OPENPUBLICA · PUBLIC MEETING RECORD
Record of Proceedings

Raleigh City Council Work Session on Data Centers – June 16, 2026

City CouncilTuesday, June 16, 2026
BodyRaleigh, North Carolina
SessionCity Council
DateTuesday, June 16, 2026
StatusNEW · FILED
Video Record
0:00 / 1:09:31
Transcript — Verbatim
0:00

That's all for this episode of Wake TV.

0:03

Keep up with all the latest Wake County news by visiting us online at Wake.gov slash news.

0:10

And be sure to follow us on Facebook, Instagram, Twitter, and YouTube.

0:14

So you can always stay in the loop.

0:16

We'll see you next time.

6:16

Well, we'll see.

11:07

Well, we're not going to be able to do it.

14:20

All right.

14:21

We will call the work session to order.

14:25

And on our canes photo to the uh data center information and infrastructure.

14:33

We have Mark Wittenberg Information Technology.

14:37

Good morning.

14:38

Uh mayor, members of the council, and Madam City Manager.

14:41

Yeah, that's a hard act to follow.

14:44

But I am also joined by my colleagues here.

14:47

I want you to know that this presentation was a collaborative effort between the Office of Sustainability, City Attorneys, IT, Planning and Development at Raleigh Water.

14:57

First of all, just an agenda for today.

14:59

We're going to talk a little bit about data centers, like what they are, the sizes, the different types, what they're used for.

15:05

We're going to talk about the impacts on the community, both the benefits and some of the challenges that we're seeing.

15:11

We're going to talk about zoning regulations and the tools that you have in your toolbox to be able to regulate where these are built and the sites.

15:20

We're going to talk a little bit about the utility considerations, the energy, the power that these consume.

15:26

And then we're going to bring it home with what to watch.

15:29

In other words, what is staff looking at?

15:31

What are we observing?

15:32

What are the things that we are looking forward as far as uh developments in this space?

15:39

So let's start off with what's driving data center expansion.

15:43

And I'm going to start with today.

15:45

If you drove into the council meeting, took a bus, uh, searched for something, used your phone, you've used a data center.

15:54

If you watched the canes just dominate Vegas the other day, you most likely used a hyperscale data center somewhere.

16:04

The point is that every day we're using these uh cell phones, we're using uh information, we're providing services, all of these consume and use data centers.

16:15

One of the big expansions started with cloud computing.

16:18

So, what is cloud computing?

16:20

Cloud computing is basically just computers that are somebody else's.

16:24

It's where you're leasing computer use or space in a data center.

16:30

The reason that cloud computing and data centers kind of came to be is that by consolidating all of these things into one place, we can share the care the cooling, we can share the backup protection, and we can share the facility instead of us all building our own and consuming more resources.

16:48

And so cloud computing was kind of that first foray into that market of consolidation of those resources.

16:56

Digital services, I love to say, you know, there's an app for that.

17:00

And there is.

18:00

Poor, that's not what we're seeing today.

18:02

What we're seeing now in this new AI world is we're seeing very directed cyber attacks that look very, very real.

18:11

And so in the cybersecurity space, we need to make sure that we're staying up with this.

18:16

Autonomous systems, these are these agents that are assisting and doing like daily tasks and automate uh providing automation around things that may have done uh been done by humans in the past but now are being done by computers.

18:32

But I would say the primary driver for data center expansion and the reason that it's kind of all the talk is around artificial intelligence.

18:40

Over the last two years, the development in artificial intelligence has just exponentially increased with generative AI, and now we're seeing these autonomous systems coming up, and you're just going to see this continue to grow exponentially.

18:56

These systems require a lot of compute power.

19:01

So different types of data center.

19:03

I'm not gonna go through this all of this information, but I just wanted to give you a little brief overview that all data centers are not created equal.

19:12

And I'm gonna start with the edge data center.

19:14

And so these are the smallest size data centers, so think of something that's like a shipping container or smaller.

19:22

And these are typically located near the service that they're providing.

19:27

For example, you might see these outside of a cell tower.

19:30

It's because it's providing the data resources for that particular infrastructure.

19:35

Or if it's a particular business that needs this type of data center, it would be relocated right next to the business.

19:43

When we talk about size of data center in development, we talk about the size or the square footage or the land requirements.

19:51

However, data centers often talk about it in the peak power demand.

19:55

And so if you hear like the hundred kilowatt to five megawatt for a data center, what that's saying is that peak capacity, like it's going at a hundred percent, that's the power that it's using.

20:09

And you'll notice that the edge data center has a low water demand and a low power demand.

20:14

Also, for your reference, we've put like the equivalent, like think of a big box warehouse, and so we've tried to put like an equivalent power demand for this particular type of data center, just for reference.

20:26

Then you've got the enterprise corporate.

20:29

So imagine uh a data center like an ATT, which is actually right next door to us here.

20:35

This is a data center that is specific to a particular corporation or a particular purpose, like banking, uh telecommunications, or even the city of Raleigh, we have our own small data center.

20:48

Typically, these data centers, and I'm gonna go into the cooling, but again, low power consumption and then also low water usage because of the type of cooling system, and I'm gonna talk about cooling systems here in a minute.

21:01

And then there's co-locational and co-location and cloud regional, these are the larger data centers.

21:08

And a co-location, the idea is that they build the data center and then they lease out space to different corporations to be able to utilize the resources of that data center.

21:20

And the reason that the water demand is a little bit higher is because of the cooling system and the amount of resources that are in there.

21:28

The big thing that's been happening is these hyperscale data centers, that's kind of the big news, which is these large mega centers, which are large in size, require a lot of power, and then a significant demand on the water.

21:44

A little bit of background also on data centers, which is we've had data centers for many, many years, and typical data centers kind of idle most of the time.

21:56

If you think about it, if your computer, oftentimes, if it sits on your desk, you're really not utilizing your computer, you're just using it once in a while, and while you're using it, you're using resources, and that's kind of the way many of these data centers work, as opposed to these hyperscale AI data centers where they're running 100% of the time.

