1
00:00:02,400 --> 00:00:06,760
Speaker 1: Bloomberg Audio Studios, podcasts, radio news.

2
00:00:06,960 --> 00:00:10,360
Speaker 2: Amazon also announcing half a billion dollar investment in nuclear

3
00:00:10,400 --> 00:00:15,000
Speaker 2: clean energy. Here to discuss it, Matt Goarman, CEO of AWS,

4
00:00:15,040 --> 00:00:18,400
Speaker 2: and Matt welcome back to Blimberg Technology. It's so interesting

5
00:00:18,880 --> 00:00:21,520
Speaker 2: to see that. Thank you to see the hyperscaler is

6
00:00:21,880 --> 00:00:26,120
Speaker 2: so focused on this. I think we start with y SMRs, right,

7
00:00:26,239 --> 00:00:28,479
Speaker 2: that's a very specific avenue to go down.

8
00:00:29,520 --> 00:00:29,720
Speaker 1: Yeah.

9
00:00:29,720 --> 00:00:32,040
Speaker 3: I think SMRs are a great new technology and they're

10
00:00:32,040 --> 00:00:35,360
Speaker 3: an important part of the energy landscape that we're looking at.

11
00:00:35,760 --> 00:00:38,280
Speaker 3: We made a five hundred million dollar investment in a

12
00:00:38,280 --> 00:00:41,520
Speaker 3: company called x Energy that are really I think they're

13
00:00:41,600 --> 00:00:43,880
Speaker 3: one of the leaders in this space, and they're really

14
00:00:43,920 --> 00:00:45,559
Speaker 3: pushing on the technology SMRs.

15
00:00:46,320 --> 00:00:47,680
Speaker 1: They have a couple of advantages.

16
00:00:48,080 --> 00:00:51,440
Speaker 3: They're able to be constructed in a smaller footprint, which

17
00:00:51,440 --> 00:00:52,320
Speaker 3: means you can put them.

18
00:00:52,240 --> 00:00:53,760
Speaker 1: Closer to where the energy needs to be.

19
00:00:54,400 --> 00:00:58,280
Speaker 3: They're also incredibly safe and easily manufactural, and so they're

20
00:00:58,440 --> 00:01:02,560
Speaker 3: very promising technology and we're quite excited about our investments

21
00:01:02,600 --> 00:01:04,880
Speaker 3: in this space and the partnerships that we have coming up.

22
00:01:06,240 --> 00:01:09,880
Speaker 4: Small modular reactors SMRs for those getting in the acronym lingo.

23
00:01:10,000 --> 00:01:13,360
Speaker 4: Matt it's great to have you. What about other deals,

24
00:01:13,440 --> 00:01:15,480
Speaker 4: I mean, is this the first of many?

25
00:01:15,840 --> 00:01:18,760
Speaker 3: Yeah, And so a couple of details on the things

26
00:01:18,760 --> 00:01:21,319
Speaker 3: that we announced yesterday. One, we have this investment with

27
00:01:21,720 --> 00:01:23,680
Speaker 3: x Energy, who's a producer of these. But we also

28
00:01:23,760 --> 00:01:27,160
Speaker 3: have announced two partnerships with utilities, and so we're actually

29
00:01:27,160 --> 00:01:30,959
Speaker 3: well down the path, ongoing and actually implementing some of

30
00:01:30,959 --> 00:01:34,760
Speaker 3: these deals. So we have a partnership with Dominion in

31
00:01:34,880 --> 00:01:39,160
Speaker 3: Virginia to go implement an SMR as part of their

32
00:01:39,280 --> 00:01:42,480
Speaker 3: energy footprint here in Virginia, and with Energy Northwest in

33
00:01:42,600 --> 00:01:45,920
Speaker 3: Washington and Oregon, and our goal is to deliver more

34
00:01:45,959 --> 00:01:48,800
Speaker 3: than five gigawatts of nuclear energy. We think that it's

35
00:01:48,840 --> 00:01:51,920
Speaker 3: a fantastic opportunity for us to keep pushing on the

36
00:01:52,000 --> 00:01:54,840
Speaker 3: carbon energy footprint that's needed as we continue to scale

37
00:01:54,880 --> 00:01:55,560
Speaker 3: around the country.

