Latest / Make Water Work Podcast / Peter Yolles & Clara Drysdale: Data Center Innovation & Cooling the Future | Make Water Work 024

Transcript

Peter Yolles & Clara Drysdale: Data Center Innovation & Cooling the Future | Make Water Work 024 · Make Water Work Podcast

0:00
There are NOS's out there that are sort
0:02
of watching what's happening. Data
0:04
center watch, right, is a community
0:06
group that shares information and
0:08
they're out to stop or delay data center
0:11
development. So, it's in the interests
0:13
of everyone to um make data data centers
0:17
more water efficient. Welcome to the
0:19
Make Water Work podcast where we dive
0:22
into the people, ideas, and innovations
0:24
shaping the future of water. I'm your
0:27
host, Isaac Pellerin, and alongside my
0:29
co-host, Megan Glover, the founder of
0:32
120 Water, we bring you insights from
0:35
hundreds of incredible conversations
0:37
we've had over the past 9 years with
0:39
inspiring and energizing leaders in the
0:42
water industry. Each week, we share
0:45
those stories and conversations with
0:47
entrepreneurs, community champions, and
0:49
innovators who are making water work for
0:52
everyone. Whether you're a water
0:54
professional, a curious innovator, or
0:56
simply passionate about our most
0:58
precious resource, this podcast is for
1:00
you. Let's get started.
1:04
Welcome to the Make Water Work podcast,
1:06
where we celebrate the innovators and
1:09
the leaders and the servants who are
1:11
serving and making water work. Um, we
1:15
love this time that we get to have
1:17
conversations with amazing guests. We've
1:20
had so many amazing guests so far and we
1:22
have two amazing guests today. So,
1:24
Megan, let us know um who we're speaking
1:28
with today.
1:28
>> Yes. Well, this topic um is so important
1:32
to us and we have been so uh nerding out
1:34
on this topic so much that it required
1:36
us to have two amazing guests for
1:38
today's podcast. So, thrilled to bring
1:40
to the show um the authors of a report
1:44
that we'll talk about here in a minute
1:46
um Claraara Dale and Peter Yalis. and um
1:50
they just authored a report called um
1:52
the data center innovation report and we
1:55
will link to it in the show's um notes
1:57
and we'll make sure that we share but um
2:00
essentially in my opinion it's the best
2:02
thing out there that I've seen that
2:04
really breaks down the economy that's
2:07
going to be driven from data centers not
2:09
only from like what's the so what like
2:11
what's the the big number impact but
2:13
what are the technology sectors and
2:15
segments um and they they do a really
2:17
good job of framing out like the
2:19
technologies and their level of um
2:22
maturity and and so on and so forth. And
2:24
so I'm just super excited to not only
2:26
talk about the report, but to dig into
2:28
some of those findings with these uh two
2:30
incredible uh guests.
2:32
>> That's awesome. Well, let's bring them
2:33
in. Welcome, Peter and Claraara. Um
2:36
welcome to the Make Waterwork podcast.
2:40
>> Thanks, Isaac and Megan. We're delighted
2:42
to be here.
2:44
>> Yes.
2:45
Well, we are going to just pick this
2:48
report um pick uh to pieces because
2:50
there's so much goodness that we need to
2:52
cover um in the show. But first, we
2:55
can't deviate too much. Um we've never
2:58
met uh folks in water that don't have an
3:00
amazing story of how water found them.
3:02
So, we would just love uh briefly to
3:05
understand Clara and um Peter, what's
3:07
your story? How did water find you both?
3:09
Claraara, do you want to start?
3:11
>> Yeah, of course. And first of all,
3:13
thanks for hosting this podcast. I
3:15
really respect the work you do putting
3:17
these conversations out there. Um, for
3:20
me, I think I just kind of grew up in a
3:22
very water conscious place, Colorado, at
3:25
a water conscious time. Um, growing up
3:28
in school, they talked a lot about
3:29
climate change and water and the
3:31
Colorado River. So, I just got into that
3:34
at a young age. I started water testing
3:37
when I was 15. And then coming now as an
3:40
undergrad at Stanford, I've always been
3:42
interested in climate, I've learned that
3:45
water is one of the first ways that
3:46
climate change is going to um create
3:49
impact in the world and realize itself.
3:51
So that's why I'm interested in water.
3:54
>> Love that.
3:56
>> Yeah. Like Claraara, I grew up in the
3:58
west um in California. Grew up my family
4:01
were sailors. So I grew up my earliest
4:03
memories are on a sailboat. So I watered
4:05
sort of in my in my jeans, if you will.
4:07
And uh and also I grew up in the in the
4:09
time of scare water scarcity in
4:11
California where we had lived through
4:14
droughts and it was always uh front page
4:16
headlines. Um and so just having
4:19
awareness that water was a political and
4:22
a business issue uh really infused me
4:24
with um an interest in seeing how water
4:27
is you know really part and parcel of
4:29
our economy and the way we live.