22:15

And so think about it like a truck going uphill dragging a huge load with the high-engine RPM.

22:24

That's what these larger data centers are running, as opposed to these enterprise co-located co-location and cloud regional centers.

22:33

I'd like to also point out that uh we currently have no hyperscale data centers in the city of Raleigh, nor do we have any uh requests for a hyperscale data center.

22:45

So some of the community benefits uh increased tax base.

22:49

These data centers are often hundreds of millions of dollars or even a billion dollars to be able to build, which means that the tax base on that particular development provides huge tax benefits for us.

23:05

And then on the municipal services, and I'll just mention that many of these have pros and cons.

23:11

And so like the municipal services, based upon the land use or the land that this particular facility is taking up, and then the resources it's using from the city, when you think about police, fire, transportation impacts, it just doesn't have the impact on municipal services that some other land uses might have.

23:32

But on the converse, you think about maybe that land use could be for something that has a different excuse me, a different benefit to the city.

23:41

Same thing with jobs is while these data centers are being built, you're talking about construction jobs, you're talking about infrastructure improvements, you're talking about investments into the community, and then resources that are being brought into the community.

23:57

However, after they're built, many of these are autonomous and so require a small job footprint.

24:03

So you have both a benefit of creating jobs during the construction, having jobs while it's running, but it is a small amount of jobs once they are running.

24:14

Another plus and minus is the utility revenue growth.

24:17

Is because they are consuming electricity, they are consuming water, there is revenue that we're getting for the water side and Duke Energy, others would be getting from the energy side.

24:30

And then research and technology.

24:35

We have top-notch universities, we have medical research, we have technology companies that are here.

24:42

All of those companies and all of those industries require the compute power that we're talking about in these data centers.

24:49

And so for that community development for those industries, they definitely require these types of data centers or data centers like them to be able to do their work.

25:02

So some of the challenges, land use, we're going to talk about land use again, which is again that ratio of land use, these large hyperscale data centers require a lot of land.

25:14

You can go vertical, for example, ATT next door is multiple levels, and so they're not using as much land.

25:21

But again, the land use ratio to job to creation potentially could be a problem.

25:27

Then there's obsolescence.

25:29

These data centers and the technology is just moving so rapidly.

25:35

And what's this going to look like in three to five years?

25:38

In technology, we often saw that our servers that run our infrastructure, we could replace every five to six years.

25:48

What we're seeing now in the AI world is these things being replaced every two to three years.

25:54

And as these are improving, they're also improving on heat dissipation, energy use, and just the compute power itself.

26:03

The other thing is NVIDIA and Dell just announced a device that allows you to do AI workloads actually, like at a workstation.

26:12

So things that required a data center, now you have the capability of potentially doing those like at we call the edge or from your device.

26:23

The biggest impact though is the environment and energy.

26:26

So we have noise and vibration.

26:29

There's been a lot of talk about the noise and vibration from these data centers.

26:34

What I would just say is as far as vibration, computers really don't like vibration.

26:40

And so the data centers are typically designed to reduce the vibration.

26:45

And so the noise and vibration that the communities are hearing and potentially concerned about are really generated from the cooling systems and from electrical production.

26:57

So it's the cooling fans.

27:00

Again, if you just go next door, I happen to like walk around the data center, and you could hear the cooling fans, but once you get out of the range of the cooling fans, you really can't hear the building, or you can't feel any vibration.

27:13

There's the energy use.

27:15

Again, if you think about the car analogy and the car going uphill and dragging these huge AI data loads, they're going to consume a lot of energy.

27:25

And when they consume that energy, they produce a lot of heat, and that heat has to get dissipated some somewhere.

27:29

And typically that heat would get dissipated into the atmosphere through a cooling system.

27:29

And then there's the water use.

27:39

So to talk about water use, I'm gonna give you a little uh HVAC and data center cooling system uh primer to talk about the different types of cooling system because what I want you to remember is that not all data centers are created equal.

27:56

I'm gonna start with there's two types of cooling system.

28:01

There's a closed loop system, and then there's an open loop system.

28:06

So think of a closed loop system like your car cooling system.

28:10

You have a radiator, you have coolant in your car, and what it does is it's circulating that through the engine to keep the engine cool.

28:17

Every once in a while, you may have to add some water or some antifreeze to the system because over time it will just leak a little bit.

28:25

It's the same thing with a cooling system and a data center.

28:29

If it's air-cooled and closed loop, it really does not use much water or much coolant.

28:36

It's just whatever is lost in the system over time.

28:40

And then there's evaporative cooling, which is an open loop system.

28:45

So imagine that.

28:48

Who grew up with like an evaporative cooler where uh there's water that drips onto the pads and then you blow air through the pad, and the evaporation of the water is what creates the cooling.

29:00

And so that's exactly what is happening with evaporative cooling.

29:04

Now, what I'm gonna mention is as you start to look at these different systems and go across from left to right, it's least expensive to most expensive.

29:16

And then look at the water usage and the power usage.

29:21

Because those are kind of your three levers or controls that these data centers have: the cost of installation, the cost of running the system, the water usage, and then what's the power usage.

29:33

Many data centers will do the evaporative cooling because it's the least expensive to install, and it's also the least expensive to run because the energy efficiency is actually very good.

29:45

However, it does consume the most water, and that's why you often see these data centers putting in that type of cooling system.

29:53

Smaller data centers typically use air cooling, and air cooling is a closed loop system, so it doesn't really use a lot of water.

30:03

It uses a moderate amount of electricity, and it's a little bit more expensive.

30:08

And so, with air cooling, it's just like it sounds like it's literally you've got like a hot aisle and a cold aisle, and you've got cold air that's drawn in through the front of the equipment, and then the hot air comes out the back, and then it's cooled again.

30:22

Some of the developments coming is this liquid cooling and immersion cooling, and that's what we're gonna start to see in the next couple of years, and this is where the major chip manufacturers are heading, which is you're taking the whole thing and immersing it into a liquid, and then in a closed loop, you're cooling the liquid.

30:44

Hopefully, that all made sense.