38
00:01:56,200 --> 00:01:58,240
Speaker 2: Matt, if you're well down the path, and based on

39
00:01:58,280 --> 00:02:00,560
Speaker 2: what you just said, do you have a clear line

40
00:02:00,600 --> 00:02:03,720
Speaker 2: of sight on what the dollar per mega what hour

41
00:02:03,800 --> 00:02:06,520
Speaker 2: is going to be out of those facilities.

42
00:02:07,640 --> 00:02:08,560
Speaker 1: We don't know that yet.

43
00:02:09,000 --> 00:02:11,919
Speaker 3: We're still we are still early and we're still learning

44
00:02:12,360 --> 00:02:15,679
Speaker 3: many of these technologies for these reactors or probably won't

45
00:02:15,680 --> 00:02:17,919
Speaker 3: come online until twenty thirty, so.

46
00:02:17,880 --> 00:02:20,760
Speaker 1: We'll still be learning there, but we feel that as

47
00:02:20,800 --> 00:02:21,480
Speaker 1: we scale and in.

48
00:02:21,400 --> 00:02:24,440
Speaker 3: The fullness of time, these energy sources are going to

49
00:02:24,440 --> 00:02:26,520
Speaker 3: be quite competitive from a cost perspective, and it's one

50
00:02:26,560 --> 00:02:28,680
Speaker 3: of the reasons that we're so excited about them.

51
00:02:29,120 --> 00:02:31,920
Speaker 2: Would you bring back a full nuclear plant like three

52
00:02:31,960 --> 00:02:35,280
Speaker 2: Mile Island in Pennsylvania And if so, which.

53
00:02:35,080 --> 00:02:39,040
Speaker 3: One, well that I wouldn't have done that one in particular.

54
00:02:39,080 --> 00:02:41,639
Speaker 3: But we actually have a partnership with Talent where we're

55
00:02:41,639 --> 00:02:45,240
Speaker 3: bringing more energy online in Pennsylvania and that's we're actually

56
00:02:45,320 --> 00:02:48,040
Speaker 3: the very first company to announce a deal with Talent

57
00:02:48,760 --> 00:02:50,760
Speaker 3: several months ago. And so I think that some of

58
00:02:50,800 --> 00:02:55,040
Speaker 3: these large nuclear plants are also excellent source of energy

59
00:02:55,040 --> 00:02:56,919
Speaker 3: and they're an important part of that. And as I said,

60
00:02:57,120 --> 00:02:59,480
Speaker 3: I think small modular reactors are going to be a

61
00:02:59,520 --> 00:03:02,920
Speaker 3: component of what we need for a carbon free energy world,

62
00:03:03,040 --> 00:03:05,560
Speaker 3: but large nuclear plants are going to be an important

63
00:03:05,560 --> 00:03:09,000
Speaker 3: part of that too. And we've said that we'll take

64
00:03:09,040 --> 00:03:11,799
Speaker 3: as much as almost a gigawot of power that we're

65
00:03:11,800 --> 00:03:15,200
Speaker 3: bringing back online from this large plant in Pennsylvania, where

66
00:03:15,240 --> 00:03:17,560
Speaker 3: we'll also be building data centers nearby. So I think

67
00:03:17,560 --> 00:03:20,640
Speaker 3: it's all an important part of the portfolio as we

68
00:03:20,680 --> 00:03:22,120
Speaker 3: move to decarbonize the energy.

