1
00:00:00,200 --> 00:00:04,640
Big tech companies are driving a
nuclear resurgence, bringing 

2
00:00:04,640 --> 00:00:08,119
shuttered facilities like 3 Mile
Island back online and 

3
00:00:08,280 --> 00:00:11,960
announcing plans to build small 
modular reactors. 

4
00:00:12,320 --> 00:00:15,480
The new data centres being built
today require so much 

5
00:00:15,480 --> 00:00:19,480
electricity that national power 
grids can't keep up with them. 

6
00:00:19,800 --> 00:00:23,240
Tech firms have worked out that 
they need to source their own 

7
00:00:23,240 --> 00:00:26,200
power if they're going to build 
the kind of data centres that 

8
00:00:26,200 --> 00:00:28,640
will make them competitive in 
AI. 

9
00:00:29,120 --> 00:00:32,800
The amount of power generated 
globally by nuclear power plants

10
00:00:33,000 --> 00:00:38,200
peaked in 2006, with new plant 
construction stagnating in the 

11
00:00:38,200 --> 00:00:42,320
1990s. 
The 2011 Fukushima disaster in 

12
00:00:42,320 --> 00:00:46,000
Japan led to cancelled projects 
around the world. 

13
00:00:46,200 --> 00:00:50,240
In Germany, Angela Merkel, a 
trained physicist, announced 

14
00:00:50,240 --> 00:00:54,280
that Germany would phase out 
nuclear power, and Siemens 

15
00:00:54,280 --> 00:00:57,120
announced that they would 
withdraw entirely from the 

16
00:00:57,120 --> 00:01:00,200
nuclear industry to focus on 
renewables. 

17
00:01:00,560 --> 00:01:04,360
For a while, it seemed that Iran
and North Korea were the only 

18
00:01:04,360 --> 00:01:08,760
countries with a suspicious 
interest in building new nuclear

19
00:01:08,760 --> 00:01:11,440
power stations. 
In most of the rest of the 

20
00:01:11,440 --> 00:01:15,280
world, projects were delayed or 
cancelled due to public safety 

21
00:01:15,280 --> 00:01:20,280
concerns, regulatory hurdles, 
financing obstacles and fuel 

22
00:01:20,280 --> 00:01:23,240
bottlenecks. 
It's not just the fear of 

23
00:01:23,240 --> 00:01:26,440
disaster that slowed down 
nuclear energy, however. 

24
00:01:26,760 --> 00:01:30,280
The global demand for 
electricity grew by just two 

25
00:01:30,280 --> 00:01:35,840
point 2% in 2023, which was 
slightly less than the 2.4% 

26
00:01:35,840 --> 00:01:39,440
growth seen the prior year. 
While there had been strong 

27
00:01:39,440 --> 00:01:43,440
growth in China, India, and the 
developing world, advanced 

28
00:01:43,440 --> 00:01:47,480
economy showed substantial 
declines in their need for new 

29
00:01:47,480 --> 00:01:50,920
power stations. 
For much of the 20th century, 

30
00:01:51,040 --> 00:01:55,440
America's electricity demand had
grown steadily and utilities 

31
00:01:55,440 --> 00:01:59,080
built power stations to keep up 
with the growing demand. 

32
00:01:59,400 --> 00:02:03,840
Starting in the mid 2000's, the 
demand for additional capacity 

33
00:02:03,840 --> 00:02:07,600
in the developed world flattened
out while the US economy and 

34
00:02:07,600 --> 00:02:11,800
population continued to grow and
while we continued to add new 

35
00:02:11,800 --> 00:02:16,480
electrical devices to our homes,
factories, light bulbs, TV's, 

36
00:02:16,480 --> 00:02:20,960
computers and home appliances 
all became more and more energy 

37
00:02:20,960 --> 00:02:24,840
efficient, meaning that there 
was no real need to build new 

38
00:02:24,840 --> 00:02:28,880
generation capacity. 
The growth of data centers over 

39
00:02:28,880 --> 00:02:31,960
the last few years, driven by 
working from home, 

40
00:02:32,080 --> 00:02:36,640
cryptocurrency mining and now 
AI, boosted demand for 

41
00:02:36,640 --> 00:02:39,680
electricity. 
As a side note, working from 

42
00:02:39,680 --> 00:02:44,120
home which increased the need 
for data centres also reduced 

43
00:02:44,120 --> 00:02:48,200
the amount of power being used 
overall as the server racks at 

44
00:02:48,200 --> 00:02:51,760
data centres are more energy 
efficient than the office based 

45
00:02:51,760 --> 00:02:54,840
servers which businesses had 
relied on in the past. 

