we need to retain a vision as well, and

335
00:20:49,240 --> 00:20:53,560
Speaker 3: that starts with the recognition that a problem that humanity

336
00:20:53,640 --> 00:20:57,560
Speaker 3: created is also a problem that humanity can solve, and

337
00:20:57,680 --> 00:21:01,280
Speaker 3: let's never lose sight of that. Yeah, it's not like

338
00:21:01,359 --> 00:21:03,960
Speaker 3: this is running the Kentucky Derby, where you get up

339
00:21:04,119 --> 00:21:06,560
Speaker 3: on Monday and the race is over and there's no

340
00:21:06,720 --> 00:21:09,960
Speaker 3: race to run. This is the race of the century,

341
00:21:10,000 --> 00:21:12,680
Speaker 3: and it needs to be run every day throughout the century.

342
00:21:12,760 --> 00:21:18,040
Speaker 3: That's our perspective on it. A THEAI race is more important.

343
00:21:19,000 --> 00:21:21,639
Speaker 1: I don't want to make some people. There are definitely

344
00:21:21,640 --> 00:21:25,080
Speaker 1: people that think that climates should take precedence over rapid

345
00:21:25,080 --> 00:21:26,600
Speaker 1: deployment of AI.

346
00:21:27,440 --> 00:21:30,800
Speaker 3: Yeah, but I think the question you should ask is

347
00:21:30,840 --> 00:21:35,560
Speaker 3: what percentage of global steel consumption in twenty thirty is

348
00:21:35,600 --> 00:21:38,200
Speaker 3: going to be devoted to the construction of data centers

349
00:21:38,280 --> 00:21:42,920
Speaker 3: versus everything else that steel goes into. What percentage of

350
00:21:43,080 --> 00:21:47,560
Speaker 3: the world's concrete consumption is going to go to data centers?

351
00:21:47,880 --> 00:21:51,800
Speaker 3: And then ask yourself this. Every time you go to

352
00:21:51,880 --> 00:21:56,080
Speaker 3: a meeting, every time you see a new innovation, whether

353
00:21:56,119 --> 00:22:01,400
Speaker 3: it's in the technology changes or the market innovation, I'm

354
00:22:01,400 --> 00:22:05,280
Speaker 3: willing to bet you will find someone from Microsoft very

355
00:22:05,400 --> 00:22:09,320
Speaker 3: likely to be there. The day we should worry is

356
00:22:09,359 --> 00:22:11,119
Speaker 3: the day we stop showing up.

357
00:22:11,720 --> 00:22:13,760
Speaker 2: Okay, what do you make of Brad's answer? Do you

358
00:22:13,800 --> 00:22:16,080
Speaker 2: really believe Microsoft's reasoning on this point?

359
00:22:16,280 --> 00:22:20,840
Speaker 1: My takeaway from following net zero from the start to

360
00:22:21,040 --> 00:22:24,920
Speaker 1: company setting these goals, to Microsoft, which has been ahead

361
00:22:24,960 --> 00:22:27,840
Speaker 1: in the game at least in setting goals, is that

362
00:22:28,359 --> 00:22:33,040
Speaker 1: it is hard. Nobody has the perfect answer. But if

363
00:22:33,080 --> 00:22:36,000
Speaker 1: anybody can try and answer that question, it's a company

364
00:22:36,119 --> 00:22:38,520
Speaker 1: with money and a company that is committed to try

365
00:22:38,560 --> 00:22:41,600
Speaker 1: and reach these goals. So I give Microsoft the credit

366
00:22:41,680 --> 00:22:43,880
Speaker 1: that it is still trying to meet these goals, as

367
00:22:44,000 --> 00:22:47,520
Speaker 1: hard as they might be. But Microsoft is the world's

368
00:22:47,520 --> 00:22:51,439
Speaker 1: most valuable company. And if Microsoft can't reach these goals

369
00:22:51,760 --> 00:22:55,280
Speaker 1: or isn't taking them as seriously, then can we expect

370
00:22:55,320 --> 00:22:58,720
Speaker 1: any other company to do so? As a climate reporter.

371
00:22:59,040 --> 00:23:04,120
Speaker 1: That is the bigger question to me, and it remains unanswered. Lukshet,

372
00:23:04,160 --> 00:23:06,080
Speaker 1: thank you, Thank you, This is fun.

373
00:23:14,320 --> 00:23:17,000
Speaker 2: Thank you for listening to Zero. If you liked this episode,

374
00:23:17,080 --> 00:23:19,040
Speaker 2: please take a moment to rate or review the show

375
00:23:19,080 --> 00:23:22,760
Speaker 2: on Apple Podcasts and Spotify, Share this episode with a

376
00:23:22,760 --> 00:23:26,000
Speaker 2: friend or with a Microsoft Windows user. You can get

377
00:23:26,000 --> 00:23:30,240
Speaker 2: in touch at zero pod at Bloomberg dot Net. Zero's

378
00:23:30,280 --> 00:23:34,400
Speaker 2: producer is me Mightily Raw. Our theme music is composed

379
00:23:34,440 --> 00:23:38,600
Speaker 2: by Wonderley Special thanks to Kiara Bindram, Dina Bass and

380
00:23:38,640 --> 00:23:40,640
Speaker 2: Alicia Clanson. Back soon.