Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:02
Hello and welcome to the Energy Gang
0:04
with the special edition from the distributor
0:07
conference in Orlando and it cracks. Distributor.
0:14
Is a big event for the electricity transmission
0:16
and distribution and straight in fact it's the
0:18
biggest in North America. Being.
0:20
Here is a fantastic opportunity to talk to
0:22
many of the leading figures from companies that
0:24
provide technology for moving and managing electricity, and
0:27
from the companies that use that technology, including
0:29
particularly utilities. Later on, in this episode, you
0:31
can hear my conversations with Antony Allah Hu,
0:33
the head of Hitachi Energy in North America
0:35
with Tom Dietrich of I Trump, with Quinn
0:38
that Yamamah from the utility Pgd and with
0:40
Zach Cast is a futurist until last year
0:42
was an executive at Open A I. As
0:45
you might imagine, A I has been a hot
0:47
topic of this event. It's been talked about both
0:50
as a source of new demand for electricity and
0:52
of a tool for managing the new strains on
0:54
the grid. To discuss what
0:56
these latest breakthroughs in A I mean
0:58
for energy I spoke first with Hossain.
1:00
Shell is the chief technologist for eight
1:03
of us. Amazon Web Services has I
1:05
think so much for joining us today.
1:07
Thank you for having me excited! So
1:09
Ai does seem to be the hot
1:11
topic already have just come from the
1:13
plenary session, the opening keynote speeches in
1:15
the main hall A I was a
1:18
consistent theme running through everything and everyone
1:20
was talking about their when you think
1:22
about A W S his approach to
1:24
a I in the. Energy Industry to
1:26
Power and utilities. What are you
1:28
offering? Pay? Absolutely. So it's really
1:30
exciting times. It's definitely the topic
1:32
of the moment and I believe
1:34
in the company really believes that
1:36
this gonna be transformative technology for
1:38
the next decades. Our approach really
1:40
around make in Gen Vi practical,
1:42
cutting through the hype and helping
1:44
our customers do the work that
1:46
they want to do with it
1:49
right and trying to find something
1:51
that is secure, flexible, and capable
1:53
of doing the things that they
1:55
want to. Do and all of those
1:57
are available through our services in the Thompson.
1:59
So a. Have you talked about
2:01
quite a bit more? Manage again Hogarth
2:03
do with had quality people. Come on
2:05
and talk about the possibilities of a
2:08
technology opens up and very much with
2:10
a theme here distributed among excitement. the
2:12
expression game changer I think have been
2:14
used about three or four times the
2:16
rate in my hearing you've just been
2:18
talking about the exciting potential that you
2:21
see. I have tended personally to be
2:23
a bit of a skeptic. I have
2:25
felt like there's a lot of excitement
2:27
around the sort of when you Michael
2:29
the. Latest iteration of a i
2:31
intend zones large language models and
2:33
so on which have clearly have
2:36
enormous implications for language face to
2:38
ignite. It was eventually for journalism
2:40
and the law and teaching and
2:42
so I can see really where
2:44
these new ai tim of these
2:46
make a massive difference to vote
2:48
What I feel skeptical about his
2:50
verification to energy where it seems
2:53
like maybe some of the use
2:55
cases are not so well demonstrated.
2:57
Yes and as as a lot
2:59
of. Talk about from it is
3:01
essential in principle, not so much
3:03
really concrete demonstration. The oh what's
3:05
that he beings on and what's
3:07
really being changed on the ground
3:10
right now by I am I
3:12
wrong to have been impressions you
3:14
think it is changing a lot
3:16
already. Eat or is this more
3:18
kind of something? Much snow for
3:20
the future and something that people
3:22
are kind of speculating about. I'm
3:24
looking forward to Rome that you
3:26
demonstrating right now. I think we're
3:28
in the early stages. Right as
3:30
I said sen the eyes as a discipline
3:32
has been around for a while but from
3:35
a technology and enablement and the ability to
3:37
use it and is large scale is fairly
3:39
new with advantage of Tennessee be t that
3:41
going to put it up in the market
3:44
and it's most people to the possibilities. but
3:46
to be honest with you the year to
3:48
the twenty three has been a lot of
3:50
feel season pilots and people trying to understand
3:53
what they can do with just like as
3:55
any new to college ethics. And
3:59
so. So I would say 2023 has been
4:01
a significant amount of piloting and
4:04
POCing and trying different things and
4:06
different ideas. Working with partners like
4:08
Accenture, they use CodeWhisper, which is
4:10
a GenAI based code suggestion system
4:12
that we have. They've seen 30%
4:14
more productivity from
4:17
their engineering developers. So we're
4:20
seeing some of those cases, right? Other
4:22
customers are using it around data, like
4:24
you said. It's very text specific and
4:26
also there's multi-modular, right, with video and
4:28
voice, et cetera. And energy customers and
4:31
energy industry is rich with data and documents
4:33
that we don't know what to do
4:35
with and where, what information, what insights are
4:37
there. So some of the use cases that
4:39
we're looking at are looking at this corpus
4:42
of data sets and documents that have just
4:44
been sent there and drives for decades
4:46
and trying to glean insights from it and
4:48
see what we can do with it and
4:51
what our customers can actually find value, right?
4:53
But I believe 2024 is going to see
4:56
a significant scale of deployment to
4:58
production, applications and solutions. 2023,
5:00
mostly POC. Right,
5:03
got it. And POC then proof
5:05
of concept. Proof of concept, yes, absolutely. So
5:07
as you say, so when people are moving
5:09
then, you think in 2024 to actually deploying
5:11
these applications practically, and you
5:13
talk about in energy sort of going through the
5:16
vast reams of text and pages that are out
5:18
there that nobody ever looks at at the moment.
5:20
So this is what, typically regulatory filings, things like
5:22
that. That, but also like in the case of
5:24
oil and gas, for example, there's a lot of
5:26
drilling reports, there's a lot of
5:28
production reports, there's a lot of content that is generated that
5:31
simply is forgotten after a project
5:33
is done. And a lot of
5:35
insights and data driven decisions are
5:38
buried within these PowerPoints or PDFs or Word
5:40
documents and people are starting to find it.
5:42
You've mentioned also law, we're seeing customers looking
5:45
at warranties and looking at documents that could
5:47
potentially allow them to have profit recovery use
5:49
cases. And so we're seeing a lot of
5:51
that, but yes, absolutely. That's very interesting. And
5:53
that's something you think then we should really
5:55
be looking out for this year to see
5:58
those practical implications taking effect. Absolutely.
6:00
So something else I often think about
6:02
in this context of AI is the
6:04
terminology seems to be very significant and
6:06
the terminology goes through these kind of
6:09
fashions and things become buzzwords and they
6:11
go into fashion and go out of
6:13
fashion. So now it's fashionable to talk
6:15
about AI. It feels like people have
6:17
been talking about AI in the context
6:19
of business in general, industry in general,
6:22
energy in particular for at least a
6:24
decade, maybe 15 years or so. It's
6:26
often been known as machine learning there
6:28
and people have used the term AI
6:30
and machine learning pretty well interchangeably.
6:34
But now it's being called AI and people
6:37
don't seem to talk about machine learning so
6:39
much because AI is exactly, people talk about
6:41
generative AI as being the kind of the
6:44
new new thing and that's the hot thing
6:46
that people are excited about. What are the
6:48
real differences here? Is what we're talking about
6:50
fundamentally different from the machine learning that people
6:52
have been talking about for many years now
6:55
or is it just essentially the same thing
6:57
being rebranded? Well, it's really
6:59
different disciplines when it comes to
7:01
using data and technology to glean
7:03
insights into the information that's provided.
7:06
Machine learning is the traditional machine learning.