69
00:03:22,480 --> 00:03:24,920
Speaker 4: Let's talk about the decarbonization because back when you announce

70
00:03:24,960 --> 00:03:27,760
Speaker 4: that deal in March with Susquehanna, you talked about how

71
00:03:27,800 --> 00:03:30,880
Speaker 4: still the aim is to be by twenty twenty five

72
00:03:31,120 --> 00:03:34,040
Speaker 4: one hundred percent renewable energy. Is that attainable with a

73
00:03:34,080 --> 00:03:36,760
Speaker 4: sudden desire to have AI and the energy that goes

74
00:03:36,800 --> 00:03:37,080
Speaker 4: with it.

75
00:03:38,080 --> 00:03:40,280
Speaker 3: Yeah, In fact, we announced that this year we hit

76
00:03:40,600 --> 00:03:43,600
Speaker 3: one hundred percent carbon free energy in our systems, and

77
00:03:43,640 --> 00:03:45,880
Speaker 3: I think that now our goal is how do we

78
00:03:46,000 --> 00:03:48,119
Speaker 3: as we look at all of the energy demands that

79
00:03:48,240 --> 00:03:51,480
Speaker 3: the business and the world frankly needs, how do we

80
00:03:51,520 --> 00:03:55,240
Speaker 3: get to a continued path on growing carbon free energy.

81
00:03:55,240 --> 00:03:57,120
Speaker 3: And so that's what a lot of these deals are about,

82
00:03:57,480 --> 00:03:58,280
Speaker 3: is for the future.

83
00:03:58,320 --> 00:03:59,040
Speaker 1: As we look.

84
00:03:58,880 --> 00:04:01,160
Speaker 3: At power that we're going to need in twenty five,

85
00:04:01,240 --> 00:04:04,320
Speaker 3: twenty seven, thirty, twenty forty, how do we make sure

86
00:04:04,320 --> 00:04:07,480
Speaker 3: that we have a ramp of power so that as

87
00:04:07,520 --> 00:04:10,520
Speaker 3: the energy needs grow, whether they're from electric vehicles or

88
00:04:10,560 --> 00:04:14,720
Speaker 3: from companies increasingly digitizing, or from generative AI, that we

89
00:04:14,760 --> 00:04:16,599
Speaker 3: have enough carbon free energy to support them. And so

90
00:04:16,640 --> 00:04:18,359
Speaker 3: it's really a long term view that we have and

91
00:04:18,400 --> 00:04:21,000
Speaker 3: part of this is on that long term view, how

92
00:04:21,040 --> 00:04:22,000
Speaker 3: we continue.

93
00:04:21,680 --> 00:04:23,159
Speaker 1: To have enough power well into the future.

94
00:04:23,320 --> 00:04:25,000
Speaker 4: Let's talk about the short term view if we come

95
00:04:25,000 --> 00:04:27,640
Speaker 4: for a minute, Matt, though, because right here, right now,

96
00:04:28,000 --> 00:04:30,839
Speaker 4: do you have the power that is necessary from the

97
00:04:30,920 --> 00:04:33,039
Speaker 4: places you want to be getting it from.

98
00:04:33,960 --> 00:04:35,560
Speaker 1: We work really hard to make sure that we do.

99
00:04:36,040 --> 00:04:40,600
Speaker 3: Our businesses continuing to grow rapidly, and we're adding new

100
00:04:40,640 --> 00:04:43,080
Speaker 3: power and new data centers all of the time, and

101
00:04:43,160 --> 00:04:46,159
Speaker 3: so we continue we look further out. We have power

102
00:04:46,160 --> 00:04:48,240
Speaker 3: that we're adding in twenty four, power that we're adding

103
00:04:48,279 --> 00:04:51,440
Speaker 3: in twenty twenty five, and continue out. And as you

104
00:04:52,520 --> 00:04:55,560
Speaker 3: may have heard, over the last five years, Amazon has

105
00:04:55,600 --> 00:04:58,479
Speaker 3: been the single biggest purchaser of renewable power contracts in

106
00:04:58,520 --> 00:05:00,520
Speaker 3: the world every year for the last five years, and

107
00:05:00,560 --> 00:05:01,600
Speaker 3: so we.