46
00:02:55,240 --> 00:02:59,040
Other changes however, like the 
roll out of electric vehicles in

47
00:02:59,040 --> 00:03:02,440
recent years and plans to 
reassure the semiconductor 

48
00:03:02,440 --> 00:03:06,320
industry, further boost the 
demand for electricity in 

49
00:03:06,320 --> 00:03:09,680
advanced economies. 
Big tech companies like 

50
00:03:09,680 --> 00:03:14,440
Microsoft, Amazon and Google are
driving the nuclear revival as 

51
00:03:14,440 --> 00:03:18,400
to get new power hungry data 
centres built in the United 

52
00:03:18,400 --> 00:03:21,760
States and Europe, they have to 
solve the problem of power 

53
00:03:21,760 --> 00:03:25,160
generation. 
Their net 0 pledges mean that 

54
00:03:25,160 --> 00:03:28,280
the sources of power that 
they've pledged to use have to 

55
00:03:28,280 --> 00:03:32,200
be low carbon, and they've 
already invested heavily in wind

56
00:03:32,200 --> 00:03:36,240
and solar, but their data 
centres still need a steady base

57
00:03:36,240 --> 00:03:38,720
load. 
So for big Tech, investing in 

58
00:03:38,720 --> 00:03:41,520
nuclear energy makes a lot of 
sense. 

59
00:03:42,200 --> 00:03:45,920
If we look at the pledges that 
big tech firms have made, almost

60
00:03:45,920 --> 00:03:49,600
all of them have promised to 
reach net 0 carbon emissions by 

61
00:03:49,600 --> 00:03:52,800
2040. 
These pledges often involve 

62
00:03:52,800 --> 00:03:56,640
transitioning to renewable 
energy sources, investing in 

63
00:03:56,640 --> 00:04:01,200
carbon capture technology, and 
optimizing their operations and 

64
00:04:01,200 --> 00:04:04,560
supply chains for reduced energy
consumption. 

65
00:04:04,920 --> 00:04:08,240
Their pledges are somewhat 
believable, as they they're so 

66
00:04:08,240 --> 00:04:11,600
profitable that they can 
actually afford to do this, 

67
00:04:11,600 --> 00:04:14,520
unlike smaller, less profitable 
businesses. 

68
00:04:14,920 --> 00:04:19,240
At the most extreme end of the 
scale, Microsoft pledged almost 

69
00:04:19,240 --> 00:04:24,000
five years ago to be carbon 
negative by 20-30, additionally 

70
00:04:24,000 --> 00:04:28,360
promising that by 2050 they will
have entirely stripped all of 

71
00:04:28,360 --> 00:04:31,920
the carbon they emitted over the
life of the firm from the 

72
00:04:31,920 --> 00:04:35,000
environment. 
As difficult as that goal will 

73
00:04:35,000 --> 00:04:39,680
have sounded in 2020, it's only 
gotten more difficult over time 

74
00:04:39,680 --> 00:04:43,200
thanks to AI. 
Microsoft now says that they 

75
00:04:43,200 --> 00:04:47,880
expect to consume five to six 
times more electricity than they

76
00:04:47,880 --> 00:04:51,440
had planned on when they first 
made that promise, but they are 

77
00:04:51,440 --> 00:04:55,360
still sticking to their goal. 
The reason for the huge step up 

78
00:04:55,360 --> 00:05:00,120
in energy consumption is that AI
consumes an awful lot of power. 

79
00:05:00,360 --> 00:05:03,360
A ChatGPT query, for example, 
uses. 

80
00:05:03,440 --> 00:05:07,320
This is about 10 times as much 
power to process as a simple 

81
00:05:07,320 --> 00:05:11,320
Google search does, and 
generating a single AI image 

82
00:05:11,480 --> 00:05:14,800
consumes the same amount of 
energy as fully charging a 

83
00:05:14,800 --> 00:05:18,600
smartphone. 
Microsoft revealed earlier this 

84
00:05:18,600 --> 00:05:23,480
year that it's carbon emissions 
have increased by 30% since 

85
00:05:23,480 --> 00:05:26,400
2020, so they're well off 
target. 

86
00:05:26,680 --> 00:05:30,560
The data centres being planned 
or built today consume more 

87
00:05:30,560 --> 00:05:35,040
power than can be bought from 
utilities, and so if Big Tech 

88
00:05:35,040 --> 00:05:40,200
wants a lot of 0 carbon power, 
their only realistic option is 

89
00:05:40,280 --> 00:05:43,200
nuclear power. 
Three Mile Island in 

90
00:05:43,200 --> 00:05:46,800
Pennsylvania was the scene of 
the worst commercial nuclear 

91
00:05:46,800 --> 00:05:51,840
accident in U.S. history, when a
cooling malfunction in 1979 

92
00:05:52,040 --> 00:05:56,480
caused part of the core to melt 
in reactor 2, releasing 