7:08
You feed it data, you train it, it
7:10
comes back with learning information from that data
7:12
set. AI is when it learns
7:15
on its own and there's different reinforcement
7:17
learning techniques. There's very different disciplines. Generative
7:20
AI in particular is using large
7:22
language models and large
7:24
content models to generate new text.
7:26
Simply put, we can go for
7:28
days describing those different disciplines but
7:30
at least for today we'll talk
7:33
about that. What's interesting like
7:35
you said is that in our engagement
7:37
and a lot of times when we
7:39
sit down with our customers and work
7:41
backwards from their problems, 40% of the
7:43
problems that they're trying
7:46
to solve are easily solvable by
7:48
existing machine learning models and machine
7:50
learning algorithms that have been used
7:52
for decades. It is
7:54
our job as technologists to sit down
7:56
with our customers and explain the differences
7:58
and explain where. One works better
8:00
than the Of and so you know
8:03
We get passed by many of our
8:05
executive board some customers on for me
8:07
to them and explaining to them the
8:09
differences helping them and of id a
8:11
new ideas and use cases that we
8:13
can use and then we can target
8:15
each one of those use cases with
8:17
the proper technology in the proper implementation
8:19
of machine learning arcs. Sir
8:22
Francis whatever uses people often since talk
8:24
about is using data to optimize of
8:26
assholes of a powerful on the site.
8:28
Second, every industry example that's something that's
8:31
really a machine learning job, as in
8:33
not really what you call something for
8:35
generative. I'm correct. A Traditionally yes, you
8:37
read a lot of sensor data from
8:39
Atlanta You know than any kind of
8:41
machinery that generates a significant amount of
8:43
inputs and bait on temperatures and pressures
8:46
censoring fit in as a model. The
8:48
mama learns from it and then provides
8:50
predictions. An inference and zone that.
8:54
On when people for instance talk about
8:56
grid optimization as well some of the
8:58
increasingly complex problems that imminent ending with
9:00
managing grids about a use. that's also
9:03
what again less radio machine learning all
9:05
it is. It's complicated to they think
9:07
you can merge different disciplines together to
9:09
create the right solution. or you can
9:12
use machine learning for the typical censor
9:14
nina and ingestion data that is to
9:16
be used in the past. Learn and
9:18
educate the model and train the model
9:20
to be within such as you can
9:23
also merged with. New datasets, synthetic thing
9:25
and were seen him some of the use
9:27
cases that is blending physics based model with
9:30
this is to go and database models and
9:32
so you're seeing that world merging and
9:34
becoming very gray is when it comes to
9:36
that. But you know I'm really really looking
9:38
forward to see what customers are going to
9:41
do and how we can use all
9:43
of these disciplines that available today for our
9:45
customers to come up with. His sons are
9:47
that's really interesting thanks so much and that's
9:50
really helpful and clarifying they term plank or
9:52
so when you think longer term. About
9:54
the potential of a in energy come
9:56
back to the super thick and five
9:58
years or ten years. Twenty nine,
10:00
Twenty Thirty Four, what are people
10:03
going to be talking about Van
10:05
and mean, what is that longer
10:07
term potential and how do you
10:09
think ultimately I will change the
10:12
emphasis veins reputable. I'm hoping and
10:14
five years we can see a
10:16
lot of these use cases to
10:18
fruition. You see a lot of
10:21
business empire that's driving better energy
10:23
efficiency, better grid optimization scenarios, and
10:25
really see the fruits of experimentation.
10:27
What's happening now in the industry.
10:30
To be true and we're going to
10:32
see a lot of different implementations of
10:34
these Muslims as models and I'm really
10:36
really hoping that customers have been a
10:38
see the benefit in our partners and
10:40
continue to grow the business for them
10:42
and see a difference in the lives
10:44
to their customers but also their operations
10:46
their existing for you mentioned good integration.
10:48
The other aspects of Ai that's been
10:50
much discussed already here at the Service
10:52
Egg is the question of the increasing
10:54
load. That's gonna be great advice over
10:56
new data centers that are going to
10:58
be delivering an Ai. Applications that future the
11:01
was an executive from you Can as he
11:03
just speaking of the my and whole saying
11:05
that it used to be a big deal
11:07
for them when they had something we'd like
11:09
to add ten megawatts of twenty megawatts of
11:11
alone and now they've got a number of
11:14
data centers talk about Would like to add
11:16
oh gigawatt of like. It's clearly something that
11:18
is making a big difference to projects and
11:20
both. It was a smart in the future
11:22
on I guess some of that demand is
11:24
coming from you as a W S, how
11:26
do you manage? That in particular has given
11:29
that you have very. Demanding commitments in
11:31
terms of reducing emissions from relying
11:33
on renewable energy? How do you
11:35
square those two things away together
11:37
and managed to me that increased
11:39
demand for cure that extra power
11:41
you're gonna need while also reducing
11:44
emissions. I'm glad you mentioned Duke.
11:46
Duke is one of our partners
11:48
with and working with them significant
11:50
harness on some of the grid
11:52
optimization work. were helping them build
11:54
cloud native technology based sweet of
11:56
applications to really look at things
11:59
that are. Enabling Technologies right
12:01
in the meeting before the large,
12:03
upscale and updated efforts are going
12:05
on of infrastructure. So for example,
12:07
we're working with them on leveraging
12:09
not only machine learning, but also
12:11
traditional I Performance computing to communities
12:13
that was available in eight of
12:15
us to do simulations into power
12:17
flow simulations and look at how
12:19
they can optimize apology optimization. At
12:21
such and such, there's a significant
12:23
amount of different solutions that I'm
12:25
a foreigner. To
12:28
look at like on an optimization. Advance
12:30
our flow simulations and really trying
12:32
to see where can you take
12:34
already on more capacity without necessarily
12:36
burdening the grid that exists today
12:39
in the hope of improving in
12:41
the future. And so we realized
12:43
of this fake partnerships with our
12:45
partners with technology providers like ourselves
12:47
but also with regulators, providers and
12:49
of and senses. Yeah
12:52
and those grid enhancing themselves you something
12:54
particular important given how hard it is
12:56
to fill infrastructure for a in most
12:59
of the devote. Wells I guess you'd
13:01
say Utterance Power transmission a particular is
13:03
one of those things that seems to
13:05
be really hard to get done and
13:07
clearly given of his new demand for
13:09
the coming on the great post on
13:12
the demand side and because of increased
13:14
verbal renewables more went on the great
13:16
and so on the exercising whole new
13:18
challenges in needs a lot more transmission.
13:20
It's hard to build that transmission so
13:22
severe that you can do anything to
13:25
upgrade the great Improvement network without actually
13:27
adding physically more capacity and other really
13:29
hold it's. absolutely important and there's a
13:31
lot of things that were doing internally as
13:33
well for example were doing a lot of
13:35
work around our water consumption cooling in our
13:38
our innocence site a game has been a
13:40
subject will not be writing increasingly concerned about
13:42
the brightness yeah right in with our commitment
13:44
to be want a positive by twenty thirty
13:47
will be given more water out of beauty
13:49
that we consume no data centers we are
13:51
as you know one of the largest mobile
13:53
and as if providers we will be ah
13:55
hundred percent plans and twenty twenty five or
13:58
way had were ninety five percent at the
14:00
moment, we can help our customers
14:02
decrease their carbon footprint by using our services
14:04
up to close to 80% and 95% once
14:06
we're 100% renewable. So
14:10
we're doing a lot of things to encourage. Our
14:12
hope is that we compete in that
14:15
space. We're not going to be able
14:17
to solve, unfortunately, climate change by ourselves,
14:19
but we want our competitors and others
14:21
to also innovate with us. And we
14:23
want to push that limit for what
14:25
we're doing. A lot of our climate
14:27
pledges are ambitious goals, net zero by
14:29
2040, and that's going to take a lot of
14:31
innovation, a lot of technology. So you would be absolutely
14:33
confident then that you will be able to get to net
14:36
zero emissions by 2040. You're still totally
14:38
on track for that. Absolutely, we are. We're
14:40
seeing results. Like I said, our carbon intensity
14:42
reduced by 7% in 23. We
14:45
are on our way to be 100% renewable by 2025. We're
14:49
working on, across Amazon, not only AWS,
14:51
for example. We've reduced package waste by
14:54
close to 2 million tons by simply
14:56
reducing our package size by about 41%.