4:31
>> Absolutely. Absolutely. and and and
4:33
Peter, you have such a wonderful story
4:36
um in your background as an entrepreneur
4:38
in water and I'd love um for you to
4:40
share also with the audience um you know
4:42
the the founder CEO of WSmart that was
4:46
uh acquired by Vertex One and um
4:48
remember meeting Peter during Imagine
4:50
H2O and listening to that story and you
4:52
were s certainly one I put on a pedestal
4:55
uh starting 120 waters and would you
4:56
mind just sharing a little bit about
4:58
that background as well?
5:00
>> Yeah, absolutely. Before founding Echo
5:02
River Capital, I I founded Water Smart
5:05
Software back in 2009. We were really
5:08
the first software as a service offering
5:11
to drinking water utilities really
5:13
trying to help customers reduce their
5:16
water use, become more water aware and
5:19
help them with um paying their bills. So
5:21
we um today water smart software
5:24
operated by vertex one uh it serves over
5:27
200 drinking water utilities today um
5:30
with water use intelligence uh and data
5:33
analysis software. So it was a really
5:35
terrific run for 10 years. WSmart was a
5:38
winner of the very first Imagine H2O
5:41
competition back in 2010. Um and we have
5:44
served each other and I still serve as a
5:46
a mentor and a and a judge for the
5:48
Imagine H2O competition today. So, it's
5:51
great um that 120 water and other um
5:54
companies um are finding their way to
5:56
Imagine H2O is a great way to help them
5:58
accelerate um their water companies.
6:00
>> Yeah. and you just gave um us some great
6:04
advice and I I mean I I remember I can
6:07
remember it like it was yesterday
6:08
sitting in front of you the advice you
6:09
gave and so um you were certainly paying
6:12
it forward not only on your time but
6:14
also um as the founder and to your point
6:16
managing partner of um Echo River
6:19
Capital which is solving a need in in
6:22
the water sector as well in terms of
6:24
capital and impact investing and tell us
6:26
a little bit about that fund. Sure. You
6:28
know, Echo River Capital invests in um
6:31
inception to seedstage water tech
6:34
companies around the world. The thesis
6:36
is to invest in companies that digitize,
6:40
decentralize, and decarbonize the water
6:42
cycle. We've made 27 investments so far.
6:46
Um many of them are doing very well. Um
6:49
things in in desalination like um active
6:52
membranes or in decarbonization like
6:54
Crewarbon are some of our companies. Um
6:57
and we're um uh we're excited to
7:00
continue on in uh raising another fund
7:02
and uh continue to support uh
7:05
entrepreneurs in their water work.
7:07
>> Yeah.
7:08
>> Yeah. And so tell us a bit um you know
7:11
because I I you we've talked Peter
7:13
offline a little bit about the uh
7:15
importance that data centers right the
7:17
economy that that's going to drive a lot
7:18
of the in innovation that that you're
7:20
you're speaking about. So how did this
7:23
report come about? Um mind sharing a
7:26
little bit the origin?
7:27
>> Sure. Um so uh last year we were
7:31
thinking about where should we continue
7:32
to lean into and invest and so we did an
7:35
analysis of where the new water demand
7:39
will come from over the next 20 years.
7:41
And by far the largest new increment of
7:44
water demand will come from cooling of
7:47
data needing for the data centers. So
7:50
they're huge energy consumers, data
7:52
centers and hyperscalers uh to cool
7:55
their chips. Um water is a really big
7:58
part of that. And it was clear in the
8:00
data analysis that we did that um if we
8:02
were going to be part of an impact story
8:04
to help the sustainability of water and
8:08
in data centers, we would need to invest
8:10
in the next generation of technologies
8:13
that help data centers cool their chips
8:17
um and their room. So that's was the the
8:20
idea for um this report. And so we
8:23
assembled um a sort of it's call a tiger
8:25
team of of four associates, Claraara
8:28
being our our our lead. Um and we worked
8:33
pretty hard for about four months to
8:35
scour universities, the um uh innovators
8:40
uh and elsewhere. What are the next
8:43
emerging technologies? Not those that
8:45
are being implemented today like
8:47
computer room air conditioning and
8:49
evaporative cooling but the next
8:51
generation. What are the technologies
8:53
that will be used in five years or
8:55
beyond um and to invest at these
8:58
inception stage companies that are being
9:01
spun out of university labs for example.
9:05
So that was the uh the the notion and
9:07
that's why and we wanted to share it
9:08
with the world. We wanted to um you know
9:10
let people know where things are
9:12
beheaded and also to um invite people to
9:16
um suggest companies or ideas that that
9:18
Echo River should be investing in or
9:20
learning about.
9:20
>> Absolutely. Absolutely. Um okay. So um
9:24
let's um let's dumb it down for the uh
9:28
my my 16-year-old who's hopefully
9:30
listening to this episode. um talk about
9:33
the most important technology components
9:35
of a data center and then maybe we can
9:37
talk a little bit about what these
9:38
emerging technologies are but can
9:40
someone break it down at the most basic
9:42
level like what are the technology needs
9:46
um at the data center center level that
9:48
we should be focused on?
9:50
>> Yeah, Claraara go ahead.