93
00:05:56,480 --> 00:06:00,480
radioactive gases and 
radioactive iodine into the 

94
00:06:00,480 --> 00:06:03,320
environment. 
Schools near the reactor were 

95
00:06:03,320 --> 00:06:06,800
quickly closed, pregnant women 
and small children were 

96
00:06:06,800 --> 00:06:10,880
evacuated from the area, and 
residents were urged to stay 

97
00:06:10,880 --> 00:06:14,600
indoors. 
Within a few days, 140,000 

98
00:06:14,600 --> 00:06:18,400
people left the area. 
In the aftermath, an 

99
00:06:18,400 --> 00:06:22,360
investigation found that the 
average radiation dose to people

100
00:06:22,360 --> 00:06:26,520
living within 10 miles of the 
plant was about equal to that of

101
00:06:26,520 --> 00:06:29,840
a chest X-ray. 
Further studies found that the 

102
00:06:29,840 --> 00:06:33,640
accident had no observable long 
term health effects on those 

103
00:06:33,640 --> 00:06:36,800
living near the plant. 
Nevertheless, it was a big 

104
00:06:36,800 --> 00:06:39,840
scare. 
The Three Mile Island accident 

105
00:06:39,960 --> 00:06:42,920
was a turning point in the 
growth of of nuclear power. 

106
00:06:43,200 --> 00:06:46,360
Prior to the incident, the 
number of reactors under 

107
00:06:46,360 --> 00:06:50,960
construction globally had grown 
steadily almost every year, but 

108
00:06:50,960 --> 00:06:55,320
following the event, 52 plant 
nuclear reactors were cancelled 

109
00:06:55,320 --> 00:07:00,760
in the United States alone. 
At the time of the incident, 129

110
00:07:00,760 --> 00:07:03,840
nuclear power plants had been 
approved for construction 

111
00:07:03,840 --> 00:07:08,120
worldwide, but of those only 53 
were completed. 

112
00:07:08,600 --> 00:07:12,000
The Chernobyl disaster in 
northern Ukraine seven years 

113
00:07:12,000 --> 00:07:15,880
later had an even bigger impact 
on public and political 

114
00:07:15,880 --> 00:07:20,000
attitudes towards the safety of 
nuclear power and ended the 

115
00:07:20,000 --> 00:07:22,480
growth in active reactors 
globally. 

116
00:07:23,280 --> 00:07:27,880
Last month it was announced that
Constellation Energy will reopen

117
00:07:27,880 --> 00:07:31,680
a nuclear plant, a Three Mile 
Island, to provide power to 

118
00:07:31,680 --> 00:07:35,840
Microsoft's data centres. 
The plant being reopened is Unit

119
00:07:35,840 --> 00:07:39,880
1, which was just decommissioned
five years ago, and Unit 2, 

120
00:07:39,880 --> 00:07:44,680
where the accident occurred in 
1979, will still remain closed. 

121
00:07:45,240 --> 00:07:49,280
The recommission plant at Three 
Mile Island will provide more 

122
00:07:49,280 --> 00:07:53,440
than 800 megawatts of power, all
of which has been bought by 

123
00:07:53,440 --> 00:07:57,840
Microsoft from Constellation 
Energy under a 20 year power 

124
00:07:57,840 --> 00:08:00,880
purchase agreement. 
The plant is expected to be 

125
00:08:00,880 --> 00:08:06,560
brought online in 2028 and to 
run until at least 2054. 

126
00:08:06,840 --> 00:08:10,280
The location of the data centres
that will use the electricity 

127
00:08:10,360 --> 00:08:14,680
has not yet been announced. 
Constellation says that the deal

128
00:08:14,680 --> 00:08:20,040
will create more than 3000 jobs,
add $16 billion to Pennsylvania 

129
00:08:20,040 --> 00:08:25,080
GDP, and bring in at least $3 
billion in state and federal 

130
00:08:25,080 --> 00:08:28,680
taxes while providing safe, 
reliable, carbon free 

131
00:08:28,680 --> 00:08:31,800
electricity that helps reduce 
emissions. 

132
00:08:32,200 --> 00:08:36,280
Not to be outdone, Amazon last 
week announced that they would 

133
00:08:36,280 --> 00:08:40,760
be spending more than $500 
million to support development 

134
00:08:40,760 --> 00:08:44,320
of four next generation small 
modular reactors. 

