Subscribe
Share
Share
Embed
Minnesota Law Now is a new podcast from the University of Minnesota Law School that brings together some of today’s most insightful legal scholars in conversation about law, policy, current events, and the ideas that shape our world.
Hello, and welcome back
to Minnesota Law Now,
a production of the University
of Minnesota Law School.
I'm your host, William McGeveran.
I'm the dean and William S. Pattee
Professor of Law.
The massive expansion
of AI has triggered a surge
in new data center projects,
forcing municipalities in Minnesota
and nationwide to navigate choices
regarding the placement
and regulation of these facilities.
Local determinations involve
a complex web of electricity needs,
water management, economic
growth, and regional policy.
Our conversation today focuses
on a critical new initiative
led by our guests.
James Coleman is a law professor here
at Minnesota Law and a prominent
scholar of energy law.
Elise Harrington
is an assistant professor
from the University of Minnesota
Humphrey School of Public Affairs,
whose work explores how communities
and regulators respond
to new energy systems
and environmental challenges.
James and Elise are
spearheading a joint effort
alongside University
of Minnesota Extension,
the Freshwater Society,
and the Minnesota cities
of Byron, Lewiston, and Stewartville.
We're here to learn
more about this project's mission
to create a practical framework
that empowers
local officials to tackle
the difficult trade-offs inherent
in the data center boom.
James and Elise, it's great
to have you both here.
Thanks for having us.
It's so great to talk with you.
Elise, let's start with the question
that prompted this project
in the first place.
Why is data center
development such a complicated
and multi-factor decision for a city?
I think what was brought to us
was that the water
and electricity demands
of data centers are often
at a magnitude
that many local governments,
especially small local governments,
might not be that familiar with.
They came to us asking,
"How do we make these kinds
of decisions?"
I think the questions
that the three cities
that you mentioned brought
to us are representative
of those challenges
facing municipalities
when it comes to data centers.
The magnitude of both the water
and electricity needs
is also represented
by the huge land requirements
and just the large scale
of investments that
these projects represent.
That's putting municipalities
in a situation
where they need to learn,
and they need to learn
relatively quickly,
about new systems and new scales
that equate to data
center development.
It sounds like this would
be challenging for all kinds
of communities, but that
if they're smaller or more rural,
they might be at a particular
disadvantage, I'm guessing.
The smaller and often
more rural communities
are those that are often seeing
some of our larger data
center proposals.
These hyperscale data centers
are many of the projects
that you will see
in more rural areas
because of that land.
Yes, exactly.
The staff for smaller municipalities
might not have
as many resources or time,
capacity, to be able to answer
these really complicated questions.
Yes, I can imagine
it would be overwhelming
potentially in some cases.
Absolutely.
James, you've studied
energy infrastructure for years.
These hyperscale data centers
that Elise referred to, the power AI,
they can have enormous
peak electricity demands.
What does that, in particular,
mean for a community and its grid?
It's a huge deal,
and that's probably the single
thing about these facilities
that is most different
from any other project
that they're potentially
looking to attract.
Of course, there are lots of cities
that have been looking
to attract investment, and maybe
some of that investment
would have been in a project
that would have used a lot of water.
You could think of some kind of food
processing facility or otherwise.
It might have been a facility
that would have used a lot of land.
The scale of electricity demand
for these projects
is really very unlike anything
that they would normally consider.
We're talking about projects.
When we talk about
a hyperscale data center,
there's different cutoffs
that people would use.
Sometimes people would use
a cutoff of 50 megawatts
of electricity demand.
Sometimes under a Minnesota law
or under Xcel’s tariff,
we might use a cutoff
of 100 megawatts.
When you hear those
figures—about 50 megawatts
or 100 megawatts—this
is just an approximation,
but the way I would think
about it is that a data center
with 100 megawatts of demand
is roughly like a city
of 100,000 people.
If you can imagine being
a town of 5,000 people,
and all of a sudden
somebody's talking about
opening a 300-megawatt facility,
that's like plopping down,
in terms of electricity demand,
a city of 300,000
people in your city.
It's really simply a different scale
of electricity development
compared to pretty much
any other kind of project
that they might be looking
to attract to town.