15:00
We've introduced a lot more EVs
15:02
into our shipping and delivery
15:04
system, about 9,000, and
15:06
145 million packages were delivered using these
15:08
trucks. So we're definitely seeing results. And
15:10
we will continue to innovate, of course.
15:13
It's not going to take only us,
15:15
and we're hoping to see what the
15:17
innovation continues across the industry. And those
15:19
challenges that obviously people see in terms
15:21
of being renewable 24-7. Variable
15:24
renewables obviously don't get you all the way
15:26
there. People look at a lot
15:28
of different solutions, like using biofuels, maybe
15:31
high-diesel and generators, buy gas. People
15:33
talk about nuclear maybe as being
15:35
part of the solution. How do
15:37
you see the potential for getting
15:39
to truly 100% renewable
15:42
energy 24-7, around the clock, given
15:44
that, as I say, you can get a lot
15:46
of the way with wind and solar, but you
15:48
can't get all the way there? Well, I think
15:51
technology is going to play a significant part in
15:53
that, but it's also going to take regulators and
15:55
it's going to take partners and customers themselves to
15:57
transform. In Amazon, we are dedicated to working across
15:59
the energy. industry all up. We work
16:01
with customers to innovate what they're doing
16:03
today in the renewable space, but also
16:05
in the oil and gas space. We
16:07
will continue to do so, we'll continue
16:09
to help our customers be as efficient
16:11
and as innovative as possible using our
16:13
technologies and we will continue to innovate
16:15
ourselves across all of the different stacks
16:17
that I mentioned from infrastructure to services
16:19
to the projects that we're doing in
16:21
this space, whether it's renewable projects or
16:23
cutting our own footprint ourselves. I'm
16:26
saying, Gail, thank you very much indeed. Absolutely. Thank you so
16:28
much for having me. One of
16:30
the keynotes because of the event was a former
16:32
executive of a company that's been at the heart
16:34
of the latest breakthroughs in artificial intelligence, OpenAI.
16:38
Up until last year, Zach Cass was the head of
16:40
GoToMarket. Zach, thanks very much for joining us today. Thanks
16:42
for having me in. So you're
16:44
not an energy specialist by background, but you're coming
16:47
here talking about this energy event. What's the message
16:49
you want to give people? What do you think
16:51
the energy industry needs to know about AI? Generally,
16:54
my message is that of abundance,
16:56
is giving people the permission to
16:59
explore a world where we might
17:01
have plentiful energy, foodstuffs, water, education,
17:04
healthcare, which is my mission, right?
17:06
To provide a counter narrative to
17:08
a dystopian or an otherwise
17:10
dystopian prevailing sentiment. At an energy conference,
17:13
it's particularly interesting because energy will be
17:15
both a beneficiary and a benefactor of
17:17
AI insofar as we're going to need
17:19
a ton of it to build these
17:22
machines and run them. And
17:24
presumably AGI will end
17:26
up producing fusion and other major
17:28
energy breakthroughs which should change how
17:30
we view energy forever. So let's
17:32
unpack that a bit and talk
17:34
about those two aspects of AI
17:36
energy in turn. Firstly, on the demand
17:39
side, if you like, so what AI needs
17:41
from energy and what its needs are in
17:43
terms of energy. You talk
17:45
about energy then being a beneficiary, you
17:48
could also say, obviously, that's one of the big
17:50
problems with AI is the scale of the energy
17:52
demand, particularly obviously at a time when we're trying
17:54
in general to reduce energy demand or at least
17:57
improve energy efficiency in the economy and a lot
17:59
of companies. They have goes for reducing
18:01
emissions. The world as a whole is
18:03
aiming for that zero emissions The unemployed
18:05
in the not too distant future in
18:07
order to avoid the worst effects of
18:09
climate change I looked attention like a
18:12
big problem Last, how do you think
18:14
about that challenges and you have a
18:16
sense of what you think the scale
18:18
of the increase energy demand for my
18:20
eyes gonna be So first will say
18:22
I think it's pretty clear that Ai
18:24
is going to present the biggest challenge
18:26
faced by the grid or by energy
18:29
providers. Yes, It doesn't mean
18:31
that we should try way from. I
18:33
don't think that of all the things
18:35
that you could compromise on right now
18:37
in terms of supplying energy A I
18:39
feel like one of the few things
18:41
in this world that actually give us
18:44
a path out of the climate crisis
18:46
that we put ourselves in a D.
18:48
I almost certainly begets few him among
18:50
many other really positive things. So if
18:52
you are going to rate limit something
18:54
I wouldn't start with a I for
18:57
there are from as reasons. That being
18:59
said, the. Inside Baseball suggests that
19:01
the scale of energy required is going
19:03
to outstrip the current supply by like
19:05
an order of magnitude current that what
19:07
we think we need is so much
19:10
more than the grid currently has, and
19:12
way more than we know how to
19:14
store and distribute to. the answers have
19:16
to be found and either major breakthroughs
19:19
or massive improvements in the efficiency by
19:21
which we train and run these models
19:23
and I think both one of happen.
19:25
It's been fascinating to me just walking
19:27
around. the conversations I've had with the.
19:30
Whole so far this time for the
19:32
kind of the two worlds colliding says
19:34
you have utility world which essentially and
19:36
heresy conservative by nature, no compromise like
19:38
to say that and have very much
19:40
got used to a period all essentially
19:43
zero demand growth for power in the
19:45
Us. so many things a little with
19:47
the demands of people coming to nothing
19:49
really to build always new data centers
19:51
and estate centers com added extra gigawatt
19:53
of load feel grid and not really
19:56
being a shock to people and people
19:58
having to kind of change. Hurt.
20:00
I'm think right different ways just about
20:02
the challenges they face. is that what
20:04
you're seeing as what if you may
20:06
want to be with other part of
20:08
what I talk about in circles like
20:11
this is this idea of said of
20:13
this abusing people of this malthusian approach
20:15
to the world which is that we
20:17
somehow convinced ourselves in the eighties and
20:19
nineties that the way to preserve our
20:21
species or to grow our species was
20:23
actually status that more energy was bad,
20:25
more population was sad, more consumption was
20:28
bad when in fact what we're. Discovering.
20:30
The opposite may be true and in
20:32
fact the opposite on a certain is
20:34
true. That thesis is is in the
20:36
beginning of the end and you start
20:39
to see that in how demographics play
20:41
out all over the world. energy is
20:43
just a proxy to growth, and an
20:45
increase in energy demands should actually signal
20:47
a lot of good things in this
20:49
guy. Selling singles, really exciting things, how
20:51
we produce and energy is the most
20:54
important thing and you know the word
20:56
nuclear is becoming more and more prevalent
20:58
and more more positive. And it's
21:00
pretty clear to most be working at
21:02
this point that the renewal of half
21:04
isn't actually sufficient to meet the demand
21:06
that we need to look beyond our
21:08
current sources of energy. Says actually solve
21:10
these problems and you bridge and fusion.