9:52
>> Yeah, definitely. So to try to dumb it
9:55
down a little bit and these are pretty
9:56
complicated buildings. When creating
10:00
training these AI models and then using
10:03
them to answer questions like on chat
10:05
GBT you need to use um compute or you
10:09
need to run operations on computer chips
10:12
and servers. Doing this produces a lot
10:14
of excess heat and this heat has to be
10:17
removed in order for the chips to
10:19
continue working efficiently. So it
10:22
starts at the chip level. You have to
10:24
take the energy, the heat out of the
10:26
chip and move it somewhere else so they
10:28
can continue operating. Then at the
10:30
building level, you need to expel this
10:32
heat uh usually outside. And then um
10:36
even at the region level, you need to
10:38
choose where to place these data centers
10:40
to reduce the overall water use. So
10:43
there's innovation um at every scale. We
10:46
talked to researchers working at the
10:48
chip scale and those working even like
10:50
in regions. Um so the innovation is
10:54
there's a range of new research and
10:57
commercialization and I'd be happy to
10:59
talk about any of those scales that
11:01
you're interested in.
11:02
>> I know no ex exciting and safe to say
11:04
that these are 247 continuous operation
11:08
environments that are always on and and
11:11
have to um operate at that at that kind
11:13
of aggressive capacity as well. Uh-huh.
11:17
>> Yeah, definitely. Um, the these chips
11:21
are very high power and you don't want
11:23
to lose any data. They're very high
11:25
security buildings. Um, so the operation
11:27
is really important to the customer.
11:30
>> Yeah, I should also say, you know, some,
11:32
you know, what where is the water being
11:34
used today in these systems, right? A
11:35
lot of the the concentrated water demand
11:38
is coming from evaporative cooling,
11:40
right? We all know that's how buildings
11:42
get cooled today. they they they bring
11:44
in water, they evaporate the the notion
11:47
of evaporation consumes the water as it
11:50
moves into the atmosphere and that the
11:52
use the concentrated use of evaporative
11:54
cooling really is the highest demand. So
11:57
what we're looking at is are there
11:59
alternatives to evaporative cooling,
12:02
alternative sources to use for
12:03
evaporative cooling or improve the
12:06
efficiency of chips. I mean there are a
12:08
whole range of different topics and
12:10
innovations that we cover in this
12:11
report.
12:12
>> Yeah, it is. Yeah. And um and again
12:14
we'll link to and we'll try to maybe do
12:16
some images and graphics too because the
12:18
um the life cycle for lack of better
12:20
word that you you described Claraara is
12:22
it safe to say I mean there's multiple
12:23
innovation categories within each of
12:26
those life cycles of a data center the
12:28
the the chip the the the heat then the
12:31
what to do after I mean um and it's
12:33
nicely broken down into all of those
12:35
different life cycle phases. Are you
12:37
able to speak um speak to that a little
12:39
bit those different phases that you
12:41
you've identified?
12:43
Yeah, definitely. Uh, another big part
12:45
of this that could be surprising is a
12:48
lot of the water use actually comes from
12:49
energy production that these chips use.
12:52
Actually, the majority probably comes
12:54
from these non-re renewable sources of
12:56
energy requiring water to produce uh the
13:00
electricity. Um, so yes, you can make
13:04
these trips and buildings more water
13:06
efficient, but energy efficiency could
13:08
also aid in your goals of reducing that
13:10
scope to water use as opposed to one.
13:15
>> H that's fascinating. Yes.
13:17
>> Yeah, it's a it's it is fascinating. We
13:20
um we've had George Hawkins on the
13:23
podcast and um it's one of the things
13:25
that I always tie back and forth between
13:29
um water utilities and now data centers
13:32
is um the public is becoming more aware
13:34
of how much power it takes to actually
13:36
move water.
13:37
>> It's it take water is heavy and it just
13:40
takes a lot of energy and so water
13:42
utilities tend to be the highest that
13:45
they've got the the largest um uh
13:48
electricity bill. um because of how much
13:50
power it takes to move water. Um and so
13:53
we're we're kind of seeing that play out
13:55
um as data centers
13:57
>> come to the front of the the forefront
13:59
of the mind of the public is to say, "Oh
14:01
wow, it takes it not just produces a lot
14:04
of power. It also takes a lot of water
14:06
and these things um are intricately
14:08
connected." Um, I'm curious too like how
14:12
you think about uh th those two things
14:15
together from a technology perspective
14:18
um and a bias toward water versus power
14:21
or um how they relate to each other.
14:24
>> Uh yeah, I can speak to that and then
14:26
Peter, if you have anything you want to
14:28
share, go ahead. But um there definitely
14:30
is that trade-off between energy and
14:32
water and we found that right now these
14:35
technology companies are prioritizing
14:37
the energy. it seems to be a limiting
14:39
factor in a lot of places. So as they
14:43
could deprioritize water and a lot of
14:46
places with energy availability have
14:48
very limited water availability
14:50
especially when you're building data
14:51
centers in places like Phoenix, Arizona
14:54
or West Virginia. Um so that problem is
14:57
something people are running into more
14:59
and more. Um Peter, do you have anything
15:01
to add?
15:03
Yeah, I think that water is sort of an
15:06
afterthought today just in the way that
15:08
permitting is done today. That power is
15:11
much more um of a limiting factor or a
15:14
constraint.