135
00:08:44,560 --> 00:08:49,600
They took a stake in X Energy 
and SMR Developer Google 

136
00:08:49,600 --> 00:08:54,240
additionally agreed to buy 500 
megawatts of power from six to 

137
00:08:54,240 --> 00:08:58,200
seven small modular reactors, 
which will be built by Cairo's 

138
00:08:58,200 --> 00:09:00,760
Power. 
None of these companies revealed

139
00:09:00,760 --> 00:09:04,440
the financial details of the 
agreements or where in the US 

140
00:09:04,440 --> 00:09:08,320
the plants will be built. 
The idea behind small modular 

141
00:09:08,320 --> 00:09:12,840
reactors, known as SMRS, is that
the components of each unit can 

142
00:09:12,840 --> 00:09:16,680
be small enough to be mass 
produced on an assembly line, 

143
00:09:16,840 --> 00:09:19,840
making them cheaper to 
manufacture than full size 

144
00:09:19,840 --> 00:09:22,960
reactors. 
Each power plant could start 

145
00:09:23,080 --> 00:09:26,160
with just one reactor and add 
more as needed. 

146
00:09:26,440 --> 00:09:30,720
SMRS can, in theory, be cited in
locations not suitable for 

147
00:09:30,720 --> 00:09:35,520
larger nuclear power plants. the
US Department of Energy defines 

148
00:09:35,520 --> 00:09:39,840
SMRS as advanced nuclear 
reactors with a power capacity 

149
00:09:39,840 --> 00:09:44,280
of up to 300 megawatts per unit,
which is about 1/3 of the 

150
00:09:44,280 --> 00:09:47,080
generating capacity of 
traditional nuclear power 

151
00:09:47,080 --> 00:09:50,240
reactors. 
Rolls Royce's SMR design 

152
00:09:50,440 --> 00:09:55,480
generates 470 megawatts, which 
is actually bigger than many of 

153
00:09:55,480 --> 00:09:58,680
the 1st generation nuclear 
reactors that were built in the 

154
00:09:58,680 --> 00:10:02,280
UK. 
According to Goldman Sachs data 

155
00:10:02,280 --> 00:10:07,800
center, power demands will grow 
160% between now and 2030. 

156
00:10:08,040 --> 00:10:10,280
They say that when this is 
combined with the 

157
00:10:10,280 --> 00:10:14,280
electrification of transport and
the reassuring of manufacturing,

158
00:10:14,520 --> 00:10:17,640
electricity demand can be 
expected to double. 

159
00:10:17,840 --> 00:10:22,440
Since 2021, American companies 
have announced plans to spend 

160
00:10:22,600 --> 00:10:26,720
more than $500 billion on 
factories for semiconductors, 

161
00:10:26,720 --> 00:10:29,560
batteries, solar panels and 
more. 

162
00:10:29,880 --> 00:10:33,040
As great as this sounds, there's
not sufficient generating 

163
00:10:33,040 --> 00:10:36,560
capacity to power all of these 
factories today. 

164
00:10:37,080 --> 00:10:41,520
The climate author Solomon 
Goldstein Rose argued in a 2001 

165
00:10:41,520 --> 00:10:45,520
Ted talk that while the world is
on track to replace all fossil 

166
00:10:45,520 --> 00:10:50,080
fuel power plants by 2050, that 
this would not be good enough as

167
00:10:50,080 --> 00:10:54,920
we're likely to need five times 
more power in 2050 as we use 

168
00:10:54,920 --> 00:10:59,000
today if the world is actually 
switching to electric vehicles, 

169
00:10:59,240 --> 00:11:02,680
Electric heating and 
electrifying various industrial 

170
00:11:02,680 --> 00:11:06,000
processes like steel making and 
concrete manufacturing. 

171
00:11:06,280 --> 00:11:10,360
He points out that over 700 
million people today have no 

172
00:11:10,360 --> 00:11:14,400
access to electricity worldwide,
and that billions more have 

173
00:11:14,400 --> 00:11:18,160
access to only small amounts of 
intermittent power, and that 

174
00:11:18,160 --> 00:11:21,840
providing these people access to
energy would lift them out of 

175
00:11:21,840 --> 00:11:24,680
poverty. 
His estimate of needing five 

176
00:11:24,680 --> 00:11:29,280
times as much power came before 
these AI data centres were being

177
00:11:29,280 --> 00:11:32,800
built. 
Nuclear power is well suited to 

178
00:11:32,800 --> 00:11:36,760
data centres because it provides
a reliable source of power 

179
00:11:36,760 --> 00:11:39,760
around the clock, unlike wind 
and solar. 

180
00:11:40,000 --> 00:11:42,920
There are a few excellent 
YouTube videos by Practical 

181
00:11:42,920 --> 00:11:46,560
Engineering explaining the 
importance of a base load on 

182
00:11:46,560 --> 00:11:50,560
power grids and why it's much 
harder than you would expect to 

183
00:11:50,560 --> 00:11:54,760
connect solar and wind power to 
our existing electrical grids. 