21:12
Do we actually need workable fusion power
21:14
to kind of unlawful responsibilities necessary to
21:16
make growth and energy demand a good
21:18
thing runs about? I mean, recently nuclear
21:20
To make growth and energy man a
21:22
good thing? not a bad thing. How
21:24
big of a solar field can you
21:26
actually build to serve the next generation
21:28
of Ai models? I. mean you can
21:30
build one the size arizona texas which
21:33
is what am i require how many
21:35
wind farms can you actually build before
21:37
it everybody will die as or whatever
21:39
right i should say whatever that's weapon
21:41
but there are real cost to the
21:43
traditional renewable energies we have the generation
21:45
nuclear reactors today that we know are
21:47
safe and we know our fish him
21:49
as good as they go out to
21:51
build them but it seems like this
21:53
has to details okay but the counter
21:55
argument to latvian is that the current
21:57
generation of nuclear technologies that are available
21:59
have been tried and essentially project by
22:01
the market. So we saw, for instance,
22:03
plans to build new SMRs in Idaho
22:05
last year, getting scrapped, that new scale
22:08
SMR project was abandoned. We've
22:10
had huge cost overruns, delays
22:13
to reactors being built in Europe and
22:15
the US in particular, projects going way
22:17
beyond schedule and so on. And if
22:19
you compare that with actually very low
22:21
costs of wind and solar power right
22:23
now and very low cost of natural
22:25
gas generation as well, it just seems
22:27
like it's really hard to economically make
22:30
the phase more nuclear. So how do you
22:32
get around that? I'm not an energy economist,
22:34
so I'll start by prefacing that. I will
22:36
say the equation seems untenable otherwise. Demand is
22:38
going to outstrip supply by such an incredible
22:41
margin that's simply saying we should build more
22:43
wind and solar energy
22:45
isn't sufficient. Even if
22:47
natural gas is truly abundant, you can only
22:50
put so much onto the grid. And ultimately,
22:52
look, I'm again not an economist, but it
22:54
seems like the solution has to
22:57
be some outside resource. So fusion power then
22:59
does seem to be the thing that sort
23:01
of unlocks all of it. This is where
23:03
I hang my hat. Right. And if you're
23:05
being cynical, then you say this is the
23:07
kind of the Deus Ex Machina, this is
23:09
the magic one, you know, how fusion power
23:11
that'll make everything all right. Sure. But you're
23:13
saying you believe that AI will be the
23:16
crucial tool that will help us get to
23:18
viable fusion power. What do you mean by
23:20
that? It seems almost certain that AI will
23:22
again be a beneficiary of fusion and a
23:24
benefactor of fusion. And so far as the
23:27
experiments that we run today, in general experiments
23:29
that we run in energy are becoming much
23:31
easier, much less expensive because of AI, we're
23:33
just able to do a whole lot more
23:35
testing. And you heard it this
23:37
morning, it's talked about all over the floor,
23:40
the things people are doing to actually understand
23:42
better how to move energy on and off
23:44
the grid. Fusion is no exception to that.
23:46
AGI almost certainly solves a lot
23:48
of technological breakthroughs. One of them is going to
23:50
be fusion. By the same token, fusion almost certainly
23:52
solves AGI because if we can solve fusion and
23:54
we can build a bunch of fusion reactors, well,
23:57
then suddenly we can build a bunch of these
23:59
massive machines. assuming we can get the
24:01
compute, which is a separate issue, and run them
24:03
abundantly. That is very interesting. It does seem to
24:05
be a real harmonization of interest then between people
24:08
who are interested in AI and
24:10
people who are interested in fusion power, a lot
24:12
of overlap there. It's not a coincidence that Sam
24:14
Altman is an investor in both Helion and the
24:16
CEO of OpenAI and sort of has called this
24:18
one of the major problems that we need to
24:20
solve, the energy deficit problem. And the other thing
24:22
I think a lot of people might say is,
24:24
well, this sounds kind of like science fiction to
24:26
us. And that combination in particular of AI plus
24:29
fusion, these are technologies
24:32
that people talk about for some kind
24:34
of far distant future. How
24:36
far away are they really, do you think? I
24:38
mean, is this, you know, we come back in
24:40
Disturbitec 2034, in 10 years time. I
24:44
think there's a fusion reactor in 2034. The
24:46
prevailing sentiment is that it takes six
24:48
years to build a fusion reactor from
24:50
time of discovery. We're probably
24:53
between three and five years out,
24:55
from discovery. I'll take the under.
24:57
I'm generally an optimist. I think
24:59
things move way faster on the large
25:01
time scale than we appreciate. And again, I
25:03
just remind everyone, try to explain the world
25:05
today to someone in 1950. I mean, the
25:07
things that we have discovered and the ways
25:09
we have moved ourselves forward is so remarkable.
25:11
Imagine the scenario in which you will struggle
25:14
to understand the world in 20 years. A
25:16
lot of things will have happened, principally among
25:19
them, I think, abundant energy and AI.
25:21
There is that good Bill Gates quote,
25:23
which someone, as you can probably
25:26
tell, they're intrinsically quite skeptical about a lot
25:28
of this. Well, you're British. So yeah, exactly.
25:32
It's not a country that tends to
25:34
be enthusiastic about embracing the future. But
25:36
Bill Gates Said famously, everyone tends to
25:38
overestimate what they can do in one
25:40
year. underestimate what can be done in
25:42
10. I Do think there is something
25:44
in that. Yeah. I Mean, you're talking
25:46
about a species that figured out how
25:48
to fly a plane and then put
25:50
people on the moon 60 years later.
25:52
We are quite good at things once
25:54
we give ourselves the permission to imagine
25:56
their possibilities. And I Think the public
25:58
perception is, actually. Critical to all
26:01
of this is why I talk about
26:03
we stop building mega projects as a
26:05
species and I think that give us
26:07
an incredible to service the know and
26:09
preferred. The interesting thing about mega projects
26:11
isn't simply that they stand and you
26:13
can remarket them if it reminds us
26:15
all the incredible things were capable and
26:17
these are really important to do to
26:19
remind ourselves how we can progress. And
26:21
thirty four the climate perspective. You could
26:23
say climate change. My view a mega
26:25
challenged exactly and eighty something. This is
26:27
the ultimate mega projects for for right
26:29
now. And given everything that makes
26:32
it so difficult to address with over technologies
26:34
that we have today, it says he would
26:36
make a big difference if we have. My
26:38
name may be the ultimate difference er ist
26:40
and well let's hope we can both be
26:42
back here. In turn your family for yearn
26:45
for the phone for I would love to
26:47
see whether that prediction came true. Eleven cycle
26:49
of junior. Thank you. For. Perspective
26:51
from one of the utilities the will be
26:53
making the practical decisions about how far to
26:56
adopt a I I talked to Quinn not
26:58
a Yamaha, the Senior Director of Grid Research
27:00
Innovation and Development Genome Id at Pg in
27:03
a weekend. For new to this a complex
27:05
homogeneous where things are have. One of the
27:07
big issues that everyone's been talking about here
27:09
at For Protected is a I my cat
27:12
of their to buy this is it seems
27:14
to be both the cause of and the
27:16
solution to all the problems of the gravy
27:19
and I but it's of the creating new.
27:21
Demands on the great lot of
27:23
and you know been granted it's
27:25
not. Also people are talking about
27:27
Ai as being a waiter manage
27:29
the great much more efficiently as
27:31
a deal with the various the
27:33
challenges including the city and do
27:36
authors and co on the they
27:38
renewable on creating whole new set
27:40
of challenges. What's your view on
27:42
I'm in how you thinking about
27:44
Ai at Pg in a what
27:46
do you see as the opportunity?
27:48
that fair amount of a potential
27:50
pitfalls and. way i have and
27:52
ml is a very big most
27:54
where that's going on on an
27:56
interesting i think it also has
27:58
a higher level called automation.
28:00
And so not all of that has to
28:03
be done by AI-ML, there's automation
28:06
that can happen without that. And so
28:08
the broader question needs to be where
28:11
do we need to automate? And I
28:13
think from our perspective is do we
28:15
understand all of the different type of
28:17
use cases that we as a utility
28:19
want to create further and further automation?