30:48

Now we're gonna talk about a case study.

30:50

Uh, you were asking about uh other communities that have these hyperscale data centers or data centers in general and what they've done.

30:57

So in Mesa, Arizona, they created an economic hub.

31:01

They had an opportunity where they had open land when the General Motors Proving Grounds closed, and they had the resources.

31:08

There was open land, there was the fiber optic cable that we need in order for the high-speed data centers to be able to talk.

31:16

Uh there was power, there was a substation nearby, they had natural gas, and they built a brand new 24 million gallon per day water reclamation plant, and so they had reclaimed water and freshwater availability.

31:30

So by gathering all these resources into one particular area, they were able to create this particular corridor.

31:38

In this corridor, uh you've got Apple, Google, Amazon is building, and Meta has a hyperscale data center.

31:47

I want to talk a little bit about Meta's hyperscale, hyperscale, that's a hard to say.

31:52

Meta's hyperscale data center, because Meta partnered with the community and partnered with Mesa in order to reach a net positive water usage by 2030.

32:06

And you're hearing this a lot from data centers, and what does that really kind of mean?

32:11

It doesn't mean that they're going to consume less water than they put back into the system, but instead, what they've done is they've looked at efficient ways of cooling the data center that don't only involve evaporative cooling and the use of water.

32:27

So they're actually using AI to look at like the temperature of the outside air.

32:32

Is that outside air cool enough where we can just bring that in to cool the data center?

32:37

And so they're using different types of technologies to create a more efficient way of cooling the data center.

32:44

The other thing they're doing is they're investing with Salt River Project and the APS, which are the two main utilities for water and energy into conservation, and how do we how do they put more water into the aquifer than they're taking out?

33:02

And so this is a good example of where there was a good partnership between the city and the data centers.

33:09

And just on a personal note, I don't know if you all know, but I'm actually from Mesa, Arizona.

33:15

And my house is located just uh within a mile of these data centers.

33:22

With that, I'm gonna turn it over to Pat who's gonna talk about zoning and regulations.

33:30

Thanks, Mark.

33:31

Good morning.

33:31

Pat Young, planning and development.

33:33

Uh to talk a little bit about how we regulate data centers from the zoning ordinance in the UDO.

33:39

You heard from Mark that these facilities, these data center Fazellies can have impacts.

33:43

They can uh produce vibration, they can create noise, and they uh can create heat islands.

33:49

So we uh have zoning regulations in place to help address and manage those concerns.

33:54

I'll go into a little bit of detail here on the on those.

33:58

Um the small-scale edge centers, that was that first column, and I'll show the graphic here in a moment that you saw from Mark.

34:04

These are uh under uh usually under 20,000 square feet and five megawatts of energy usage.

34:10

These are normally co-located with cell towers, as you heard from Mark.

34:14

They help boost cell towers so that they can handle the data demands of AI associated apps and other um data usage.

34:24

We classify those in the zoning code as a minor utility, um, and those are allowed by right in all zoning districts.

34:31

Again, the uh as you heard from Mark, the technological limitations currently are that they need to be very proximate to the use they're supporting.

34:40

So these are typically going to be near um existing data centers that need additional support or um cell towers or other similar things.

34:49

All other data centers other than these very small scale, what are called small scale edge centers are classified as a major utility.

34:57

Major utilities are limited to the um IX, industrial mixed use, and IH heavy industrial zoning districts.

35:04

And very importantly, and relevant to our um controls and the protections for our community is only after the issuance of a special use permit by the board of adjustment.

35:14

So, as most of you are aware, the special use permit process is very rigorous, it's quasi-judicial, which means there's evidence and fact finding done.

35:22

This is put on by the Board of Adjustment, which is of course a body appointed by council that is working with our legal team, is expert at making findings of fact and analyzing and assessing uh any potential impacts uh on the site, and that if there are negative impacts that are not able to be mitigated, the board must deny the application.

35:46

If they are able to be mitigated and meet the findings that go with the use permit, then it then it must be approved.

35:52

But these issues of noise, water usage, um, uh vibration and heat can be addressed through the special use permit process and the criteria therein.

36:04

Um so here's that uh same chart that you saw from Mark showing that last row shows that again, just those edge data centers, the small ones, and it's that uh graphic there is a really good example of what those typically look like would be permitted as a minor utility.

36:22

All the other ones would require uh a major utility uh um be classified as major utility and require special use permit and are limited to those industrial zoning districts.

36:34

It's also I'll go ahead and throw in too that it's my professional opinion, and it's just that 31 years of being an urban planner that the land values in Raleigh are too high for hyperscale centers to make economic sense.

36:46

It's certainly possible.

36:48

There's always evolutions, but at this point, what the demand we see are are for that that enterprise and co-located level, uh, and they would have to go through that major utility uh assessment through the special use permit process.

37:02

Um, state law requires that we uh as administrating the zoning code fit uses into the closest category we have in our code.

37:11

We do not have a category for AI data centers.

37:14

Optimally, we would want to develop some very clear standards that are directly targeted at AI data centers and bring those to you through a tax amendment process.

37:22

Um there was a law passed in 2024 that you all are all very familiar with at this point, cited on the slide here that would prevent us from doing anything that would be considered downzoning, and uh really any changes that we could contemplate to the standards uh would be considered downzoning.

37:39

It is staff's opinion though that the rubric that we have in place that I showed in the previous slide, I'll put up again for a moment, really does address all those interests through all of those areas of concern and interest by requiring the special use permit process for anything other than the small scale edge centers.

37:58

So addition uh and so in addition to the zoning controls, the uh our public utility department is working on some controls that will address water usage and protect our water resources.

38:09

I'm gonna turn it over to Witt Wheeler with Riley Water to talk a little bit about those and we'll be available for questions after what presents.

38:18

Thanks, Pat.

38:20

Okay, so just to recap here, we do have existing data centers in our system.

38:25

They've been around for a long time.

38:27

Uh, but the water use on those data centers is less than 25,000 gallons per day in all in all situations.

38:36

And these data centers, as Mark mentioned, they support um telecommunication utilities, they support our banking community, and so that is the purpose.