184
00:11:55,360 --> 00:11:59,320
I was surprised to learn that 
many grid tight solar systems 

185
00:11:59,320 --> 00:12:03,040
shut down whenever the grid goes
down, which is not really what 

186
00:12:03,040 --> 00:12:05,880
you would want. 
This happens both for the safety

187
00:12:05,880 --> 00:12:10,160
of utility workers and because 
many of the inverters are grid 

188
00:12:10,160 --> 00:12:14,560
following inverters which need 
to sense the voltage of the grid

189
00:12:14,560 --> 00:12:18,200
so that they can synchronize 
their output to match the grids 

190
00:12:18,200 --> 00:12:22,880
frequency and phase. 
A more obvious barrier to a 100%

191
00:12:22,880 --> 00:12:26,640
wind and solar grid is the 
intermittency of these power 

192
00:12:26,640 --> 00:12:29,400
sources. 
The wind doesn't always blow and

193
00:12:29,400 --> 00:12:33,600
the sun doesn't always shine. 
On top of that, the windiest and

194
00:12:33,600 --> 00:12:37,880
sunniest places are not all that
close to most major population 

195
00:12:37,880 --> 00:12:41,360
centers. 
Grid scale battery backup has 

196
00:12:41,360 --> 00:12:45,080
been put forth as a solution to 
this problem, but it's by no 

197
00:12:45,080 --> 00:12:49,440
means a perfect solution either,
as rapid cycling degrades the 

198
00:12:49,440 --> 00:12:52,240
lifespan of utility scale 
batteries. 

199
00:12:52,560 --> 00:12:57,040
A 2021 wind drought in Europe 
meant that windmills generated 

200
00:12:57,040 --> 00:13:01,000
significantly less electricity 
than was expected, sending power

201
00:13:01,000 --> 00:13:04,280
prices to record highs as 
utilities had to scramble to 

202
00:13:04,280 --> 00:13:07,560
find alternative fossil fuel 
power sources. 

203
00:13:08,080 --> 00:13:12,160
This map shows the potential for
solar power in different parts 

204
00:13:12,160 --> 00:13:15,640
of the world, and you'll notice 
that Canada and Western Europe 

205
00:13:15,800 --> 00:13:19,280
are particularly I'll suited to 
solar generation. 

206
00:13:19,840 --> 00:13:23,280
British businesses, as an 
example, already face the 

207
00:13:23,280 --> 00:13:27,240
highest energy costs in the 
developed world, and if the new 

208
00:13:27,240 --> 00:13:30,600
government is hoping to boost 
growth, dealing with the weak 

209
00:13:30,600 --> 00:13:33,240
energy infrastructure might be a
good start. 

210
00:13:33,520 --> 00:13:37,480
Unfortunately, Britain is not 
only I'll suited for solar and 

211
00:13:37,480 --> 00:13:41,040
wind generation, it's also the 
most expensive place in the 

212
00:13:41,040 --> 00:13:44,040
world to build a new nuclear 
power station. 

213
00:13:44,360 --> 00:13:47,920
According to Sam Dimitro 
Hinckley, Point C, which is 

214
00:13:47,920 --> 00:13:52,360
currently being built, will cost
six times more to build per MW 

215
00:13:52,560 --> 00:13:54,600
than it's South Korean 
equivalent. 

216
00:13:54,840 --> 00:13:59,160
France and Finland, which used 
the same reactor design, managed

217
00:13:59,160 --> 00:14:03,000
to build them for about 50% less
per MW. 

218
00:14:03,360 --> 00:14:07,040
South Korea have kept their 
costs lower by building fleets 

219
00:14:07,040 --> 00:14:10,120
of reactors benefiting from 
economies of scale. 

220
00:14:10,400 --> 00:14:14,560
The same reactors is being built
in the UK, is cheaper to build 

221
00:14:14,560 --> 00:14:18,840
in Finland and France because of
a simpler approvals process. 

222
00:14:19,080 --> 00:14:23,480
UK rate regulations require a 
more complex planning process 

223
00:14:23,640 --> 00:14:28,680
and requested 7000 changes to 
the original design, essentially

224
00:14:28,680 --> 00:14:32,920
making it entirely bespoke. 
One of the alterations required 

225
00:14:32,920 --> 00:14:36,880
to make it through the planning 
process involves spending 10s of

226
00:14:36,880 --> 00:14:41,520
millions of pounds on underwater
loud speakers to scare fish away

227
00:14:41,520 --> 00:14:45,960
from the cooling intake. 
This fish disco is not required 

228
00:14:45,960 --> 00:14:49,960
in other countries. 
Even outside the UK, nuclear 

229
00:14:49,960 --> 00:14:54,080
power can be extremely expensive
and time consuming to build. 

230
00:14:54,280 --> 00:14:57,240
It's much more expensive than 
wind or solar. 