15:15
>> That's why some companies are sort of
15:16
bringing their own power, if you will,
15:18
to the data centers. But when it comes
15:20
to water, if a city council needs to
15:22
permit something, they just say uh
15:25
basically yes. Um and then they tell
15:27
their water department, hey, go find go
15:29
hook them up and go find the water. and
15:32
it is creating more of a problem. Um
15:34
these companies right are the big a lot
15:36
of name brands but some that you that we
15:38
don't know that build them and then
15:40
lease them out to uh the brand name data
15:43
center companies and they do need a
15:45
social license to operate. Um y and so
15:48
there are you know community concerns
15:50
about um you know diverting water from
15:53
natural systems or from human uses uh
15:56
that you know big companies need to be
15:59
concerned about and think about the
16:01
operational water risk um that goes
16:05
along with supplying um or diverting
16:08
water. Uh so you know companies like um
16:12
um water plan work with large brands to
16:15
help them manage operational water risk.
16:18
Um on the flip side you know there are
16:20
NOS's out there that are sort of
16:22
watching what's happening. Um uh so data
16:26
center watch right is a community group
16:29
that shares information and they're out
16:31
to stop or delay data center
16:34
development. So it's in the interests of
16:36
everyone to um make data data centers
16:39
more water efficient um so that they can
16:42
continue operating and provide the
16:44
services that we are all depending on
16:46
like you know the chat GPTs and the
16:48
clouds and perplexities of the world.
16:50
Mhm.
16:50
>> Yeah. Our use is certainly of those not
16:52
they're not going away, right? So, uh
16:54
the technology needs to evolve. So,
16:56
let's let's talk about then some of the
16:58
what are some of the um technologies
17:02
uh that that are most interesting to you
17:05
all along that front especially on the
17:07
efficiency uh front. What what
17:09
technologies or technology segments um
17:11
emerged on this that are most kind of
17:13
exciting for you all?
17:15
>> Yeah, there are definitely a few
17:16
examples starting at the tip level.
17:19
people are developing new materials that
17:21
can allow better heat transfer, more
17:23
efficient heat transfer. Um, at the
17:26
building level, you want the power of
17:28
evaporative cooling without losing as
17:31
much water as you lose today.
17:32
>> So, people are developing membranes like
17:35
holo fiber membranes that are actually
17:37
used in medicine, but are really good at
17:39
holding on to water so you lose less.
17:42
And then last, at a larger scale, you
17:44
can, like I've mentioned, once you have
17:47
more data about water use and um water
17:50
availability in certain regions, you can
17:52
focus on citing your data centers in a
17:54
sustainable way.
17:55
>> Yeah. Yeah. And and what on the um
17:58
anything on the water capture side uh
18:00
that that would go into the the reuse, I
18:03
think um or how does the the membrane
18:05
with the the medical grade membrane
18:07
work? I'm sorry, I'm getting kind of
18:08
nerdy on this. Yeah, pull me back up,
18:10
Isaac. But I am curious on that piece
18:13
again. Let's do it. Yeah.
18:17
>> Yeah. But is there is there any kind of
18:19
circular economy thing going on here? If
18:22
we can capture more, then we can put it
18:23
back to to use and and kind of in
18:26
containerized way.
18:28
>> Yeah, for sure. Let me let me mention a
18:29
couple things. So, one of them is around
18:31
um atmospheric water harvesting. uh
18:35
where um companies like Uruvu Labs which
18:38
is in our portfolio captures atmospheric
18:41
water um and then they you it's powered
18:44
by the waste heat from a a data center
18:47
for example which comes out at just the
18:49
right temperature to power it and then
18:50
once the water is captured then it is
18:53
sent back to the data center for
18:54
evaporative cooling purposes.
18:56
>> Yeah, I think that's really cool.
18:57
>> So that's sort of the cir the the
18:59
circular economy in in a data center for
19:01
for heat and and water capture. But then
19:04
in the traditional way there you know a
19:06
lot of data centers are using um
19:09
recycled water from the m municipality
19:12
um to use for their once through
19:15
evaporative cooling systems. So you know
19:18
there's a way to capture that and treat
19:20
their municipal waste water and send it
19:23
uh to the data centers for example. So
19:25
that's another um circular usage.
19:28
>> Yeah. And I think something to point out
19:30
um I was just at a conference in cal not
19:33
California a few months ago as I think
19:37
for us as you move further east this
19:39
notion of recycled water isn't as mature
19:41
where it is on the on the west but I
19:43
think you know Peter and I'm sure
19:44
Claraara you can speak to you know
19:46
recycled water is is being and has been
19:49
used in California for multiple
19:51
applications for both potable and
19:53
non-pottable use for a while now. Um,
19:56
and so this notion, if you're hearing
19:58
this in Indiana and you're like, "What?
19:59
You can, you know, use recycled water
20:01
for municipalities, like no, this is
20:03
actually a thing,
20:05
>> right?"