28:22
Whether it's through AI, whether it's through ML,
28:24
whether it's generative AI, or just pure nuts
28:26
and bolts automation. Once you figure out
28:29
all of those use cases that you
28:31
could particularly envision, the second question is,
28:33
is this really a technology problem? I
28:35
think too many utilities jump directly into
28:38
AI-ML if it's going to resolve all
28:40
of your related issues. And when you
28:42
peel that back, you might actually find
28:44
that yeah, you can do
28:46
AI-ML for this particular thing, but it's a
28:48
process issue stupid. Right? And it's like, well
28:51
in that case, are we just applying technology
28:53
to solve a higher level issue? Once
28:55
you peel that back, then it's really
28:58
about well, where are we going to
29:00
get the most bang for the buck
29:02
to apply AI-ML to achieve our 10-year
29:04
targets on our strategy or our 30-year
29:06
objectives? And we haven't gone through that
29:08
exercise. What ends up happening is you
29:11
get all these spot AI-ML conversations. If
29:13
we used AI-ML in this area, it
29:15
would be amazing. Or here, here, I think
29:17
we need to take a step back and ask ourselves
29:19
what are going to be the biggest benefits? Now
29:22
I can tell you, generally speaking,
29:24
that certain foundational things need to
29:26
happen before you do more of
29:28
the advanced AI-ML. So I'll give
29:30
you an example. AI-ML can really
29:32
help you with identifying where your
29:34
assets are and cleaning up your
29:36
asset registry. Once you get there,
29:39
then you can start stacking on all
29:41
these other AI-ML related use cases. But
29:43
there are certain situations where you just
29:45
got to get the basics right and
29:47
then be able to go into
29:50
how can AI-ML help you with
29:52
inspections? How can AI-ML help you
29:54
with optimization of dispatches? How can
29:57
AI-ML help you do simple estimations
29:59
with without having to have an
30:01
engineer draw up those
30:04
drawings. There are definitely places
30:06
that we can target, but we're
30:08
going to undertake an effort to
30:10
really understand what is the world
30:12
of the use cases, and then
30:14
create a multi-year view into what
30:16
is going to have the biggest
30:19
bang for the buck. How
30:21
do we think about data? How do we think
30:23
about data security? How do we think about cyber
30:25
security? These are all things that we're going to
30:27
have to tackle. Right, so as you say, when
30:29
you're thinking about security, privacy, some of those
30:31
issues, that's really where you start
30:33
to get into a discussion about the pitfalls and
30:36
the risks in AI. What
30:39
do you see those as being? I mean, is there a
30:42
danger in creating a system?
30:44
I mean, as you say, you think about
30:46
automation at large, a system
30:48
that kind of escapes from human control.
30:50
I'm not so concerned about a system
30:53
that escapes from human control because I
30:55
think that we're going to step into
30:57
the AI and ML in a very
30:59
methodical way. I'll
31:01
give you one example where AI and ML could
31:03
be very detrimental to a particular
31:06
utility's business, and let's take a look
31:08
at it through the inspection process. Right
31:11
now, utilities have the ability due
31:13
to the cost of photos becoming
31:15
significantly cheaper and visualization becoming significantly
31:18
cheaper that you can get imagery
31:20
of all of your assets almost
31:22
all the time. Now, AI
31:25
and ML done incorrectly could create a
31:27
whole bunch of false positives that
31:29
you now need to go track down. If
31:32
you don't do this in a
31:34
very careful way, you can't unsee
31:36
some sort of degradation on the
31:38
system. However, if it's wrongfully prioritized,
31:40
you put yourself at significant risk,
31:43
both from a cost perspective and a
31:45
liability perspective. So really, how do you
31:47
make sure that you're training it, retraining
31:49
it, getting the right results before
31:51
you start to really roll it out across
31:54
the system? And this is where, I think
31:56
one of the biggest things that utilities struggle
31:58
with is this fast pilot, fast. fail,
32:01
right, mentality, where you take
32:03
particular use cases, you do
32:05
a quick hyper agile development
32:07
on that, you test that out and if it doesn't
32:09
work, you move on to the next thing. Sometimes
32:12
what we do is we tend to get too
32:14
invested in our ideas, that we have a hard
32:16
time letting go and we just pump money into
32:18
a failed idea. And I think with this AI
32:20
ML, it's going to be a lot of that.
32:23
We're going to try things. Something's going to work,
32:25
something's not going to work. And as you kind
32:27
of take a step approach, really
32:29
try to make it so that
32:31
the outputs that you're getting are
32:33
having the material benefit that you
32:35
want for your 10 year strategy
32:38
is going to be key. Now, is
32:40
there a world where AI ML
32:42
will take over what the utilities are going
32:45
to do? I mean, who
32:47
knows what the 30 year future is going
32:49
to look like at the current time? I
32:51
can't see it. Anything's possible with
32:53
this technology. Skynet may take over the
32:56
world. I mean, there are endless possibilities
32:58
of where this can go. I
33:00
think what we're going to do is
33:03
we're going to walk into this very
33:05
deliberately, push the realms of what we
33:07
can do, but always keep in mind
33:09
the safety of our customers, the safety
33:11
of our grid. That is always going
33:13
to be at the forefront of what we do. And
33:15
if we do that, then I believe that we can
33:17
create an AI ML vision and a
33:19
world that's really going to benefit our customers. Quinn
33:21
Nakayama, thanks very much for joining us on the
33:23
energy game. I appreciate it, thank you. For more
33:25
of my conversation with Quinn Nakayama, where we talked
33:28
about the new challenges for the grid created by
33:30
electric vehicles, look out for a special edition of
33:32
the energy gang coming soon. Now,
33:34
Quinn made the good point that AI applications
33:36
are only a subset really of the wider
33:39
range of technologies for automation and digitalization of
33:41
the electricity system. To discuss
33:43
those, I talked to Tom Dietrich, who's chief
33:45
executive of Itron, which is wise technologies for
33:47
utilities and cities to manage energy, water, and
33:50
traffic. And I talked to him about the
33:52
adoption of those technologies and what might be
33:54
holding them back. Tom, good to see
33:56
you again. Ed, great to be back with you. Absolutely, yeah, great
33:58
to have you back on the show. So, interested
34:01
first of all perhaps in what
34:03
you're hearing while you've been here,
34:05
as you're kind of walking around
34:07
meeting people, meeting other people in
34:09
the industry, meeting customers I'm sure
34:11
as well, and the
34:13
conversations you're having here, what are people
34:15
talking about? What struck you
34:17
as interesting? So, there are two mega trends
34:19
that I think are really interesting to point
34:21
out. One on the customer side, one on
34:24
the technology side. I'll start with the customer
34:26
part first. Consumers are
34:28
really, really struggling to keep up
34:30
with the pace of technology that's coming
34:32
out them. There's so many different
34:34
ways to attack a problem. There's so many
34:36
problems that they have that are unsolved yet
34:39
today, but technology is coming at them much
34:41
faster than they can absorb it. And the
34:43
corollary to that, and then the second part
34:45
of my answer, is on the technology side,
34:48
that technology companies like I-Tron have realized
34:50
that, oh, this is a real problem.
34:52
It's a problem for the customer because
34:54
they aren't solving the problem for their
34:57
consumers, you and I, but it's also
34:59
slowing down our business growth potential. So
35:01
what are we doing in response to that?