That's what makes
it a different challenge
for infrastructure,
for electricity regulators,
and for our utilities.
If I get this right,
are you saying that adding
one of these hyperscale
data centers to someplace
like Byron or Lewiston turns
their electricity demands
into Saint Paul's
electricity demands?
Yes, that's right.
St. Paul, almost 300,000 people.
You're talking about
a 300-megawatt facility
using roughly the same,
actually probably a little bit more.
We've had facilities
that people are talking
about developing here,
Minnesota, that go up
to 700 megawatts.
There's no city in Minnesota
that has 700,000 people.
Minneapolis, you get
close at various times.
It's a really tremendous amount
of demand going
into places that, again,
are towns that are much
smaller populations.
Wow.
Elise, you're the principal
investigator of the project.
You heard from these towns—Byron,
Lewiston, and Stewartville.
What did you hear
from them that helped shape
the way you thought
about the research?
I think what we've heard
from them consistently throughout
the project is a need for,
first of all, just clear,
fact-based basic
information on data centers,
and exactly the kind of information
that James was just talking about.
What is the scale?
How do we put that into a picture
that they can understand
and also communicate
with their constituents about?
Then, actually, a second feature
that I think was really driven
by these three cities
was also a question to us about
how do we think about the data center
development in a regional way?
How do we think about it not just,
right, we're working
with three places;
they wanted to work together on this,
not just working
on it siloed one by one.
What does it look like to think about
both these water and electricity
demands of data centers
across the southeast
region of the state?
Wow.
James, I want to talk a little bit
more about these energy questions.
One thing that can be confusing about
this is individual municipalities
aren't usually the ones
that are controlling electricity
rates or utility decisions
or construction
of infrastructure, right?
What kind of ability,
what kind of leverage do
they have, and how does
this complicate the fact
that these are normally
statewide policy decisions?
That's absolutely right.
Typically, it's not regulated
at the local level
unless they had a municipal utility.
Typically, they're having to work
with their utility
or their cooperative about
ensuring that these new projects
don't raise electricity rates
for other users.
That is quite naturally
one of the biggest questions
and talking points
about these new projects.
It's frankly complicated.
Although everyone agrees
we would like to ensure
that these projects
don't raise electricity rates
for other customers,
ensuring that's the case
requires regulation
and utility decision-making
that maybe some of these cities
don't participate
in that actively usually.
If they want to go ahead
with one of these facilities,
or frankly, if one
of these facilities is going ahead
in a neighboring town, they may want
to be engaging with their utility
or their cooperative to say,
"What are you doing to ensure
that this doesn't change rates
for other customers
or ideally makes other rates go
down compared to what they
otherwise would have been?"
One of the tasks
that we've taken for ourselves
in developing this toolkit is helping
these communities both understand
what the electricity problem is,
how regulators would
or should try to protect them,
and how they can participate
in those decisions,
either even just through making
comments or otherwise,
to help ensure that their utility
or their cooperative
is protecting them.
Elise, you mentioned water before,
in addition to electricity.
Water presents a different
kind of a challenge.
In this region,
groundwater and surface water
are closely connected.
I guess the water is used
to cool the data centers, right?
How does that figure
into the complexity
of these decisions for communities?
I think water is often
one of the things that cities,
for example, what James
was describing,
is an area that many cities
have to educate themselves on,
in many ways on the fly,
but when it comes to data centers.
Water is an area that they have
a much longer history
of understanding,
both the process
of how they're allocated water,
but also just the governance of water
is usually something
that municipalities
have closer ties to already.
You're right.
The data centers, the demand
for water is typically for cooling.
Now, what we've seen is some shift
in terms of industry practices
from cooling systems
that are very water-intensive.
You can think of when you drive
by a large industrial site,
and you see the cooling towers,
you see all that water evaporating.
Shifting from systems
that do use a lot of water
to some that use much less water.
The trade-off there
is that systems that use
less water are using
more electricity.
Think about your air
conditioning, perhaps,
that you would turn
on during the summer.
Many people can feel
that electricity spike
that they pay during the summer
when they're using
that air-cooled system.
Now, that means that, onsite,
a data center
might be using less water.