20:06
>> Oh, absolutely. I mean, California is
20:08
making huge investments in in recycled
20:10
water today. Um, some of the largest
20:12
urban areas, LA, San Diego, Silicon
20:15
Valley are all putting in um, you know,
20:18
recycled water uh, systems that can
20:20
treat to uh, indirect or direct potable
20:23
reuse. people can can drink this stuff.
20:26
It's in fact it's even at a higher
20:27
quality than many of the source waters
20:29
are. It's polished to that level. Uh and
20:33
so it's certainly usable um for inputs
20:36
into evaporative cooling for sure and
20:39
even for um high pure um high purity
20:42
water usage in industrial or fab you
20:45
know uh semiconductor fabrication things
20:48
like that.
20:48
>> Yeah. Yeah. I just wanted to double
20:50
click because I think some folks that
20:51
may be a little further removed from the
20:53
water industry in terms of think this is
20:55
a futuristic thing and it's not at all.
20:57
I mean this is actually something in
20:58
application in practice in statute I
21:00
mean in um you know regulation even and
21:02
and acceptance and so it's just um I
21:05
want to make sure that people don't
21:06
think that's such a futurist concept.
21:08
>> Yeah. And sometimes we talk about that
21:09
idea and it's like I was it another
21:11
podcast guest that we had but we were
21:13
just talking about like all water is
21:15
recycled. It just is to what extent are
21:17
you compressing that timeline or what
21:19
process are you using um to do that?
21:22
There's there's just a natural water
21:24
cycle that happens regardless um we have
21:26
the water that we have.
21:28
>> Yeah, I think you're bringing up a good
21:29
point too, but there is a role for
21:31
regulators here, right? I mean,
21:32
California um last year allowed for the
21:36
first time direct reuse, which is the
21:39
human consumption of recycled water. Um
21:42
and so this is a new a really new area
21:44
of important investment as well. Um and
21:48
I think there may be a role for
21:50
regulators in um tracking, monitoring,
21:54
reporting on water use in data centers.
21:56
Many data center developers actually
21:58
want some sort of regulator input here
22:02
to sort of set a level playing field so
22:05
that they don't have to choose um on a
22:07
siteby-sight basis or statebystate
22:10
basis. Maybe there should be some
22:13
requirement um around that uh to
22:16
incentivize data centers to do the right
22:18
thing for water sustainability.
22:20
>> Yeah,
22:20
>> that's a great topic and we um we were
22:25
just talking and that I didn't prep you
22:27
for this question at all. So, I'm um
22:29
going to put you in the hot seat, but um
22:31
but we we had another guest when we were
22:33
talking about data centers that brought
22:35
up the concern um more that like if
22:38
water utilities are required to put in u
22:43
more equipment that um helps produce a
22:46
lot more water, but then the data center
22:48
gets more efficient. How do you it ends
22:51
could end up being the rateayer in 15 20
22:54
years that's bearing the operational
22:56
maintenance cost of maintaining
22:59
>> the footprint
23:00
>> the the footprint of um of the upgrades
23:03
that were needed to be done to the uh to
23:05
the water utility. So all of a sudden
23:08
the the population doesn't support the
23:11
the use um uh the water utility. So, um
23:17
I'm just curious like kind of your
23:19
reaction as we think about what
23:20
innovations are needed or how data
23:23
centers themselves should think about um
23:25
interacting with the with the water
23:27
utility. Um other options for them.
23:30
>> Yeah, I I'll start and then I'll pass it
23:32
to Claraara about to talk about water
23:34
use effectiveness. But the the first
23:35
point is I think is that it's a
23:37
case-byase basis like water use is very
23:40
local. It really depends on what the
23:43
situation is in that particular case.
23:44
Like are you in a water scarce area? Um
23:47
what's the, you know, the capital cost
23:50
of the water system that they're
23:51
operating in? In some cases, a new wa a
23:55
new data center that is a new water
23:57
customer can actually reduce the cost,
23:59
the variable cost of water use for
24:02
everybody because they're a new customer
24:04
and they're sharing the cost across a
24:07
bigger customer base. So that could
24:09
actually reduce that that cost. But I'm
24:12
going to pass it over to Claire to talk
24:13
about water use effectiveness.
24:16
>> Definitely. Right now the main metric
24:18
for measuring water use would be water
24:21
use effectiveness or like water use per
24:24
unit of compute.
24:25
>> Okay.
24:26
>> And talking about like things are
24:28
changing so fast. These data centers are
24:30
getting more efficient. That number may
24:32
change as the compute goes up. But
24:35
overall, my impression is that as long
24:37
as they can keep using more energy and
24:39
more water, they will until they're
24:41
stopped. Um, and these chips are getting
24:43
so powerful that they're demanding more
24:46
energy and more water um as they do
24:48
higher comp.
24:50
>> Oh, that's interesting. So, as the So,
24:53
there is So, as they get more powerful,
24:55
they're they're actually consuming more
24:57
energy. So, so we should only expect the
25:00
the need for it to go up. There is no
25:03
really notion that it would actually go
25:05
down just by the nature of how the
25:08
>> kind of yeah that where that makes sense
25:10
to me is I remember being a kid and
25:13
getting like a computer that had like
25:15
two two or eight megabytes of RAM or
25:20
like and being like who would ever need
25:22
a gigabyte?