35:04
It is spending more time collaborating among
35:06
ourselves. Let's reintegrate a solution. Let's have
35:09
two companies that work side by side
35:12
inside of the utility to supply some solutions to
35:14
them, but let's do that before we get to
35:16
them. So we present a united front by the
35:18
time we get to the customer, the utility in
35:21
this particular case. One of the
35:23
big examples of that is we've had some
35:25
announcements running up to this event where we
35:27
were going to preintegrate our solution for grid
35:29
edge intelligence with Schneider Electric,
35:31
who does ADMS, or
35:34
with GE Vernova, who does a similar
35:36
ADMS kind of solution. But if we
35:38
can't combine... Sorry, ADMS. Sorry, too many
35:40
acronyms, but Advanced Distribution Management System. So
35:44
I-Tron works close to your dwelling, close
35:46
to the home, higher up in the
35:48
transmission line and in the stack, higher
35:50
voltages where these other companies work. If
35:52
we can come and show up with
35:54
the solution together, it helps the customer
35:56
deal with this problem that they have
35:58
to provide good and reliable service. but
36:00
also deal with the pace of technological change. That's
36:03
very interesting, yeah. So going back to the first
36:05
point you were making, as you say, about the
36:07
two big trends, that's something I've also been really
36:09
struck by, and it's one of these kind of
36:12
terrible cliches that people come up with about, oh,
36:14
there's so much change and things, you know, we
36:16
live at a revolutionary time, et cetera. But actually,
36:19
in this industry, the utilities in particular, who
36:21
are crucially your customer base, it does really
36:23
seem to be true. There is something that's
36:25
very different about the world we live in
36:27
and the number of things I think of
36:29
are, I suppose, two crucial ones. One
36:31
is that we're on the demand side, is
36:34
that after a long period of essentially
36:36
static demand in the United States for
36:38
power, suddenly that's really kind of taken
36:40
off. Various trends going on here, data
36:43
centers for AI being one big one,
36:45
the revival of manufacturing industry, new plants
36:47
bringing up, particularly chip plants coming has
36:50
been a big thing, but also obviously
36:52
demands for literally heating, electric vehicles becoming
36:54
very significant. And then on the supply
36:56
side, obviously the well-known challenges of managing
36:59
variable renewables on the grid, which must
37:01
be a completely different environment for utility.
37:03
Is that how you're saying it was?
37:05
Absolutely, I identify with everything that you
37:07
said there. I had a customer event
37:09
yesterday, so picture this, a room with
37:11
about 100 to 125 customers inside
37:16
of it, and they were just talking among
37:18
themselves. And one customer said, I see my
37:20
demand doubling or perhaps tripling over the next
37:22
10 years. So all we have to do
37:25
on the generation side is build what we
37:27
built over the last 100 years in
37:29
the next Decade. Oh, that doesn't sound hard
37:31
at all.. And What was driving that
37:34
demand increase was exactly the things you
37:36
were talking about. It is manufacturing, it
37:38
is AI driven data centers, and it
37:40
is electric vehicles, which are wildly different
37:43
loads, especially EVs, than what the utility
37:45
is used to dealing with. And So
37:47
in terms of tackling those challenges, what
37:50
can technology do to help? And what
37:52
are the crucial innovations? And As you
37:54
say. So there's one particular thing you've
37:56
just been talking about in terms of
37:59
suppliers getting together. The form an alliance
38:01
isn't partnership to provide and a package
38:03
solutions utilities but one of the things
38:05
that really have make a difference and
38:07
will help them deal with they've as
38:09
like what is genuine. The I think
38:11
of a new well here I will
38:13
focus my answer on the distribution. Great
38:15
many that the stuff that's close to
38:17
where you and I were called to
38:19
our house and I think all of
38:21
these two seperate energy resources. electric vehicles,
38:23
rooftop solar, batteries behind meter. All of
38:25
those resources themselves are starting to pop
38:27
up in uncontrolled way from the utility
38:29
for stalking. Her first and foremost
38:32
honey provide visibility to the utility Were
38:34
those things are that is truly truly
38:36
helpful for the Italy Fab. I know
38:38
where they are that I can reach
38:40
out to that customers. Hey, would you
38:42
like to enroll in my program to
38:44
control charge and secondarily as understand what
38:46
capabilities you have a utility to deal
38:48
with That Hey will that person allow
38:50
me to control when their vehicle is
38:52
charged? Be able to control it turning
38:55
on and off when you need our
38:57
on the grid. and finally optimize, Please
38:59
Charge This car. When the sun is
39:01
shining on that roof. So this for
39:03
progression of visibility capability control is it
39:05
has. And the fact
39:08
that utilities don't have that have a
39:10
moment I think is often spoken to
39:12
people when they think about it need
39:14
to come from another in disarray. and
39:16
if the girls were you in the
39:18
business is serving the consumer the amount
39:20
virtues of it is just don't know
39:22
about what mechanism doing And from the
39:24
dimona she'd been increasing think of a
39:26
distributed energy supplying power back to the
39:28
great of well it may they. Agreed
39:30
was designed as a one way push
39:32
it was always assumed would just push
39:34
our down and will be there and
39:36
suddenly the world changes. Aware that isn't
39:38
relief economically possible a feasible and now
39:40
how do you deal with that from
39:42
a consumer standpoint is shocking and services
39:44
that was when you for cure another
39:46
service in your life. As a consumer
39:48
you ordered blueberries for delivery as an
39:50
example you new in real time when
39:52
a driver was leaving the restaurant and
39:54
his or her name and how much
39:56
it was gonna cost and when they
39:58
were showing up on. your flat. The
40:00
last time you dealt with your power company may
40:02
have been lights are off and you're ringing them
40:04
up saying, when are my lights coming back on?
40:06
And they may not have even known the lights
40:08
are out. That service imbalance is a huge
40:11
barrier for the utilities to overcome and something
40:13
that I think technology like Edge Intelligence that
40:15
I talked about can really help. So it's
40:18
not really a surprise that these changes are
40:20
coming, right? I mean, people have seen this
40:22
at least kind of 20 years in the
40:24
past, 30 years in the past. People could
40:27
tell that this was the way the grid
40:29
was likely to evolve and people had that
40:31
sense that the grid of 100
40:33
years ago was not going to be fit for
40:36
purpose in the 21st century. How
40:38
rapid has progress been towards adopting
40:41
these new technologies? It feels to me like
40:43
actually the industry has been often much slower
40:45
than it should have been and
40:48
now things which have been seen sort
40:50
of as looming future possibilities heavy towards
40:52
us are now actually very real and
40:54
are really affecting people and
40:56
the industry is just not really prepared in time.
40:58
Is that fair with that? I think
41:01
that's sadly extremely fair. I think
41:03
grid modernization now needs to happen
41:05
in years, not decades. And we're
41:07
starting from a disadvantaged position. We
41:09
are behind. There's lots of reasons,
41:12
the regulatory models that are at
41:14
play, how traditional buying patterns have
41:16
worked. Utilities are used to buying
41:18
things and putting them in
41:20
the field and depreciating them over 10 or
41:22
20 years. But if you don't know what
41:24
happens next month, let alone five years from
41:27
now, how do you plan for that? So
41:29
there's many things that need to change about
41:31
the buying model, the regulatory model and really
41:33
the technology itself to be able to help
41:35
solve problems. And what more could be done
41:38
to help expedite this change and to get
41:40
these new technologies deployed across the US industry?
41:42
Is this something the policy needs to help with
41:44
more? Is it something that needs to change
41:46
in the system of regulation? What can
41:49
really be done to sort of accelerate this? And as
41:51
you say, to get this change accomplished in years, which
41:53
is what we need to do. My
41:55
off-the-cup reaction is can I choose all of the
41:58
above as the answer? I think that there Our
42:00
policy changes, so governments can help
42:02
drive policy in a way that makes
42:04
it easier for utilities and for
42:06
technology companies to move faster and drive
42:09
better solutions to one. I
42:11
think regulatory policy can be adapted
42:13
to make it easier to deploy
42:16
new technologies. Today, it is very
42:18
difficult to deploy fast solution software
42:20
as a service-based solution because it
42:23
doesn't cleanly fit into the model
42:25
most utilities have. A
42:27
change in the regulatory structure to
42:30
allow capitalization of software as a
42:32
service as an example is something that's
42:34
possible that would also be helpful
42:36
to move the technology forward itself.