38:46

We do not have any AI generative uh data center sites.

38:52

I did want to mention uh our reuse water system.

38:55

So we currently operate a reclaimed water system, and that system does have customers that use this water for evaporative cooling.

39:04

Um examples that I can give you are NC State University, Centennial Campus uses uh reclaim water, their physical plant to dissipate the heat.

39:12

Uh, so does WakeMed.

39:14

Wake Med uses that for all their cooling at the main hospital campus there.

39:19

We do have about two million gallons per day of capacity in that reclaimed water system that could be used for other evaporative cooling projects.

39:30

However, um that is not an ongoing solution because our our long-term uh plan assumes that we return 90% of the water we use to the Neuse River.

39:45

Okay, water rates and development system development fees.

39:49

So data centers are considered an industrial class, and in North Carolina, our laws require that we have the same rates for all class members, which means that a data center would have would have the same water rate as a pharmaceutical plant or manufacturing plant.

40:07

So that's that's the way those would be locked in.

40:10

And then system development fees, all new users do pay system development fees, but those fees are heavily regulated.

40:18

There is a very prescriptive state process for calculating that, and we are currently working on a revised version of that that would be available for our next budget session next year.

40:35

Okay, so we were not able to get someone here from Duke Energy, but our good friend Derek Reimer sent over some talking points that we could share with you today.

40:45

Um like I said, they've been providing data, um, providing energy to data centers for over 20 years.

40:52

Um, they do not have discounted rates or develop development incentives, and they do have contract cost control.

40:59

So their large contract users are have a contract that requires them to bear the cost of provide the service being provided to them to protect the existing customer base.

41:14

And of course, most of their customers are residential and commercial, um, but those large contracts that are in place.

41:23

Um, I think Derek said they had four, I think, yeah, 4.8 gigawatts in new large customer contracts.

41:32

Um they expect the financial analysis to show about $3.6 billion in benefits, it would actually support residential customers.

41:43

Um they estimate that that would be like a five to six dollar um reduction in residential bills from now through the next 15 years.

41:54

And the new contracts also do provide a clause for power curtailment.

41:59

So if there's a power shortage, uh they can reduce power those facilities.

42:05

And lastly, less than one percent of data centers make up the peak demand, but they do expect that to grow to 13% by 2030.

42:17

Okay, going around in the General Assembly, there is a bill, Senate Bill 730, ratepayer protection act that has passed the House is currently in the um in the Senate Rules and Operations Committee for review, but that bill applies to data centers that consume more than 100 megawatts or greater of energy.

42:39

It had it provides for a rigorous site assessment process, which uh in particular sound, you've heard that mentioned.

42:47

Um the bill has a requirement that there's a there's a study looking at residents and schools within 500 feet of that boundary.

42:57

There's a comprehensive resource um impact study, uh looking at water resources, air quality, uh thermal plumes, um, agricultural resources and forestry.

43:11

And the bill also would require Department of Environmental Quality to adopt water use standards for data centers, and it also requires that the developer cover the energy expansion, which is consistent with the information that Derek gave us.

43:28

And finally, it would prohibit prohibit a local local government prov from providing incentives or using eminent domain for citing a data center.

43:39

Okay, so what to watch?

43:41

Um, as a city, we want to look for ways to minimize the environmental impacts for our use.

43:47

So IT and sustainability will be training our staff on the proper use of AI so that it doesn't create an unnecessary demand.

43:56

Um new technology would um they would be looking at vendors to have sustainability goals so that we're choosing products that are environmentally responsive, and also the environmental partnerships.

44:11

We're gonna be watching that.

44:12

So Mark mentioned the Mesa Arizona example.

44:15

We also know that both um Microsoft and Amazon have made public commitments to be water positive in by 2030.

44:26

We'll be continue to monitor state and federal regulations, and of course we'll be looking at technological advances.

44:33

So um, you know, microchips that are designed to dissipate heat better, or um processing that uses edge data centers instead of putting all that back on the hyperscale data center, sharing that that load.

44:49

So those are those are things we'll be watching in the future.

44:53

And with that, we'll take questions.

44:56

Okay.

44:57

Thank you all for presenting, and we will um I'll start with Council Silver and then go to Mayor Pro Tem Harris.

45:04

So I have two questions.

45:05

One is for Patrick and the other from Mark.

45:08

Um Patrick, just as a follow-up, could you just provide us?

45:11

I know we can do it by going to IMAPs.

45:14

Just the location of IX and IH.

45:17

Within the city.

45:18

So I'm I pretty much could figure out where it's located near RDU and in district C, but it'd be helpful just to know those locations of where data centers, whether minor or major, like well, you said minor is throughout all districts, where some of the major uh data centers may uh appear.

45:40

Sure.

45:29

So great question, councilmember.

45:42

So the um generally we'll share a map um highlighting those areas with you all um following this session.

45:48

But uh in general, you mentioned some areas of Southeast Raleigh.

45:51

There's also Capitol Boulevard, we have some IX uh mixed mixed use IX.

45:56

Um it's the criteria for establishing these industrial districts is that they are almost exclusively distant from existing residential or planned residential, and um because they were originally designed almost exclusively for higher intensity and traditional smokestack industrial uses, so they were uh are located in areas that are not proximate to existing or planned residential.

46:25

Um they are scattered throughout the community but kind of focused on major corridors, um, such as Capitol Boulevard, uh uh South Saunders, Garner Road, areas like that.

46:36

Okay, thank you.

46:37

Now I don't know this is a question for you from Mark, but you had mentioned the ATT data center.

46:42

Uh I don't know if the trend because land could be constrained, that's a very tall structure, and of course, it's not a lot of windows, but are you seeing a trend anywhere across the country because of the limited availability of land?

46:57

People still prefer to fill out a footprint on one story, but it places a land constraint.

47:02

Are we seeing any approaches where they're going maybe two or three stories versus occupying, you know, 100 acres of land for a data center?

47:11

The trend is typically to keep the data center on one floor because the equipment is very uh heavy, large, and hard to move.