25:23
>> Yeah.
25:24
>> And now I have like 512 on my phone and
25:26
I still fill it up. Um, so yeah, you can
25:29
always
25:29
>> Yeah, we will always fill the space. Um,
25:33
>> yeah, I'm going to date myself and say
25:35
my first computer, my Apple 2, I think,
25:36
had 32K.
25:39
>> Okay, so we're just want to
25:41
>> Yeah, I was trying to remember. I'm
25:42
like, yeah,
25:44
>> but um but I think there the water use
25:46
effectiveness is one part about it,
25:48
right? Is if they can lower lower that
25:50
then you can put more compute in with
25:52
the same energy demand. But if the
25:55
energy demand of the shell of a data
25:57
center goes up, you're going to need
25:59
more cooling. Also, it's really about
26:02
the total energy consumption for the
26:06
building. And that will set that will
26:08
determine, you know, the the cooling
26:10
demand
26:11
>> of which then can be set and determined
26:14
or engineered in a number of different
26:16
ways. And that's really what our data
26:18
center innovation report is all about.
26:21
What are the new ways to deliver cooling
26:24
in water efficient ways?
26:26
>> Yeah.
26:27
>> Yeah. What were um what were some of the
26:29
surprises of the report? Any any
26:32
findings that uh that surprised you?
26:35
>> Definitely those trade-offs between
26:37
energy, water, and also flexibility in a
26:41
cooling system, how well it can adapt to
26:43
a building, an existing retrofit,
26:46
something like that. I was also
26:48
surprised just like speaking to
26:50
researchers here at Stanford um to see
26:53
how much more people are thinking about
26:55
this as of late. Uh Stanford actually
26:58
released a new initiative earlier last
27:01
year um called the Bits and Watts
27:03
initiative. One of those projects
27:05
includes professor Megan Mauder who is
27:08
working on optimizing data center
27:10
sighting. And then there are other
27:12
professors here working on carbon
27:14
capture and realizing that data centers
27:17
and uh atmospheric water harvesting can
27:20
all come together um in this like next
27:22
generation research.
27:25
>> I Yeah, I had the pleasure of meeting
27:27
Megan not too long ago, another great um
27:30
water uh water entrepreneur. What um can
27:33
you double click a little bit on the
27:35
building optimization and the um uh
27:39
building or region optimization that you
27:41
speak of because I think it'd be helpful
27:42
to it's it's a topical thing actually
27:45
here in Indiana. So I I'd just be
27:46
curious can you explain a little bit
27:48
more to our audience what that means and
27:49
the importance of that.
27:51
>> Yes. So her specialty or the slab
27:54
specialty um that I'm a part of. So I
27:57
maybe that's why I keep mentioning it.
27:59
Um but they are working with people here
28:01
in the Bay, a lot of these technology
28:03
companies to understand how they would
28:06
like to site their data centers, receive
28:08
data on their operation and um their
28:11
placement and then use machine learning
28:14
AI itself actually to uh optimize for
28:19
water use and when designing these data
28:21
centers and to use you know optimization
28:26
and non-convex problems to I I don't
28:28
want to get too much into it. I don't
28:30
know everything about it. But
28:32
>> in some of our research, we you know, we
28:33
went to some of the biggest data center
28:35
operators and developers out there. And
28:36
what I what was surprising was the
28:39
variety of of emphasis they were putting
28:42
on water. Um in some cases, a company um
28:46
would have over a hundred have hired
28:48
over a hundred people just to focus on
28:49
solving water for their data center
28:51
operations and development. But others
28:53
of an equal size have two. And you know
28:57
it was really interesting to see that
28:59
just the difference emphasis that they
29:00
might place on some are some are trying
29:03
to solve water use effectiveness like
29:05
Microsoft has a new campus in Wisconsin
29:07
that's almost you know zero water use
29:09
it's closed loop um and others are more
29:11
focused on um water replenishment where
29:14
they're not reducing the water demand of
29:15
a data center. They're just trying to
29:17
find water in the natural environment to
29:18
replace
29:20
that water use. So the variety of ways
29:24
they're addressing sustainability, you
29:26
know, are is quite broad.
29:29
>> So Peter, how um where's the where do
29:32
you think the standardization will come
29:34
from or do you think there will be
29:35
standardization for these things?
29:38
>> You know, there was an effort um AB93
29:40
was an attempt to standardize reporting
29:42
and monitoring of water use in data
29:44
centers in California and it was vetoed
29:46
by um Governor Nome. um unfortunately I
29:49
think but that is the type of um
29:51
standardization we should at least have
29:53
as a minimum bar is around you know
29:56
monitoring and reporting of of water
29:58
use. I think that's that's pretty that
30:00
should be done universally. And I just
30:02
think every every economic development
30:05
um uh I'm going to send it to mine here
30:07
in the Indiana uh association ecosystem
30:10
needs to read this report because I
30:11
think it does show some of the
30:13
variability of how to your point out how
30:14
people are looking at this two versus
30:16
100 and the you know and I just it's
30:19
there there needs to be some layer of
30:21
kind of standardization or where are the
30:23
best practices that we are you know
30:24
pointing to because it does seem like it
30:26
could get a little bit wild west out
30:28
there uh pretty quickly if it's not
30:30
already there in and circumstances.