231
00:14:57,400 --> 00:15:01,080
And while the climate case for 
nuclear power is strong, the 

232
00:15:01,080 --> 00:15:04,560
challenge of safely storing 
nuclear waste has not been 

233
00:15:04,560 --> 00:15:07,760
entirely resolved. 
There is, of course, plenty of 

234
00:15:07,760 --> 00:15:11,200
controversy over whether nuclear
power can be considered clean 

235
00:15:11,200 --> 00:15:15,520
energy or not, mostly driven by 
concerns around nuclear waste, 

236
00:15:15,680 --> 00:15:19,880
nuclear weapon proliferation, 
and the risk of accidents. 

237
00:15:20,360 --> 00:15:24,400
A number of studies show that 
coal power plants output more 

238
00:15:24,400 --> 00:15:27,040
radiation than nuclear power 
plants do. 

239
00:15:27,360 --> 00:15:32,440
The fly ash emitted when burning
coal for electricity carries 100

240
00:15:32,440 --> 00:15:36,240
times more radiation than a 
nuclear power plant producing 

241
00:15:36,240 --> 00:15:40,080
the same amount of power. 
According to the US Office of 

242
00:15:40,080 --> 00:15:43,720
Nuclear Energy, the use of 
nuclear power in the United 

243
00:15:43,720 --> 00:15:47,080
States has reduced CO2 
emissions, equivalent to 

244
00:15:47,080 --> 00:15:51,400
removing 100 million cars from 
the road, and nuclear power 

245
00:15:51,400 --> 00:15:55,880
plants produce more electricity 
on less land than any other 

246
00:15:55,880 --> 00:16:00,320
clean energy source. 
They say that all of the used 

247
00:16:00,320 --> 00:16:04,640
nuclear fuel produced by the US 
nuclear energy industry could 

248
00:16:04,640 --> 00:16:08,840
fit on an American football 
field at a depth of less than 3 

249
00:16:08,840 --> 00:16:12,160
yards. 
Because this video is about AI 

250
00:16:12,160 --> 00:16:15,440
and energy, I thought it might 
make sense to use Microsoft 

251
00:16:15,440 --> 00:16:18,920
Copilot to convert those 
measurements into something more

252
00:16:19,080 --> 00:16:21,920
understandable for an 
international audience. 

253
00:16:22,360 --> 00:16:26,320
And I'm not actually joking. 
It came back with the reply that

254
00:16:26,320 --> 00:16:30,680
this volume is equivalent to 18 
Olympic sized swimming pools 

255
00:16:30,880 --> 00:16:36,360
filled up with Tesla's. 
So yeah, it's a big box, I 

256
00:16:36,360 --> 00:16:39,840
guess. 
The US Office of Nuclear Energy 

257
00:16:39,840 --> 00:16:43,640
goes on to say in their report 
that a lot of that waste could 

258
00:16:43,640 --> 00:16:47,680
be reprocessed and recycled and 
that many of the new reactor 

259
00:16:47,680 --> 00:16:50,400
designs can operate on used 
fuel. 

260
00:16:51,040 --> 00:16:54,200
Now there are not enough 
decommissioned nuclear reactors 

261
00:16:54,200 --> 00:16:58,760
like 3 Mile Island for Big Tech 
to resurrect, but they have 

262
00:16:58,760 --> 00:17:02,360
shown a growing interest in 
using small modular reactors. 

263
00:17:02,640 --> 00:17:06,280
The problem with that is that 
there are only three operational

264
00:17:06,280 --> 00:17:10,520
SM Rs in the world today, one in
China, one in Russia, and 1:00 

265
00:17:10,520 --> 00:17:12,880
in India. 
Three more are under 

266
00:17:12,880 --> 00:17:16,800
construction and 65 are still in
the design stage. 

267
00:17:17,160 --> 00:17:22,000
GE, Vernova, Rolls Royce and 
Westinghouse all have SMR 

268
00:17:22,000 --> 00:17:26,240
designs, as do a long list of 
startups like the Bill Gates 

269
00:17:26,240 --> 00:17:30,160
backed Terror Power, which broke
ground on its first plant in 

270
00:17:30,160 --> 00:17:34,840
Wyoming this summer, X Energy, 
which is backed by Amazon, and 

271
00:17:34,880 --> 00:17:39,680
Oaklow, which counts Sam Altman 
of Open AI among its backers. 

272
00:17:40,160 --> 00:17:43,480
Bill Gates has invested more 
than a billion dollars in Terror

273
00:17:43,480 --> 00:17:47,360
Power, which is working to 
develop SMRS in partnership with

274
00:17:47,360 --> 00:17:50,360
Warren Buffett's utility 
company, Pacificorp. 