47:19

And so when you start to go vertical, it adds cost to the structure to support all of that weight and also to move that equipment.

47:26

So uh as Pat mentioned uh in his thoughts, is there really looking for lots of land that's cheap and near the resources that they need, and if possible, keeping it onto one floor.

47:42

Mayor Perton.

47:44

Yeah, and thinking about AI data centers, I have four concerns in mind.

47:48

First, I want to protect customers from higher utility bills.

47:52

So, you know, energy bills, they're increasing dramatically, and we obviously have finite water supplies, so these race sources have to be safeguarded.

48:01

I appreciate this information about the variety of centers and different uses or usage of our resources.

48:09

I also want to avoid negative environmental impacts, so whether that's protecting drinking water supply, the noise, the vibration, the heat.

48:17

I don't want to see further heat islands in our city.

48:20

Um, third, I definitely want to see more clean energy for our grid.

48:23

So the more that we can double down in investments in water, sorry, uh wind and solar, um, so that we're not using all these fossil fuels for these types of centers.

48:32

I think that is critical for a sustainable future for our city.

48:36

And then fourth, I want to make sure that we're listening to any uh frontline communities that are impacted by these data centers.

48:42

So when there are ideas or requests coming forward in our city for these, I want to ensure that there's not people experiencing outsized impacts if they're next door.

48:52

Um again, I really appreciate the presentation and just the spectrum of different centers out there.

48:57

It gives me a better sense of what to be, I'd say not so worried about and maybe more worried about.

49:03

Power and water demand, they do vary greatly, and this body is gonna be tasked with evaluating whether these different types of centers are appropriate at times for our city.

49:13

And I know there are other communities enacting moratoriums on AI data centers, and I am open to that conversation here.

49:21

I think if we are to be responsible stewards for our resources, we do need to look at those high-impact data centers, especially if state law hasn't kept up with the advances yet.

49:32

How are we going to be proactive?

49:34

If 13% of peak energy demand is going to be used for these by 2030 in our state, that does, you know, give me some heartache.

49:42

And so I guess two questions.

49:44

Um, first, is the Board of Adjustment well informed enough about these data centers to make rulings on special use permits?

49:52

Are they gonna receive this presentation?

49:54

Do they have the expertise to evaluate impact?

49:58

Yeah, thank you, Councilman, for that question.

49:59

Pat Young with planning and development.

50:02

I, to my knowledge, none of the board adjustment members are expert on this topic, but as is typically the case with any of the cases they hear, they receive testimony, evidence, information that helps ground them in facts.

50:16

Their role and their legal requirement is to do fact finding, and they um are not shy about asking for data information to make sure that they have the information they need to make good decisions.

50:27

So we certainly can prov provide this to them proactively, but certainly if they are going to receive a case, we would make sure they have all the information they need to make good decisions.

50:37

Okay, yeah, I think that is definitely critical if that is kind of our our main safeguard.

50:42

And then also curious, and this might be also for you, Pat, or maybe for Karen, what is the status of the no downzoning rule at the state level?

50:51

I know there's been legislation considered.

50:54

Where are we with that?

50:56

That legislation has not passed, so it is still effective.

51:00

Okay, so I think that's another tool that we don't have access to at the moment, but I want to make sure that we're just keeping uh keeping up on.

51:08

Thank you.

51:09

Absolutely.

51:10

Okay.

51:12

Council member, could I have one?

51:13

Um, you mentioned moratoria, and I wanted to just cite there is a state statute on moratoria that really contemplates the idea that a local government will pause development of a certain kind to create criteria and standards that allow the use rather than permanently banning it.

51:30

So certainly we encourage council to have any conversation they want to have, but I just want to note that staff's position is that we have between the water use standards that are being worked on and the zoning limitations that the public interest is protected, but we we certainly want you to continue to have that conversation, and I just want to get on the record that there are um considerations under state statute for moratorium.

51:54

Thanks.

51:56

I guess I would just add if council is interested in that, that probably is a discussion that's more appropriate for a closed session than at this time at the table.

52:07

Counselor Patton.

52:09

Yeah, I have a bunch, but I'll start with a couple.

52:12

Um first, I'm interested in how all the water is connected.

52:19

Uh, you know, and so while we don't have hyperscale data centers on our system at this time, and we can, you know, do our best to influence the policies that were under our control.

52:30

If our neighbors on a different water supply enact different policies or or take a different approach, is how like how does the broader water table all work together?

52:43

Yeah, so the um let's say the Cape Fear River Basin adjacent to our news river basin.

52:50

Um, because of the geology in this area, the the water tables will not depend it on each other.

52:57

So we're not really concerned about um the Cape Fear consuming water that would affect the Nice River Basin.

53:04

Okay, so in like Jordan Lake, if they users of Jordan Lake adopt different policies, that's not gonna be that will not affect Falls Lake or News River.

53:13

Got it.

53:14

Okay, helpful.

53:15

Um then I do you know anything about the gosh, the waste byproduct that's produced by the evaporative cooling, like the water that is evaporating off, is it contaminated with?

53:32

Um I don't know specifically um from a traditional a traditional cooling tower, I would not think it would be contaminated.

53:40

It would be, you know, like the ones that we serve now.

53:43

It's just you know, there's a lot of heat absorbed when water goes from liquid to gas phase.

53:49

That's why it's that's why it's so effective.

53:51

Okay, great.

53:52

Mark, no care.

53:54

And then uh maybe one last one for water.

53:57

You mentioned data data centers would be in an industrial class, along with like pharmaceutical and manufacturing.

54:04

What are the other classes and if you know like general percentages of uses, like how much is residential use of our water, etc.?

54:13

Yeah, so um so we don't we do not have an industrial rate.

54:19

You know, we only have two rates now.

54:21

We have a commercial rate and we have a residential rate.

54:24

And so uh the majority of our use is residential.

54:30

So the top, let's see, 10 consumers would be about 10 percent of our system.

54:39

That is that that gives you a balance.

54:41

So, okay.

54:43

And but we do have, and so we only have two classes now, residential and commercial, right?