30:33
>> Yeah, I mean data centers represent a
30:35
huge increase in the you know the GDP of
30:38
our country. Like I think it was
30:39
reported that 92% of the incremental
30:42
growth in the economy last year was due
30:44
to data centers and AI. Um so if we're
30:47
going to be spending that kind of money,
30:48
we should at least be knowing what the
30:50
water use effectiveness and the water
30:52
intensity and total water use are for
30:54
the these systems. Um you can't um
30:58
manage what you don't measure. Yeah,
31:00
that's one of my favorite lines. I love
31:01
that. And you can and that's and that's
31:03
the thing. There are technologies out
31:04
there to do just that. So why not why
31:06
not adopt them and build upon them,
31:08
right?
31:09
>> Yeah.
31:10
>> Um so how help us frame up then Peter?
31:15
Uh you talked about the GDP the what
31:17
where do you see the growth of this
31:18
space? I mean crystal ball where do you
31:21
see this how big do you see this space
31:22
getting?
31:23
>> Yeah, we're in early innings in AI,
31:25
right? It's it's transforming our
31:27
economy. every company's becoming an AI
31:29
company or is using AI. Um and uh so you
31:33
know we're still uh going to be building
31:35
this out for the next 20 years at least.
31:38
And so I think it's important that um
31:40
investors like Echo River are supporting
31:43
the the new entrepreneurs and innovators
31:45
and that there are data center operators
31:47
who want to adopt and deploy uh new
31:50
innovations for for water s you know
31:53
water efficient um cooling technologies.
31:56
That's the point of this report. We're
31:58
hoping others will um will take a look
32:00
at this and invest in it um and reach
32:03
out to us, you know, for ideas where uh
32:06
of the the next big thing.
32:08
>> Absolutely. Are you guys um Peter just
32:11
curious, are you running into much
32:12
competition? What's the Are there others
32:15
as interested in this space or do you
32:17
feel like it's um a little bit of green
32:19
space right now?
32:21
Oh, there's lot lots of um you know uh
32:24
venture firms out there interested in in
32:26
the climate tech and AI. I think we see
32:29
a lot of collaboration and partnership
32:31
at every stage of company growth. So,
32:35
we're excited to, you know, to work with
32:37
and co-invest in um in the space.
32:39
>> Awesome. I'm curious too kind of
32:41
following up on that like generally I
32:44
get the sense that there's um uh in
32:50
in venture funding in general that's
32:52
it's kind of cooled down you know back
32:55
from when I started in 2009 there was
32:57
kind of like a like everything was
32:59
venture and now there's a lot more um
33:02
corporate uh corporations partnering
33:05
with universities and and those kind of
33:07
things. I'm curious if you have any
33:09
sense on how the models for funding
33:11
might might shift or be collaborative um
33:14
uh with with corporates.
33:17
>> Yeah, it's a it's a great um uh idea. I
33:20
think I see a lot more collaboration
33:22
happening. Um so uh for example um at
33:25
Yale University, they've established a
33:27
new partnership um with their regional
33:29
water authority there. So there's a new
33:32
incubator there called Climate Haven's
33:34
Water Hub, which is exciting. Um LA uh
33:38
clean techch incubator partners with the
33:39
Los Angeles Department of Water and
33:41
Power. Uh they have a whole facility.
33:44
They've partnered together and brought
33:45
in corporate investors and um corporate
33:48
venture capital is growing as well. Um
33:51
so that's exciting. So um you know
33:54
companies like Zyllem have have new
33:56
venture investment arms. So there's I
33:59
think there's a lot of um opportunity
34:01
for you know cross- sector collaboration
34:04
between NOS's like um Imagine H2O with
34:09
corporate venture capitalists with
34:11
utilities and they're all focusing on
34:13
solving some of these challenges
34:14
together. So I think it's really
34:16
exciting at the forefront of innovation.
34:18
>> It is.
34:19
>> I do too. And I'll I'll also kind of say
34:22
like where I see a lot of that going is
34:24
that there's a lot more focus on
34:27
meaningful impactful innovations versus
34:29
the era of who's got the next Candy
34:32
Crush that's going to go from zero to $2
34:35
billion valuation in 3 months. what's
34:37
going to make it
34:38
>> and um and I I think we're at a more
34:40
sustainable thing where we're being more
34:42
um
34:43
uh a bit more thoughtful about what
34:46
needs to exist um and what's truly
34:49
investable.
34:51
>> Yeah. Clara, what were you going to say?
34:54
>> Sorry. Yeah, just one more example of
34:56
collaboration. Amazon in partner with
34:59
the water environment federation and the
35:02
water center at UPEN as well. Leading
35:04
utility of the world just launched a
35:06
water AI Nexus center for excellence. Uh
35:09
it was announced at New York Climate
35:11
Week. So just another example of
35:12
collaboration.