The tricky part
there actually gets to,
if you look at a slightly
bigger picture of water use,
it means that offsite water use
is increasing because the water
that's used to generate
that electricity can be a part
of the picture as well.
If you look at total water
use for data centers,
a large portion can be attributed
to that offsite use
for electricity generation.
Again, that sounds
like another externality, right?
One town has the development
of a data center there,
and they can embrace a design
that's more electric-intensive,
less water-intensive,
but another town
may have water effects there.
Exactly.
Got it.
James, that's energy,
and there's water.
Then, of course, there's so many
other infrastructure questions.
There's land and sewer
infrastructure,
and then just the workforce.
How do those complications figure
into these municipalities' debates?
All of those are really
important considerations.
I think the one thing that's somewhat
helpful is that more like water
and less like electricity,
again, towns are going
to have more experience
dealing with that with new
projects as they come in.
Absolutely, they have
to figure that out.
There's often very
complicated debates about
what is going to be
the economic benefit
that we receive
from this major new project.
A lot of it's going to come
in property taxes,
but will some of those property taxes
be allocated to developing
some of the infrastructure
that's needed just
because of this new project,
and how do we weigh
those economic benefits
with some of the downsides
that we've talked about?
I think each of those create
complicated questions.
On the workforce front, I think,
for a project of this economic size
and this electricity use, these tend
not to have as many permanent jobs
as some other examples
that you'd be thinking about.
The benefits often for local
communities are more in terms
of the direct actual payments
under property tax
and otherwise
that are negotiated there,
and less under "You're
going to get 100 new jobs
at these facilities."
I think a lot of times,
the centerpiece of those negotiations
tends to be about that.
Now, Elise has done some work
talking with places
that have a longer history
of data center
development in Virginia.
Ultimately, if you keep building
data center after data center,
yes, there is some
permanent job impact,
not necessarily in a single town,
but around the areas
that are building that.
That's a future question.
It depends on the future trajectory
of data center development.
In terms of the facility
itself for their acreage
and for their electricity use,
it's going to be relatively
few permanent jobs.
Can I just add one thing
is that range of issues
that you're talking about,
one of the things that,
as we've talked with municipalities
both locally here
and across the country,
that we've heard, is to really think
about a clear picture
for your community of those kinds
of trade-offs you're talking about.
What does it mean, for example,
to have a diversified economic
base, and then how does
that translate
into our land use decisions
or our water use
allocation decisions?
Do we want extra capacity
on the grid for something else,
whether it's population
growth or a different industry?
One of the recommendations
or things that we've heard
from other places is to think
pretty proactively and early on about
what that picture might look
like for different places.
It really might look
different in different places,
but that might also help guide people
through that decision-making process.
It strikes me that some
of these trade-offs are trading
off some things
that have a more immediate
short-term benefit for community
versus long-term impacts
that maybe are not
as easy to see right away.
How does that figure
into the complexity
of this calculus for towns?
I think it's a huge part [chuckles]
of the current complexity
and a part of the deep
uncertainty that places are facing.
You have regions of the country
that have had a much
longer history with data centers.
Now they have been a different kind
of category of data centers
than we're seeing a lot
of this construction now.
There's lots of questions
about how long will this last?
What do I do?
One of the things that we're seeing,
I think, increasingly
in local conversations
is what kinds of safeguards
can we put in place
that either have to do
with something like site
remediation or protection
of benefits or making sure
that if something were to go wrong,
and you can think
about this very much
on the electricity side,
but across projects,
if something goes wrong,
what's in place to make sure
that a local community,
especially a small town or city,
is not bearing
the brunt of that loss?
James, you talked before about
the ways that local governments
do and don't have control
over various elements
of these data center decisions.
What tools do local governments
actually have when it comes
to influencing both
whether there's going to be
a data center
and what it's going to look like?
Well, one way that most communities
in Minnesota are fortunate
is that, typically,
this is not going to be as-of-right
development where you can just say,
"Okay, we meet the normal criteria
for a facility in your city,
so we can go ahead
and develop in this industrial area."
In part because most of them
are asking for 500 acres,
500 megawatts, et cetera.
There's going to be a give
and take in that.
That's where that community
has the most choice.
Now, they are going to--
You mean trends like zoning
and permits and things like that?
Yes.