54:47

But we have the we have a sufficient authority to establish industrial classes and that's what you're gonna bring us back.

54:53

Next budget.

54:54

Okay, got it.

54:55

I can pause for now.

54:57

Okay.

54:57

Uh Counselor Silver, you had another.

55:00

This may be a question for Patrick, but anybody can chime in, potentially Mark.

55:05

Patrick, you said something about land cost and why having a larger site could be problematic.

55:13

I guess my question is proximity to a metro location.

55:18

I'm sure there are many rural areas in this state that have a lot of land, uh, but I just don't know in terms of the infrastructure to support a data center.

55:28

I'm just trying to find out the nexus between the two.

55:32

While we probably have the infrastructure in the metropolitan areas throughout the state, uh land costs are higher.

55:39

If you go to rural areas, you have the land availability with a lower cost, but may not have the infrastructure.

55:45

Just trying to understand, I won't want to say threat, but potential for one of these classes of data centers coming to the Raleigh metro area.

55:55

Uh so that's kind of my question.

55:57

Yeah, that's a great question, Councilman.

55:59

I'll characterize it generally and turn over to Mark for some more color and detail.

56:03

All of the hyperscale centers that have been constructed or proposed in the Southeast are in rural areas.

56:10

But the metro areas are starting to see these kind of regional edge centers, kind of that hundred thousand square feet, and you know, closer to 10 or 20 megawatts.

56:21

Um and my understanding from talking to Mark, I'll let him elaborate is there's kind of a cascading a need for proximity to users on different types, but I'll let him elaborate.

56:33

See if I can go back to this slide.

56:36

Uh Pat, you're absolutely correct.

56:39

Uh when I mentioned like edge data centers, those need to be really close to the workload that they're supporting, as opposed to a hyperscale data center, needs fiber to connect it, but they're more of regional, and so they can be hundreds of miles from the users they support.

56:55

But then when you start to talk about like enterprise corporate co-location cloud regional, corporate uh what we do is we want to make sure that we have resiliency, and so we want one corporate data center that's in Raleigh, but then another one that is outside outside of our zone.

57:13

Uh and by zone I mean outside of like the area of we're talking about hurricanes.

57:19

You know, we want to make sure that we have an alternate site that's outside of the hurricane zone or outside of a disaster zone.

57:25

And so we have two locations, one that's near, one that's far.

57:32

Counselor Fort.

57:33

Yeah, this question I think is mostly for WIT.

57:35

Um others may waiting if they've had other conversations.

57:40

Uh I think we need to remind folks that the water um usage is not just for the city of Raleigh.

57:46

We serve, was it six other municipalities?

57:48

Six other municipalities.

57:49

And have you had conversations with the folks in those municipalities about this emerging issue as well as with folks in at the county level?

57:58

We have.

57:59

We have had the conversations with our utility partners, and uh, you know, our agreements with those um partners have capacity limitations.

58:10

So most of them are concerned about using their available capacity for such a large use.

58:17

That's a big consideration for them.

58:19

Okay.

58:20

Thank you.

58:22

Councilor Jones.

58:25

I just have two questions.

58:27

Um, first one I think is from Mark.

58:31

And I'm just wondering how are we currently tracking our impact on the any data centers that we are currently using for AI initiatives like Ask Raleigh?

58:39

I know we don't have I'm not from what I understood, we don't have them here, but wherever they are, are we tracking our impact on those communities as well?

58:47

Uh we're not, no.

58:49

And where are those?

58:51

Uh that's a really good question.

58:53

That's the hard part about cloud computing is that you really don't understand or know where your workload is happening.

58:59

It's kind of like the same thing when you pick up your phone and use your phone.

59:03

You could be using a data center in Raleigh, or you could be using a data center in Mesa, Arizona.

59:08

And so you really just don't know.

59:10

And it's based upon Amazon, Microsoft, and the big cloud providers to balance those loads.

59:16

So there's no way to find that out.

59:18

Got it.

59:19

Um, and then I'm not sure uh maybe it's for you as well, but we've talked about water and we've talked about power.

59:25

As we scale up in years to come, those are we see those impacts right now.

59:29

Do we have a way of tracking any other long-term impacts that we may see that are outside of this that might come up?

59:36

Like, side effects.

59:39

Uh that's a very good question.

59:40

Thank you.

59:41

We've been talking with directly with NVIDIA, with Dell, with Microsoft and the big players.

59:47

Uh I've been able to connect with uh Dell's sustainability team.

59:51

And so what I've been doing is tracking what they're doing because again, we're using their equipment, we're using other people's data centers.

59:59

And so for us, how we track these things are through their resources and their sustainability.

1:00:05

The other thing I would say is that uh I mentioned briefly about chip development, is we're seeing from NVIDIA and from the big uh makers of these AI chips just huge increases in the amount of uh compute power, but then a decrease in the amount of electricity, cooling, and other types of resources.

1:00:28

And that's something we've just seen throughout decades in information technology and we will continue to see.

1:00:34

Awesome.

1:00:34

Thank you so much.

1:00:36

I have some questions.

1:00:38

I'll go.

1:00:39

Um I want to just build on Counselor Silver's question about the map showing the zoning where this could happen.

1:00:47

And could you specifically highlight if there are any industrial zoning codes that are proximate to residential?

1:00:56

I know that's not typically what happens, but I'm sure there are incidences where that does, and we just want to be particularly sensitive to what happened in those circumstances and and how do we want to think about that.

1:01:09

Yep, Mayor, we can certainly do that.

1:01:10

We'll get to that map and identify the properties that are within uh proximity, say existing or allowed residential.

1:01:17

I will know I'll reiterate that that if the if there's a proposal on those sites, they'd have to go to that special use permit process and those impacts would be assessed, but uh your concern is very valid, and we'll identify that fully for you all.

1:01:28

It's great.

1:01:28

And then I don't know whether this is you, but you know, we talked about reuse and recycled water, and I know that Raleigh has, you know, plans to expand our system on that.

1:01:38

I would also be curious to know, you know, where are the where is that network of uh utility and do you know where are those relative to these sites so that we know even, right?