42:39
And then on the provider side, I think
42:41
we have to continue to push the envelope
42:43
and do a lot of pre-integration to make
42:45
it easier for our technologies to be consumed.
42:48
And as you meet people, walk the corridors
42:50
here, listen to the conversations that are
42:52
going on, has it left you more
42:54
optimistic that the progress is going to be made
42:56
to update the grid the way we're going to
42:58
need to? I would say that I'm certainly more
43:00
optimistic than I was 48 hours ago when
43:03
I set foot on the ground, not because
43:05
we've got the answer in hand today, but
43:07
I see a lot of like-mind thinking and
43:09
talking about the same things, and that really
43:11
is step one to solving the problem. Tom
43:14
Duchik, thank you very much indeed. Thanks for talking to us
43:16
about that today. Yeah, I appreciate your time. Thanks for having
43:19
me. And lastly, I talked to Anthony Allar, who's the head
43:21
of Hitachi Energy in North America. Hitachi
43:23
is an important supply of transmission equipment.
43:26
He was a pioneer of high-voltage direct current
43:28
transmission lines all the way back in 1954. Distributec
43:32
is a great opportunity for him to meet
43:34
customers that use Hitachi's technology. I'll
43:36
ask him what he'd been talking to them about. Thanks
43:39
very much for joining us on the Energy Gang. Nice to meet you. Good
43:41
to be here. So I wanted to
43:43
start off talking a bit about Distributec
43:45
and your presence here. So Hitachi Energy
43:47
is a very important company in the
43:49
transmission industry. You were the pioneers of
43:52
HVDC, high-voltage direct current transmission
43:54
lines. That's I think
43:56
70 years old this year, right? 1954,
43:59
the first one. Phil and Co.
44:01
This is supposedly the leading event
44:03
or admission and of commission and
44:05
I like to have in North
44:07
America. Whoa! A hobby! A great
44:09
opportunity for them to get together
44:11
for have the hands of com
44:13
Haitian, what are you really want
44:15
to talk to people about While
44:17
you'll have what is coming into
44:20
focus on the conversations you're having
44:22
at the city has to be
44:24
questioned assuming se han hi to
44:26
them on if you are think
44:28
harrys like com number of. Composition.
44:30
Of we love in swimming year with customers
44:32
I said it know couple of feel lonely
44:35
is still very high for good reason than
44:37
that in the mind of a catharsis of
44:39
later know. I know now
44:41
for a number of years with him
44:43
behind or equipment going up with inflation
44:46
in fact know how the height of
44:48
the point or solution so those conversational
44:50
still need it because he correctly with
44:52
seen that those challenges warring parties hurt
44:55
and part of the technology provider like
44:57
us on our hundred ethical audible order
44:59
equipment for another the second way Korea
45:01
where you're coming into it's important for
45:04
us to really share what we see
45:06
globally with with you North America with
45:08
our customers so that they can then.
45:11
Into today and get ready at did
45:13
ahead of the curve for that won't
45:15
have another topic is William Digital Solutions
45:18
which is going to be that as
45:20
being loaded more on the diesel hide
45:22
and remains or as in Haiti oh
45:25
good Rhythm of Little High Horse also
45:27
is that important topic or we know
45:29
that the great Today needs to be
45:31
more flexible and more resilient to be
45:34
able to accommodate more renewable energy. So
45:36
Digital is is an enabler whole of
45:38
the national so far as. A. T
45:41
topic that we have a look
45:43
at the dogs would be to have
45:45
been room for the center or consider
45:47
her my ankle. Much of the want
45:50
to take those in term and thinking
45:52
about supply chain challenges for so you're
45:54
saying the thing hard getting were film
45:57
and supply chains. Delays are getting
45:59
longer. It feels like this is
46:01
something that people have been talking about really
46:03
since the pandemic ended and since investment activities
46:05
started to pick up. And you might have
46:07
hoped perhaps that some of the issues with
46:09
the supply chain would have started to get
46:11
fixed by now. But no? Yes
46:14
and no. So the yes is there's hope.
46:17
So I would say the first type
46:19
of equipment which has been impacted by lead
46:21
time and if you ask around, what will
46:23
come to the mind will be transformed. Here
46:27
now fast forward now to three years.
46:30
Now what we see is we've seen a
46:32
number of announcements in terms of investment in
46:34
addition capacity. We have ourselves made a couple
46:36
of investments here in North America. So
46:38
here that additional capacity, some is already
46:40
online, some of ours is already online,
46:43
some of the industry is already online,
46:45
some is not. So I think the
46:47
market sees that we're making progress but
46:49
still the demand is much higher
46:51
than the supply chain. So we see improvement but we
46:53
still see a lot of it clearly. And just particularly
46:55
on that issue of transformers, there's been a lot of
46:58
discussion about that. Obviously people have become very kind of
47:00
fixated on the question of bottlenecks and the
47:02
transform supply chain. But there then it sounds
47:04
like you've got a reasonably positive hopeful message
47:06
for people. You think that as you say
47:08
that new capacity is getting built, we
47:11
are going to see those bottlenecks start to ease. I
47:13
wouldn't go that far. I think
47:15
we see capacity coming in. But still
47:17
the new capacity is still below the
47:19
demand. So we still expect to
47:22
see a lead time at a much higher
47:24
level compared to what they were, for example,
47:26
for your So
47:38
that's the, for the yes, it's getting better. The
47:40
no, it's not getting better. It's now what we
47:42
see is other type of equipment where now we
47:44
see the lead time increasing. And here this is
47:46
where we have a lot of conversation with the
47:48
customers about that to make sure that they understand
47:50
so that we can work with them on the
47:52
planning of the- And What
47:54
types of equipment then are affected by that? For Example,
47:56
circuit breaker is one of them. And Again, are we
47:59
going to ultimately see, They would recommend a
48:01
capacity and I would assume we'll see
48:03
the same dynamic and it also affects
48:05
time similarly that I can. It's a
48:07
com Yes, because Stages of Last Resort.
48:09
Like are you okay You want to
48:11
build a new one or two on
48:13
earth for the capacity you need to
48:15
find. People are getting too old to
48:17
find the the machine as he needs
48:19
or another machine. Or
48:23
harm or like soy turkey and
48:25
a couple that you are the
48:28
core subjects. Withdraw my girl to
48:30
pick up on the other big
48:32
topic you've had her parents talking
48:34
about the you're talking about a
48:36
button digitalization. It has become clear
48:38
up to convey hope that I've
48:40
had conversations with and how they
48:43
offer here van and is a
48:45
enormous interesting location as part of
48:47
the solution to this huge you
48:49
set of challenges that he is
48:51
anything with how effect that. Is
48:53
probably the only for broken
48:55
were bought coping with increased
48:57
proportion of February or with
48:59
on the grid a grief
49:01
distributed energy resources you kinds
49:03
of demands you load particularly
49:05
in North America from the
49:07
present Israel I need Manhattan
49:09
and a his info on
49:11
how little pressures hitting the
49:13
industry thing. When you
49:15
talk about visualization of thing the thing
49:18
that people are looking at what does
49:20
that mean physically and have the know
49:22
what they've chosen as big as a
49:24
question is what robin what you're trying
49:26
to solve with digital who's and I
49:28
thought into a couple of problem though
49:30
was it industry trying to solve one
49:32
of them his residency or we also
49:34
see any crazy know extreme weather events
49:36
are so how do you for so
49:38
hostile how did you into people how
49:40
did you bring more flexibly you agreed
49:42
to be moral of this is the
49:44
type of solution. of problem than
49:47
for solution in the were little connected
49:49
products and in software come in to
49:51
try to provide a solution then when
49:53
you talk about renewable integration right how
49:55
do you make sure that in real
49:58
time you able to keep your supply-demand
50:00
balance on your grade when you know that
50:02
the sun and the wind change so far.