35:14
>> Collaboration. Okay. Um going to end on
35:17
a media question, but I think it's an
35:19
important one because uh I think to the
35:21
general public data centers is a bit of
35:23
a controversial topic. When I talk to my
35:25
friends who maybe aren't in in water or
35:28
as knowledgeable, they know about the
35:30
data center uh they know about data
35:32
centers coming into their community, but
35:33
it's not from a point of positivity
35:36
necessarily or innovation. It's they're
35:38
going to come in and steal my water.
35:40
They're going to come in and ruin my
35:42
environment and and and take my energy
35:44
and so on and so forth. Um, but I think
35:46
those of us working in the space really
35:48
see a lot of the positive outcomes and
35:50
the innovation all of you know which
35:52
your report kind of um, you know, ties
35:55
together. So I'm curious how do we what
35:57
what's the conversation we should be
35:59
having to help people educate people on
36:02
the positive outcomes and the innovation
36:04
that's really coming from um, data
36:07
centers in our community. Do you guys
36:08
have any thoughts on on that?
36:10
>> Yeah, that that definitely makes sense.
36:12
I've read articles from the New York
36:14
Times about data center companies or
36:17
just big technology companies coming
36:19
into places building buildings without
36:21
really letting the local community know
36:23
about the water impacts cuz they
36:25
themselves weren't quite aware. Yeah,
36:27
>> but now I think as people measure these
36:29
things more um and we arrive at
36:32
consistent metrics, then it's easier to
36:35
communicate with local communities and
36:37
help people understand and hopefully be
36:40
transparent about um the impacts that
36:42
they could have and how they're
36:43
mitigating them.
36:44
>> Yeah. And I don't think that um you know
36:47
data centers are worth demonizing at
36:49
this point. They're providing you know
36:50
uh necessary services for our economy
36:53
and this these are going to be built.
36:55
They just need to be built in um you
36:56
know in a in a sustainable and
36:59
responsible way. Um where you know water
37:03
can't stay in externality. Um as AI
37:06
scales we really have to think um think
37:09
through what the effects are going to be
37:11
and to maintain their social license to
37:14
operate in a way where communities can
37:16
get behind them. But like I said, not
37:18
every place, every local is the right
37:22
place for a data center or a right place
37:25
for a data center particular type of
37:27
water use technology.
37:29
>> So it has to be the right technology in
37:32
the right place with the right effect
37:33
and impact in a transparent in a
37:36
transparent way. So you know um we just
37:40
have to do it do it in do it right.
37:42
>> Do it right. Yeah. Absolutely.
37:44
>> I I love that. And it's it's so rooted
37:47
in I think one of one of our favorite
37:49
things about water in general is just
37:52
how place-based it is. You can't get
37:54
away from being in the water industry
37:56
and having to understand the history and
37:58
the place um
38:00
>> um and how you're um yeah how water
38:03
finds you um to make it um kind of a a
38:06
corny call back but
38:08
>> um but yeah it's it's important.
38:11
>> It's a theme. Okay. what is um we are
38:14
going to link out to the report because
38:17
um there's so much goodness in that. We
38:19
just wanted to kind of tease tease it
38:21
out to the audience and and hit on some
38:23
nuggets, but what's something we didn't
38:25
cover perhaps that you want to leave
38:27
listeners with um that that you really
38:30
took away from the report?
38:33
>> Yeah, we would love any insights we
38:35
could get into data center operations,
38:38
innovation, research. So to any of the
38:41
audience members, if you know something
38:43
that could be of use to us or that we
38:45
could disseminate to other people, um I
38:48
think we gave you a link to a form that
38:50
they can fill out, you could share that
38:52
as well.
38:52
>> We will share the form.
38:54
>> Yeah. And I just want to say, you know,
38:56
water use does matter and especially in
38:59
certain places where water scarcity is
39:01
acute
39:02
>> um and where water where new water
39:04
demand may be competing with natural
39:06
systems and with human use. So the
39:09
companies need to maintain their social
39:10
license to operate. So it's really
39:12
incumbent upon them to to be the
39:14
innovators to be the adopters of new
39:16
technologies and and to invest in these
39:18
things and um we need to have partners
39:20
with them. So I think it's a really
39:21
exciting moment. There's there's great
39:23
innovation. There's great um uh
39:26
opportunities for all of us to do great
39:28
things together and to um make sure that
39:30
people and industry are using AI um in
39:33
the best way. Uh and we're really
39:35
excited about the future.
39:36
>> Yeah.
39:37
>> Yeah. Well, thank you.
39:38
>> That's awesome.
39:39
>> We are too. And um we are always excited
39:42
about the conversations that we get to
39:43
have here on the Make Waterwork podcast.
39:47
And thank you so much to Claraara and
39:49
Peter for joining us. If you've enjoyed
39:52
this podcast, please like, share,
39:55
follow, reach out to us, let us know who
39:57
else you want to hear from who makes
39:59
water work, and we'll see you next time.
40:01
Thanks for listening to the Make Water
40:03
Work podcast. If you enjoyed today's
40:05
episode, please take a moment to like,
40:08
share, and subscribe so you never miss a
40:10
conversation. We'd love to hear your
40:13
thoughts and questions. Connect with us
40:15
on LinkedIn or send us a message.
40:17
Together, we can make water work for
40:19
everyone.