You can usually decide:
do you want a facility or not?
If you had a by-right situation,
that might not be true,
but that's typically
what we're talking about here.
Now, in terms of controlling
the electricity use,
they typically don't have control
unless it were a municipal utility.
Now, of course, many of these cities
might be a very important member
of their very important constituency
for their cooperative or otherwise.
They might have a lot
of input to that decision,
but I think in some ways,
they're probably
not going to want to go forward
with the project unless they trust
their utility
or their cooperative to protect them
from potential electricity
costs, et cetera.
I think one thing that we found,
and Elise can speak
to this as well, is when you talk
to the communities
that have done this in the past,
some of them developed
on the basis that,
"Oh, our utility told us it was okay.
They're in charge of regulating this.
They can figure it out."
I think there's probably less comfort
with that as we talk about facilities
that are so outsized
compared to the city.
I think a lot of times
these communities are saying,
"Okay, I want to kick the tires
on this a little bit, and say,
is it really the case
that the utility
or the cooperative is protecting me
from potential rate rises?"
I actually think that
even if you ultimately had no say,
and the utility's going
to do what it's going to do,
it still would be
very useful for them
to participate
in that process to look
at how the utility protects itself.
Basically, the way
the utility protects
itself and its other
customers is it says,
"I want to ensure
that these facilities pay
for the costs that
we are going to have to pay
to preserve them."
We maybe have to build
new power plants.
We need to do transmission
and distribution.
We want them to pay for that.
The rate has to be high enough
so that they're paying for that.
The nightmare scenario
is that we build all that for them,
and then the product is delayed
or scrapped completely
or only has half the size.
All of a sudden, I have all
this new infrastructure
that I have to find
somebody to pay for.
Maybe it's my existing customers.
That's the nightmare scenario.
Typically, the way
the utilities protect themselves
is the first thing is the rate's got
to be high enough.
The second and third
thing are that data center
has to make a promise that
it is going to purchase electricity
for the next 10
or 15 years every month.
It has to be a certain percentage
of what they told me
it was going to be.
There's a take-or-pay contract.
In other words,
either you're going to take
the electricity and pay for it,
or you're going to pay
for it even though
you didn't take it.
It's really a pay-or-pay contract.
[chuckles] Either way, you pay.
Pay or pay.
[chuckles]
You might think,
"Well, that's not enough
protection because
what if the company goes bankrupt?"
Absolutely, that is something
every utility is looking at.
Now, I think it's appropriate
for the other utility constituents
to make sure the utility
is protecting itself because,
ultimately, if the utility
gets into trouble,
their rates may go up.
The utility is trying
to protect itself.
How it does that is it asks,
"What's the creditworthiness
of this developer?
Is this some giant multinational?"
They're not going broke.
Yes, it's not a problem.
If it's a subsidiary, do we have
some parent agreement that says
they're going to pay
even if the subsidiary goes broke?
Do we have some other assurance,
whether it's a bond or otherwise?
They want all those things.
Now, it's both useful
for the cities to make sure
that their utilities
are protecting themselves
that way because, ultimately,
that's how the city is protected.
Also, the city
could be thinking about it,
going to your previous
question of how do they ensure
that they get this long-term benefit,
because the city
has also been promised
all these long-term benefits.
Yes, we're a data center.
Yes, we need some new infrastructure,
but you've got 15, 20,
30 years of new money
coming to fund your city
through property taxes,
coming to fund your schools,
coming to fund your water system.
If you see your utility saying,
"Ooh, I don't think this developer
is necessarily good for it,
and I need some more
credit assurance,"
if I'm the city, I'm thinking,
"Wait a second,
I've got a lot of promises
for money for them in coming years.
Maybe I need the same
kind of assurance."
I think that, even if all you got
was information and you can't control
your utility decision,
it's still very valuable
for those cities to participate
in those dockets,
to see what other people are saying,
see what the utility is asking for,
see where the data center
developer is pushing back,
and try and get
a handle on how certain
can I be of these promised
benefits down the road.
I get a sense that you might teach
contracts in Europe.
[laughter]
Elise, you referred before to cities
taking a proactive rather
than a reactive approach.
What does that look like
on the ground for a municipality?