50:04
So that's another problem that can find a
50:06
solution to a digital solution. And the lead
50:09
goes on and on and on, those are
50:11
kind of two examples of where a digital
50:13
solution can support those problems. Right, and though
50:15
beyond that, it does seem very clear that
50:17
there are some problems for which digitalization and
50:20
the kind of technologies we've just been talking about are not
50:22
a complete solution. And I think no
50:24
one would deny that when you go around here,
50:26
there's a lot of people here at S U
50:39
B Tech with various kinds of technology and
50:41
digital solutions that they are offering for different
50:43
aspects of the power grid. That's
50:48
presumably, from the degree with providing that
50:50
physical infrastructure is very much the business
50:52
that the searchy energy is in. physical
50:56
infrastructure. certainly
50:59
true in Europe, it's true very definitely
51:01
in the UK. It's just very, very
51:03
difficult to get more team-commissioned lines put
51:05
in. Why
51:09
is it? That's a difficult question. I think there are a couple
51:11
of things. infrastructure is not as a
51:13
business anymore. It's a very difficult one. It's a very difficult one.
51:16
It's not a very difficult one. It's a very
51:18
difficult one. It's a very difficult one. It's not a
51:20
very difficult one. It's a very difficult one. It's a
51:22
very difficult one. I think there are a couple of
51:24
things. I think the first one that we don't see
51:26
in the United States is a national transmission planning mechanism
51:29
compared to what maybe we see in other parts of
51:31
the world. This is something that there's
51:33
a lot of talk about it, and we certainly
51:35
believe that it is the right thing to do.
51:38
Here, we hope that we will move
51:40
into this direction. It's
51:42
number one. Another challenge is around permitting. Here
51:45
also, this administration has been trying
51:47
to address that topic, and we all know it's a complex
51:49
topic. Again,
51:52
we commend what they've done so far, and
51:54
we really hope that collectively we can find
51:56
a solution to this. the
52:00
speed out which we also need to deploy. Assuming that
52:02
we have all the permitting in the world, all the
52:04
planning in the world done, then you still need to
52:06
execute on those projects. And I think here, if you
52:08
think about it, there's a number of statistics that say
52:10
we need to double, triple the grade. But let's assume
52:12
that, let's just take up the argument, we need to
52:14
double the grade in the next 15, 20
52:17
years. It has taken us almost
52:19
more than 100 years, right? The grade was first
52:21
created in the other part of the world
52:24
at the end of the 19th century. So
52:26
here, now we're talking about rebuilding the same
52:28
grade in the next 15, 20 years or so. So
52:32
we need to think differently. The technology is here,
52:35
so we have the chance to have the technology, but
52:37
we just need to think about it differently.
52:39
We cannot refuse the same scheme and assume
52:41
that we can go at a much faster
52:44
pace. So here, we really need to think
52:46
about new business models. As an
52:48
industry, it's not only one or two stakeholders,
52:50
the entire industry, the utilities, the technology suppliers,
52:52
the government. All the stakeholders need to come
52:55
around the table and find different ways, right?
52:57
And we see different ways in all parts
52:59
of the world, right? We see in Europe,
53:01
for example, a different type of business model
53:04
where some utilities have decided to find long-term
53:07
frame agreements with companies, for example,
53:09
like us, from technology. The positive
53:11
result of this is that that gives them
53:13
security of supply for the next 10 years
53:15
or so. And for us,
53:18
that allows us to know that we
53:20
have a security of demand and in
53:22
return allows us to accelerate our investment,
53:24
right? So I think this is an
53:26
example of new type of business model
53:29
that I think we need to look at more closely
53:31
in North America. And yes, the regulatory construct is a
53:33
little bit different in this part of the world. So
53:36
we cannot do a cut in terms of what we
53:38
see in our part of the world, but we need
53:40
to look at it and try to see what are
53:42
we learning from that one example I get from Europe
53:44
and how could we tweak that to apply it to
53:46
the US in a way that works with the US
53:48
regulatory construct. That is really interesting.
53:50
How optimistic are you that that kind of
53:52
new thinking in business models can be made
53:54
to think? I mean, just because
53:57
the reason I raise that
53:59
question The details like that is
54:01
a very conservative industry. The in Sleep
54:03
doesn't change much as essentially you're talking
54:05
about the physical infrastructure being the same
54:07
as it was in the late nineteenth
54:09
century. As you the business model is
54:11
also oversight. Music was in the late
54:13
nineteenth century. have not a whole months
54:15
and for good reason. People are risk
54:17
of her and competitive and they want
54:19
to shake things up in ways that
54:21
might lead to the light's going. All
54:23
so how could you persuade people who
54:25
have a duty to think differently? Your
54:27
question I am fairly optimistic for to.
54:30
Or think number one this country has demonstrated
54:32
than when there is a problem there was
54:34
a solution and at some points all the
54:36
stakeholders get together and really of find a
54:39
solution but also we'll adding very different proposition
54:41
today with our customers over to the courses
54:43
and we I'm in four years ago I'd
54:45
and us really a command or customers year
54:48
in terms of acknowledging that we need for
54:50
a qb to or differences and the said
54:52
that we are collected in making flooding here
54:54
Adidas very encouraging it was still have a
54:57
new ways to go but I'm looking his
54:59
signature will. Give their say think becomes
55:01
a distributed again in five years' time.
55:03
Authentic distributor Twenty Twenty Nine You think
55:05
things are going to seem pretty different
55:07
than I would hope that. I will
55:09
hope that we will see new business
55:11
model that we see a regulation which
55:13
has evolved and enable a faster investments
55:16
and I'm sure so we will see
55:18
a lot of a Ai solution on
55:20
the floor like we already saw to
55:22
see this year. but I'm sure in
55:24
five years from now there will be
55:26
a not a a buzzword there would
55:28
be around just a fact. That set
55:30
back when to think about the
55:32
broader context of this, it's now
55:35
one of those cliches. the past
55:37
five years ago Randall of People
55:39
Face know transition without transmission in
55:41
that building you need capacity is
55:43
absolutely essential to shifting the electricity
55:45
system over to low carbon technologies
55:48
renewables another kind. So this is
55:50
something of last. If we're serious
55:52
about addressing climate change, we really
55:54
do have to make absolutely yeah,
55:56
there was no transition with transmission.
55:58
I fully agree. and I think
56:01
now what's very encouraging is again going
56:03
back in time three years ago people
56:06
in the industry knew about it But
56:08
today this is public knowledge This is
56:10
something that you see in major newspapers
56:12
and that really the proof that now
56:14
it's really a Fight that
56:17
is well accepted and understood, but now we need to
56:19
get to work and see the yellow. Thanks for much.
56:21
Enjoy. Thank you So that's it for
56:23
us on the first full day of distributed Fascinating
56:26
to talk to all these people about the
56:29
problems they're facing the solutions They're advancing
56:32
when you stand back and think about the energy
56:34
transition at the highest level I think it's very
56:36
easy for the debate to get pretty abstract But
56:38
what we have at this event is thousands of
56:41
people who are working on the challenge of decarbonizing
56:43
the energy system And they're creating
56:45
real change in the shape of metal
56:47
and concrete and software such a really
56:49
inspiring thing to see I think many
56:52
thanks to all of our guests have been talking to
56:54
us about what they're doing Many thanks to our producers
56:56
Dan Cottrell Toby Biggins Gilchrist and Sam Nash and
56:59
above all of course many Thanks to all of
57:01
you for listening Please do keep your feedback coming
57:03
look out for the rest of my conversation with
57:05
Quinn Nakayama of PG&E You won't
57:07
want to miss that. I think it is really
57:09
a fascinating discussion And we'll
57:11
be back soon with all the latest news and
57:13
views on the energy transition until
57:16
then. Goodbye
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More