1:01:51

Additionally, are these sites going to be meeting uh I the legislation hasn't passed yet, but certainly that would be an attractive feature.

1:01:59

Right.

1:02:00

So the we have two reclaimed water systems.

1:02:03

Uh the one in there's one in Raleigh, that's the largest one, and then there is one in Zebulin.

1:02:08

Um the one in Raleigh has a backbone that goes from the Neuse River plant, it goes up through uh Walnut Creek Amphitheater down through Worthdale Park, generally following the Walnut Creek basin.

1:02:21

Okay, and makes its way all the way to Centennial Campus.

1:02:24

Okay.

1:02:25

Now you know we're we're doing a resilience project to put a water transmission main that'll go through Dix Park.

1:02:31

There's also a reclaimed water line that will be laid in parallel with that line.

1:02:36

Okay.

1:02:38

Then this is just granular, but the ATT data center next door.

1:02:42

I mean, literally, right?

1:02:43

Oh, it's right next door.

1:02:44

How can you just give us the is how many megawatts is that?

1:02:48

And that would be a corporate level center.

1:02:51

Yeah, that's a small, yeah, yeah.

1:02:54

Go ahead with the with the megawatts.

1:02:56

I honestly don't know.

1:02:59

Yes, it's in the corporate enterprise corporate uh structure.

1:03:02

Okay, and less than five megawatts.

1:03:04

Okay, that's just and then the final question uh I know that Duke Energy is not here, and we do not control a lot of what they do, but if you go back to that slide uh where you had a few points they had given you, what I don't really understand is we know they have a hundred billion dollars of investment plans, right, to expand their transmission and energy generation in the state in anticipation of growth and AI, but yet they said this would save you five to six dollars off a utility bill.

1:03:39

I know y'all aren't Duke Energy, and you're just presenting what they gave you, but that I f I don't quite understand that because my sense is as as a city, we are really gotta think about how we mitigate the increases in utility costs that come as they expand their network.

1:03:56

And I would just like to have some follow-up on how they are saying this is now gonna save us money.

1:04:04

We'll do that.

1:04:06

Okay.

1:04:08

Back to all good questions.

1:04:11

Um to add to the list related to where is the heavy like where are the permitted zoning areas um would just be like total land area, if that could be provided when the map is provided, like total land area.

1:04:25

We'll provide that, thank you.

1:04:26

Um and then I am curious, maybe for Mark, these um you're saying servers and and equipment then tended to need to be replaced every five years, but these have such intense computing power, they're being replaced more often.

1:04:41

And if you could if you know anything or can speak to um impact on like rare earth minerals, like there's chips in everything now, and it's in like use deploying chips to replace AI servers as opposed to putting them in the cars, and make you know it's creating competition, which can it's driving up even like the cost of your car and your fridge and other things like that.

1:05:03

So that piece, chips.

1:05:06

Uh absolutely we're working actually with sustainability, and so one of the things we're looking at is just what are we doing as far as uh our electronic waste.

1:05:16

So, how can we recycle the electronic waste?

1:05:19

How can we reuse it?

1:05:20

Uh again, these AI vendors, they're on this quick turnaround.

1:05:25

However, those processors still have life.

1:05:28

And so the idea, and this is what we do is we try to turn it over while there's still usable life, and so we're not sending this uh to the landfill.

1:05:38

We're reusing it, and then eventually the idea would be to return and recycle those rare minerals.

1:05:44

Got it.

1:05:45

Helpful.

1:05:46

Um, and then one more, you know, there's this is a topic of a lot of community concern, and there's, you know, both like policy changes and guardrails that need to accommodate, you know, to rise to meet the moment at federal, state, and local levels, but on an individual level for residents who may have a concern and may be interested in their um other individual actions do or don't play into all of this.

1:06:09

Are there um at the sort of like front and regular user level differences in the impacts of the ways we interact with AI?

1:06:20

So it's like a is that Google summary use using less drawing less than say like generating a cartoon image using AI.

1:06:30

And that's exactly what we're doing in our training is we're doing what's called prompt engineering, which is ensuring that as you prompt AI for information that your prompt uh is efficient, effective, and that you think a little bit about the energy use.

1:06:48

And uh I joked earlier with Witt, it's yeah, he's starting to laugh already.

1:06:53

It's like don't use AI to generate yourself a picture of yourself with Bigfoot, right?

1:06:58

Just think about that impact that that is having, and that's what people I don't think really realize at a residential is that even though these may be fun, that there is a resource impact at the end of the day.

1:07:13

Counselor Branch.

1:07:14

Yeah, definitely thank you for the information.

1:07:16

Um I live a little bit in this world in being in IT.

1:07:20

Um, and part of that is uh there are a lot of data centers in RTP.

1:07:25

So my question is what is RTP get is water from?

1:07:33

Um mostly Durham and Kerry water system.

1:07:38

Okay, okay, so that's helpful because I know I know there's a lot of reuse of some buildings and infrastructure as RTP rebrands itself.

1:07:46

Um, which leads me to the point of some of our um with equipment having a five-year and tenure, this maybe from Mark, used to be land switch.

1:07:57

Now with the new two to five year switch of technology.

1:08:01

Um, what's the possibility of buildings having a reuse at a sooner left time frame than historically in the past?

1:08:12

Where in the past you may have a data center and it lasts 20 30 years, but with technology changing, I see that turnover shrinking to maybe five to ten.

1:08:22

And I want to know what could that impact be as well.

1:08:26

The impact is really going to be the cooling system itself is if you have an older building that only has evaporative cooling, and as these AI uh chips come about that are liquid cooled, those buildings are going to have to upgrade their systems.

1:08:41

And so that's really where I believe the investment is going to come.

1:08:44

Not the building itself, but it's the components that make up the cooling system, the electrical usage, uh, and then the fiber optic cables within the building.

1:08:54

Okay.

1:08:55

Yeah, and thank you.

1:08:56

And I do thank you for the comments about you know how we use AI and generate things.