Yes, that gets a little bit
to what James was very
briefly talking about,
about things like your local code.
Turns out many places
actually don't have data
centers defined in their local code.
Maybe not surprising,
but they should.
There are some steps
that local governments can take
before a data center
is knocking at their door,
before they get a proposal,
to really think about
what is it that they want
and would need
within their community.
That is something
that you certainly see
gaining traction all over the country
is places saying, "Hey, actually,
we need a little bit
more time to be able to study
and change our local code."
That's one
of the reasons that you see
this boom in moratoria
going on around data centers as well.
Most of those are temporary
with the goal of saying,
"We need to define a data center.
We need to then apply
specific regulations
within our codes
to that data center."
That can be about the site design;
it can be about the building design.
We also see some areas
where they're very
specifically trying to think
about the resource
use of these facilities.
Now, as we've already talked about,
local governments
don't have a lot of control
over exactly
what that resource use is.
What we do see
increasingly is requirements
for studies on what
those impacts might be.
Those allow local governments
to be more proactive.
This also allows more time
for local governments
to talk with their constituents
about what kind,
if any, data center
they would be willing to have.
The earlier those
conversations can happen,
likely the better for any place.
Some of those steps,
like a moratorium or studies
or proactive regulation in advance,
sound like they're
all aiming to address
this knowledge imbalance
that you were worried about.
I have to imagine that the people
building the data centers
have teams of experts
that know all about this.
As you've said,
these small towns don't.
Is that something that you recommend
to communities in terms
of taking a pause and fixing
that knowledge imbalance?
I think that whether
that's the right course
of action or just taking the steps
to change their local codes,
I think that is absolutely
the right step.
We've seen different places
in different parts
of the state and different parts
of the country taking
a different strategy.
Some places really do need
that extra time, and some feel
like they have the capacity
to be able to address
that without something
like a moratorium.
I think that that proactive nature
is figuring out what's the right path
to be able to do things
like define data centers
and figure out what regulations
are going to apply to them
in the pathway that makes
the most sense for any given place.
Yes.
I think it's important to emphasize
how very different
different communities
are from a number of perspectives
when they look at these projects.
We've seen there are many communities
that have been looking for these big
keystone investments for a long time,
over a decade, et cetera,
and they may have already
land devoted to it.
There are some places
that are even in a more
advantageous place to look
at these where maybe
they had some facility
that was previously
using electricity.
Remember, the whole challenge
with cost-of-service
regulation for electricity
is that those costs are shared
by everybody on the grid.
If one of your big power users
goes down, all of a sudden,
everybody else has to pay
for the grid that's there,
and so that means
everybody's costs can go up.
If you can slot in,
"Hey, I've got another big user,"
that could be a huge assistance.
There are those kind of places
that might be quite eager for these.
Then there are other kinds
of areas where they say,
"Well, frankly, we have
a lot of different proposals
for development in our city,
and this is just one, and we'll take
the time to see if this really makes
a lot of sense for us," or "Actually,
we have higher value
uses that are coming to us
looking for land, looking
for electricity demand," et cetera.
I think every city is going to differ
a lot on their basic
orientation toward "We want
this yesterday,
we never want this, we got to think
about this some more."
Sure.
It sounds like you're
saying some cities,
like a place where
some kind of industrial
plant has closed, a data center,
at least in the short term,
could bring consumers' electric
prices down if they're already
bearing the cost of the grid
that that facility left behind.
Absolutely.
I think it's interesting.
The connection between
these new data centers
that demand a huge
amount of electricity
and the electricity prices
that everyone else
pays is very complicated,
and it could go up or down.
Often it's calculated
to try and have no impact,
but there are reasons
that it could lower or raise
your electricity prices.
If you look at the last decade or so,
there have been rising
electricity prices
in general across the US,
but those prices have tended
to rise more slowly or even fall
in the places with more
new electricity demand,
which is odd.
That's not what we're used
to in other energy markets.
If we all started driving
much bigger cars tomorrow
or driving further,
gasoline prices would go up
because they're set
by supply and demand.
Electricity is more complicated
because so many parts
of the electricity system are set
by that cost-of-service rate-making.