275
00:17:50,760 --> 00:17:55,760
France gets around 70% of its 
electricity from nuclear energy,

276
00:17:55,920 --> 00:17:58,520
which is the highest percentage 
in the world. 

277
00:17:58,760 --> 00:18:03,120
But the United States generates 
the most nuclear power overall, 

278
00:18:03,360 --> 00:18:07,720
mostly through plants that began
construction in the 1960s and 

279
00:18:07,720 --> 00:18:10,320
70s. 
There are more SMRS being 

280
00:18:10,320 --> 00:18:14,280
planned in the United States 
today than anywhere else, 

281
00:18:14,400 --> 00:18:16,720
largely because of the tech 
industry. 

282
00:18:17,320 --> 00:18:22,920
While promising, the technology 
behind SMRS is still unproven at

283
00:18:22,920 --> 00:18:24,840
scale. 
And the recent Wall Street 

284
00:18:24,840 --> 00:18:28,920
Journal article asked what 
customer would pay the higher 

285
00:18:28,920 --> 00:18:32,720
price for a first of kind 
project and who would order 

286
00:18:32,720 --> 00:18:35,360
enough to get an assembly line 
started. 

287
00:18:35,720 --> 00:18:38,800
The answer to that question 
today is quite obvious. 

288
00:18:39,200 --> 00:18:41,920
Big tech. 
And while power plants can be 

289
00:18:41,920 --> 00:18:44,640
thought of as boring 
infrastructure which might be 

290
00:18:44,640 --> 00:18:48,400
difficult to finance, the fact 
that these data centres are 

291
00:18:48,400 --> 00:18:52,280
needed for AI, which venture 
capital and retail investors are

292
00:18:52,280 --> 00:18:57,000
very excited about, means that 
AI companies can raise loads of 

293
00:18:57,000 --> 00:19:01,640
cheap money from VC funds and 
pump it into building not only 

294
00:19:01,640 --> 00:19:05,360
the data centres that they need,
but also the nuclear power to 

295
00:19:05,360 --> 00:19:07,880
run them on. 
As without sourcing your own 

296
00:19:07,880 --> 00:19:11,680
power, you might might be out of
luck waiting for grid upgrades. 

297
00:19:12,160 --> 00:19:16,760
Last month, Microsoft teamed up 
with BlackRock to launch a $30 

298
00:19:16,760 --> 00:19:20,240
billion fund to invest in 
artificial intelligence 

299
00:19:20,240 --> 00:19:23,200
infrastructure. 
They told the press that the 

300
00:19:23,200 --> 00:19:28,240
partnership would mobilize up to
$100 billion in total investment

301
00:19:28,240 --> 00:19:32,400
potential once debt financing 
was included, and that the money

302
00:19:32,400 --> 00:19:36,880
would be invested to enhance AI,
supply chains and energy 

303
00:19:36,880 --> 00:19:39,800
sourcing. 
Microsoft said that they expect 

304
00:19:39,800 --> 00:19:43,800
this trend to catch on with 
other tech firms and said that 

305
00:19:43,800 --> 00:19:46,760
Microsoft might participate in 
all of them. 

306
00:19:47,280 --> 00:19:50,760
Governments, after years of 
ignoring it, are now getting 

307
00:19:50,760 --> 00:19:55,480
excited about nuclear power too.
During UN Climate Week, 

308
00:19:55,640 --> 00:19:58,960
officials reaffirmed their 
commitment to triple the global 

309
00:19:58,960 --> 00:20:04,240
output of nuclear power by 2050,
and at the same event, 14 large 

310
00:20:04,240 --> 00:20:07,040
banks said that they were 
willing to finance these 

311
00:20:07,040 --> 00:20:10,840
projects. 
Joe Biden recently signed a law 

312
00:20:10,840 --> 00:20:14,560
with bipartisan support, which 
is almost unheard of in the 

313
00:20:14,560 --> 00:20:18,880
United States today, to hasten 
the development of nuclear 

314
00:20:18,880 --> 00:20:22,280
energy projects. 
Historically, many Democrats 

315
00:20:22,280 --> 00:20:26,760
opposed new nuclear plants, 
citing safety, environmental and

316
00:20:26,760 --> 00:20:30,000
economic concerns. 
Today, they see it as being 

317
00:20:30,000 --> 00:20:33,040
critical to their goals of 
reducing greenhouse gas 

318
00:20:33,040 --> 00:20:36,400
emissions and meeting the needs 
of a growing economy. 