108
00:05:01,520 --> 00:05:02,320
Speaker 1: Continue to do that.

109
00:05:02,360 --> 00:05:04,680
Speaker 3: We continue to add power to our footprint, and we

110
00:05:04,760 --> 00:05:07,200
Speaker 3: plan multiple years out because these projects that we go

111
00:05:07,279 --> 00:05:09,840
Speaker 3: and launch they take many years to come online, and

112
00:05:09,880 --> 00:05:13,760
Speaker 3: so we have lots of plans and we continue to

113
00:05:14,080 --> 00:05:15,680
Speaker 3: grow as the needs of our customers grow.

114
00:05:16,720 --> 00:05:19,160
Speaker 2: Matt, Will you have any Blackwell online by the end

115
00:05:19,200 --> 00:05:19,680
Speaker 2: of this year?

116
00:05:22,160 --> 00:05:22,240
Speaker 1: No?

117
00:05:22,440 --> 00:05:26,720
Speaker 3: Unfortunately, in videos, timelines have shifted on that and so

118
00:05:26,880 --> 00:05:28,320
Speaker 3: and so. We have samples in the lab and we're

119
00:05:28,320 --> 00:05:31,159
Speaker 3: working on them right now, but production qualities are production

120
00:05:31,279 --> 00:05:33,719
Speaker 3: quantities we expect to be early next year.

121
00:05:34,400 --> 00:05:35,080
Speaker 1: We work very.

122
00:05:34,960 --> 00:05:38,640
Speaker 3: Closely with in Vidio on GPUs and super excited about

123
00:05:38,680 --> 00:05:42,200
Speaker 3: that launch. And in Video's own large research cluster is

124
00:05:42,200 --> 00:05:45,240
Speaker 3: actually launching in as so we'll have the first production

125
00:05:45,320 --> 00:05:48,520
Speaker 3: qualities will be in AWS when they're ready, But so

126
00:05:48,600 --> 00:05:50,839
Speaker 3: right now we're working with them and we have early samples,

127
00:05:50,880 --> 00:05:52,680
Speaker 3: but as they get their yields up from the respind

128
00:05:52,839 --> 00:05:55,960
Speaker 3: of the Blackwell chip, it'll probably be early next year

129
00:05:55,960 --> 00:05:59,200
Speaker 3: before we really get production samples in real volume.

130
00:05:59,480 --> 00:06:03,560
Speaker 4: Another partnership then Anthropics work on the chips. How is

131
00:06:03,600 --> 00:06:04,039
Speaker 4: that going?

132
00:06:04,120 --> 00:06:06,359
Speaker 1: Briefly fantastic.

133
00:06:06,400 --> 00:06:11,360
Speaker 3: Anthropic continues to deliver fantastic new models. Their Sonnet three

134
00:06:11,400 --> 00:06:14,080
Speaker 3: to five model is one of the most powerful ones

135
00:06:14,080 --> 00:06:16,840
Speaker 3: that is out there today and customers really enjoying using

136
00:06:16,839 --> 00:06:19,600
Speaker 3: it on a variety of tasks. I think the reasoning

137
00:06:20,279 --> 00:06:23,400
Speaker 3: is industry and world class and a bunch of the

138
00:06:23,400 --> 00:06:26,320
Speaker 3: capabilities that enterprises are able to deliver, and the value

139
00:06:26,320 --> 00:06:29,560
Speaker 3: they get from those models is really it's fun to

140
00:06:29,640 --> 00:06:34,080
Speaker 3: watch and we really enjoy that partnership and we continue

141
00:06:34,080 --> 00:06:36,120
Speaker 3: to expect it to go for a long time.

142
00:06:36,360 --> 00:06:40,240
Speaker 4: Matt Carman, AWS CEO, we appreciate your time. Thank you,