When you have a new user
come on the grid,
that is more investment
that has to be made and paid
for by the utility,
but it's also another big user
that helps share those costs.
The sort of good news story is, yes,
that's another user
to share the cost,
and that's going to lower prices.
We've talked about
how there's a lot of reasons
that might not be the case.
If they don't pay enough,
if the nightmare scenario
is we make the investments
for them and they don't end up
paying for it because that company
goes belly up or otherwise,
or the data center boom goes away.
There's a lot of risks
that you have to protect
other ratepayers against.
Even if the ideal is that these
projects would have no impact,
it takes a lot of regulation
to have any confidence
that we'll achieve that ideal.
Maybe a little bit of luck,
it sounds like.
Yes.
Elise, this project is moving
toward a pragmatic outcome.
You're working on a toolkit
for city administrators
that will give them some guidance
on what you're calling a proactive
approach to data center siting.
What's in the toolkit
when I open it up?
We are finalizing it right now.
The first part is really trying
to answer a lot of the questions
that you've brought
for us here today.
How do we make sure
that local governments first
understand these impacts,
understand how these projects work,
and then go into
the nitty-gritty details
of what have we seen other places
do in terms of both
that development process
that James was talking about,
what are those really good questions
they need to be asking,
and can we give them a leg up,
a jump start on asking
those really good questions
in the development process?
Then, what are the things
that we've seen other places
change in their local codes
so that they can better manage
these projects?
Yes.
I would say the thing that makes
it very different from some
other large—I don't know
if they call them
toolkits—but basically studies.
There's a lot of names
for it. [chuckles]
If you see other studies
or white papers
or websites on data centers,
I think what's very different about
our project is it is not written
from the perspective
of you should have data centers
or that you should
not have data centers.
Our goal for this toolkit
is if you come
to the questions of data centers,
or if, after reading the information
that we give you about data centers,
you think, I never want
a data center, okay,
if you're constructing
a ban or a moratorium,
you still have
to have some threshold.
When you talk about
wanting to ban data centers,
do you mean everything
above 20 megawatts,
above 50 megawatts
or above 100 megawatts?
Our ambition is that you come
to our toolkit to say,
"This is what I mean.
This is what I really want
to ban or put a moratorium on."
On the other hand,
if you come and say,
"I need a data center yesterday,"
we want to be able to help you.
"You want your data
center; how do you ensure
that it doesn't have some downsides
you might not be thinking of,
and how do you ensure
that you guarantee
the benefits that are being
promised from it,
that you get the best deal
that you can as a city?"
It's designed to serve
the widest variety of communities
with different outlooks
on whether they want
a data center or not,
or those that are still
trying to determine
where they fit along that spectrum.
You can think about
it like a little roadmap
or a menu of options
depending on what
your particular situation is.
If you like a road
trip or a meal better.
[laughter]
Take your pick.
Right.
The ultimate goal here,
it sounds like,
is not a pro or anti-data
center position,
but it's pro-informed
decision-making.
That's a good way
for an academic institution
to contribute to public
debate, I think.
I hope so.
We're doing our best.
That's great.
The only other question
I had on my mind was—and I don't know
if you have an answer
for this—but it strikes me
that the technological
need for these high-scale
data centers came out of the blue.
I follow artificial intelligence,
and the scale of the electricity need
is ramping up way
quicker than I anticipated.
It could just easily
change on the other side.
There's a real risk that this boom
goes bust, isn't there?
Yes.
I wouldn't want
to predict the future.
There's a couple
of things to say about that.
One is we've already seen
a bit of an inflection up.
The electricity use
for almost 20 years
was basically stagnant,
and it has started to go up.
Even if it mostly went bust today,
it still would have potentially
permanently changed
the demand for electricity
in the United States,
which is quite remarkable.
I think it could be a bust,
but still a huge, huge story
for the economy
and for the electricity sector.
The second thing is,
it's interesting
because I have a bunch
of friends who work for AI
companies of various ilks.
They often say, "Oh,
this data buildout is crazy.
These models are going to get
so much more efficient.
That's what always happens
in computer science," et cetera.
Then I ask them,
"Oh, so are you telling me
that some of these projects
are going to go bust?"
They say, "Well, no, I'm sure
we'll find a use for all of it."