319
00:20:37,040 --> 00:20:41,080
The Department of Energy Loan 
Programs Office has a $600 

320
00:20:41,080 --> 00:20:45,080
billion allocation of funds for 
renewables and advanced 

321
00:20:45,080 --> 00:20:48,800
manufacturing, and this can be 
used to fund new nuclear 

322
00:20:48,800 --> 00:20:52,040
projects too. 
As an example of how much 

323
00:20:52,040 --> 00:20:55,720
funding there seems to be, 
Bloomberg reported this summer 

324
00:20:55,840 --> 00:21:00,840
that Sam Altman of Open AI was 
staking a sizeable chunk of his 

325
00:21:00,840 --> 00:21:05,360
personal wealth on a nuclear 
fusion start up, which he says 

326
00:21:05,360 --> 00:21:10,520
is key to an AI enabled future. 
Nuclear fusion attempts to 

327
00:21:10,520 --> 00:21:14,320
create energy not by splitting 
atoms, but by smashing them 

328
00:21:14,320 --> 00:21:18,240
together, which requires extreme
temperatures and lots of power, 

329
00:21:18,480 --> 00:21:22,560
but hypothetically could output 
with more energy than goes into 

330
00:21:22,560 --> 00:21:26,200
it without creating the 
radioactive waste associated 

331
00:21:26,200 --> 00:21:29,080
with today's nuclear fission 
reactors. 

332
00:21:29,720 --> 00:21:33,320
While currently unproven as a 
workable option for power 

333
00:21:33,320 --> 00:21:36,720
generation, nuclear fusion 
advocates claim that if 

334
00:21:36,720 --> 00:21:41,040
successful, fusion has the 
potential to be a sustainable 

335
00:21:41,040 --> 00:21:45,040
and abundant energy source. 
Skeptics, on the other hand, say

336
00:21:45,240 --> 00:21:50,080
that fusion is worse than Tesla 
robotaxis, as it's always just 

337
00:21:50,280 --> 00:21:54,200
20, 20 years away. 
Most experts say that we're 

338
00:21:54,200 --> 00:21:58,200
unlikely to be able to generate 
large scale energy from nuclear 

339
00:21:58,200 --> 00:22:02,200
fusion anytime soon. 
But Microsoft still signed a 

340
00:22:02,200 --> 00:22:05,800
power purchase agreement with 
Altman's company Helion. 

341
00:22:05,840 --> 00:22:09,080
Earlier this year. 
Bloomberg reported that this 

342
00:22:09,080 --> 00:22:13,280
deal cut some Helion employees 
by surprise, and they expressed 

343
00:22:13,280 --> 00:22:16,560
doubts about the company's 
promise to get a prototype 

344
00:22:16,560 --> 00:22:20,280
running by year end. 
Like most VC backed start-ups, 

345
00:22:20,600 --> 00:22:24,840
Elian seems to be aiming to move
fast and break things, but is 

346
00:22:24,840 --> 00:22:29,400
unfortunately being slowed down 
not just by the science, but by 

347
00:22:29,400 --> 00:22:32,400
numerous complaints of 
harassment and gender 

348
00:22:32,400 --> 00:22:35,920
discrimination. 
The fusion industry, which as I 

349
00:22:35,920 --> 00:22:39,920
said is not hugely promising in 
the short term, has still 

350
00:22:39,920 --> 00:22:43,800
managed to attract more than $7 
billion in venture capital 

351
00:22:43,800 --> 00:22:47,480
investment. 
A nuclear power resurgence may 

352
00:22:47,480 --> 00:22:51,320
not be essential for AI to get 
off the ground, but when 

353
00:22:51,320 --> 00:22:54,400
combined with the climate 
commitments big tech firms have 

354
00:22:54,400 --> 00:22:57,520
made, along with the power 
demands of electrified 

355
00:22:57,520 --> 00:23:01,560
transportation and reassuring of
manufacturing, it does seem to 

356
00:23:01,560 --> 00:23:04,640
be happening. 
From the perspective of big tech

357
00:23:04,640 --> 00:23:08,680
firms, nuclear power success 
would help calm the climate 

358
00:23:08,680 --> 00:23:12,880
driven fears about the extreme 
power demands of generative AI. 

359
00:23:13,200 --> 00:23:17,400
And after that it'll be up to 
Big Tech to develop the AGI, 

360
00:23:17,600 --> 00:23:21,200
which they for some strange 
reason are telling investors 

361
00:23:21,200 --> 00:23:24,560
will kill us all. 
Which I have to say, I'm firmly 

362
00:23:24,560 --> 00:23:26,960
against. 
Thanks for tuning into this 

363
00:23:26,960 --> 00:23:30,360
week's podcast, with a special 
thanks to my supporters on 

364
00:23:30,360 --> 00:23:32,680
Patreon who make this all 
happen. 

365
00:23:32,880 --> 00:23:36,280
If you want to support the show,
I'll put a link in the show 

366
00:23:36,280 --> 00:23:38,280
notes. 
Have a great week and talk to 

367
00:23:38,280 --> 00:23:39,800
you again soon. 
Bye.