In other words, because they are,
at the same time,
building so many
different technologies,
even if they all become
more efficient, it's like the--
What's the Jevons paradox?
The famous--
If you build another lane on a road,
you'll just get enough
traffic to fill that road.
Yes.
It's interesting,
every month, I think,
people's projections about
what these will be used for,
there were times when people say,
"Oh, this is just
for model building,"
and the actual inference,
like when you're actually using
it on your phone or whatever,
will be done in much
smaller facilities.
Well, I think the model building,
they think, "Okay,
well, maybe actually
it is going to be
that inference task, right?
That's what all
of this will be used for,
and we'll do the model
building in space or whatever.
It's like the story about
how that's going to work out,
it really changes month by month,
which is pretty amazing.
I don't have any doubt
that it's going to be transformative
for our economy and electricity
system over the long-term,
but there could be a lot
of things that,
compared to the current pace
of development, look like bust.
As somebody who studies permitting,
I just don't see how they get--
building projects is extremely tough.
One reason that we are seeing
a fair amount of development
in the Midwest and in Texas
and in some other places
is that we have been relatively
more successful at building
the kind of transmission
that basically allows you to push
more demand on the grid.
It's interesting because the way
that Texas has worked
is they just have their own power
grid and they develop it themselves.
In the Midwestern states,
we're the one part
of the country that's been relatively
cooperative in terms
of developing interstate power lines.
If I look at
the trajectory for building
interstate infrastructure
across the US,
it keeps getting more challenging.
You see these incredible proposals,
and you think,
"I've seen a lot of projects
that attracted a lot less public
opposition get caught up
for many, many more years
in litigation
and permitting, et cetera.
I just can't believe
that more than a fraction
of this would be built,
but even a fraction
of this would be transformative.
I expect it to be
a continuing large story,
even if we have what you might call,
along several measures,
especially compared
to projections, a bust.
The only thing
that I was going to say
on the growth part that I thought
was really interesting
was I was reading a report last week,
and it was just talking about,
we think, "Oh, AI does
AI," but actually,
depending on the different kind
of task that you're asking,
it uses different amount
of electricity.
If you're using AI to help you manage
your inbox is really
different than using
these models that are doing
these iterative,
almost self-decision-making
tasks—these agentic models.
It's actually quite different.
That's one of the things
where I think
about the growth over time
is that AI is moving so fast
and developing so fast.
I think there's lots of demand
that we actually don't quite know
what it is yet because these models
are changing so much.
It's hard to find this information,
but the resource demands go along
with those changing models as well.
I think there's a lot
that we don't know
about what might contribute
to at least maybe
some stability before that.
I want to close with a question
that's like you get in job
interviews about where you want
to be five years from today.
If a Minnesota city gets
an approach from a data center
proposing to build
five years from now,
what's one question
that you hope local leaders
will be asking
that they're not well-informed
enough to ask today?
The question that I really hope
that local leaders
are asking more in five years
than they are today
is I hope they're asking
for those studies.
I hope they're asking
for the developers to pay for them,
[chuckles]
and I hope they're really taking them
seriously in a way that gives
themselves more time.
There's a lot of eagerness, I think,
around these projects right
now that is making
it feel like people
have to make these
decisions very quickly.
Actually, what we've heard from a lot
of the places we've talked
to is "Slow down,
make good deals, study the impacts."
Those are the questions that I hope
local governments are asking.
How about you, James?
I think my main question would be,
what infrastructure
are we going to need
both within our city
and also for the wider grid,
and who's going to pay for it?
How do we determine that,
in any of those risky
scenarios we worry about,
those costs don't fall on us?
Conversely, how can we try
and lock in, to the extent we can,
the benefits that we're hoping
to gain from this project?
My thanks to James Coleman
and Elise Harrington
for joining us today,
and to their partners
at the Humphrey School
of Public Affairs,
University of Minnesota Extension,
Freshwater Society,
and those Minnesota communities
they're working with in Byron,
Lewiston, and Stewartville,
all of whom are contributing
to this very important work.
Thank you for listening
to Minnesota Law Now.
I'm William McGeveran,
and if you enjoyed
today's conversation,
please be sure to subscribe
wherever you get your podcasts.