1:09:01

And I do want to keep in mind too is every time you go to that local department store online, you are using a data center.

1:09:12

Any other follow-ups?

1:09:14

All right.

1:09:14

Really appreciate all the information we received and a good conversation, and we are adjourned until 1 p.m.

1:09:24

So you do what?

1:09:26

I don't even use GPT, I don't use any of it, but everything is changing.

Discussion Breakdown — Share of Meeting
Technology and Innovation███████████████████████████████████████39%
Engineering And Infrastructure███████████████████████████27%
Environmental Protection████████████████████20%
Procedural██████████10%
Water And Wastewater Management██2%
Economic Development1%
Infrastructure1%
Summary of Proceedings

Raleigh City Council Work Session on Data Centers – June 16, 2026

The Raleigh City Council held a work session to discuss the impacts, regulations, and future considerations for data centers, particularly hyperscale and AI-driven facilities. Staff from Information Technology, Planning and Development, and Raleigh Water presented an overview of data center types, community benefits and challenges, existing zoning and utility controls, and emerging trends. Council members raised concerns about water and energy use, environmental impacts, utility costs, and the need for proactive safeguards.

Discussion Items

  • Data Center Overview (Mark Wittenberg, IT): Mark described the growth drivers (cloud computing, AI, cybersecurity) and types of data centers: edge (small, near service), enterprise/corporate (e.g., AT&T facility next to city hall), co-location/cloud regional, and hyperscale (large, high-power, high-water). He noted Raleigh has no hyperscale centers or requests for them. Cooling systems vary: closed-loop air cooling (low water), and open-loop evaporative cooling (higher water, often cheapest). Emerging liquid/immersion cooling may reduce water use. A case study of Mesa, Arizona showed partnerships for net-positive water through technology and conservation.
  • Zoning Regulations (Pat Young, Planning): Small edge data centers (<20,000 sq ft, ≤5 MW) are classified as minor utilities, allowed by right in all districts. All other data centers are major utilities, limited to IX (Industrial Mixed-Use) and IH (Heavy Industrial) districts, requiring a special use permit from the Board of Adjustment—a rigorous, quasi-judicial process. Young stated land values in Raleigh are likely too high for hyperscale centers. He noted that state law (2024) prevents downzoning, but staff believe existing rules adequately address impacts.
  • Utility Considerations (Witt Wheeler, Raleigh Water): Existing data centers use less than 25,000 gallons per day. Raleigh's reclaimed water system has ~2 million gallons per day of capacity for evaporative cooling, but long-term plans prioritize returning 90% of water to the Neuse River. Data centers would pay industrial water rates (same as pharmaceutical or manufacturing) and system development fees.
  • Duke Energy Information (presented by staff): Duke does not offer discounted rates or development incentives; large contracts require users to bear costs. They estimate $3.6 billion in benefits, potentially reducing residential bills by $5–$6 over 15 years, and include power curtailment clauses. Less than 1% of peak demand currently comes from data centers, expected to grow to 13% by 2030.
  • State Legislation (Senate Bill 730): The bill would apply to data centers consuming ≥100 MW, requiring site assessments, water use standards, and developer-paid energy expansion, while prohibiting local incentives or eminent domain for siting.
  • Council Questions and Comments:
    • Councilor Silver requested a map of IX and IH districts and asked if data centers could be built vertically; Mark responded that one-story designs are typical due to equipment weight and cost.
    • Mayor Pro Tem Harris raised four concerns: protecting customers from higher utility bills, safeguarding finite water supplies, avoiding negative environmental impacts (drinking water, noise, vibration, heat islands), and ensuring frontline communities are not disproportionately impacted. She expressed openness to a moratorium on AI data centers.
    • Councilor Fort asked about water connections with neighboring systems; Witt confirmed geological separation between river basins means policies in the Cape Fear basin do not affect the Neuse.
    • Councilor Jones asked about tracking Raleigh's own AI impacts; Mark noted it is not currently tracked due to cloud computing's opaque load balancing. He added work with vendors on sustainability and chip efficiency.
    • Mayor Williams-Butler requested mapping of industrial zones near residential areas and reclaimed water infrastructure locations. Staff will provide maps and total land area data.
    • Councilor Branch noted many data centers exist in RTP, which uses Durham and Cary water, and asked about building reuse as technology turnover shortens; Mark said cooling system upgrades will be the primary investment.

Key Outcomes

  • Staff will provide council with a map of IX and IH zoning districts, including proximity to residential areas and total land area.
  • Staff will follow up on Duke Energy's claimed $5–$6 residential bill reduction.
  • Council expressed interest in exploring a potential moratorium on high-impact AI data centers, though City Manager noted such discussion is more appropriate for closed session.
  • No votes were taken; the work session was informational to guide future policy.

Meeting Transcript

That's all for this episode of Wake TV. Keep up with all the latest Wake County news by visiting us online at Wake.gov slash news. And be sure to follow us on Facebook, Instagram, Twitter, and YouTube. So you can always stay in the loop. We'll see you next time. Well, we'll see. Well, we're not going to be able to do it. All right. We will call the work session to order. And on our canes photo to the uh data center information and infrastructure. We have Mark Wittenberg Information Technology. Good morning. Uh mayor, members of the council, and Madam City Manager. Yeah, that's a hard act to follow. But I am also joined by my colleagues here. I want you to know that this presentation was a collaborative effort between the Office of Sustainability, City Attorneys, IT, Planning and Development at Raleigh Water. First of all, just an agenda for today. We're going to talk a little bit about data centers, like what they are, the sizes, the different types, what they're used for. We're going to talk about the impacts on the community, both the benefits and some of the challenges that we're seeing. We're going to talk about zoning regulations and the tools that you have in your toolbox to be able to regulate where these are built and the sites. We're going to talk a little bit about the utility considerations, the energy, the power that these consume. And then we're going to bring it home with what to watch. In other words, what is staff looking at? What are we observing? What are the things that we are looking forward as far as uh developments in this space? So let's start off with what's driving data center expansion. And I'm going to start with today. If you drove into the council meeting, took a bus, uh, searched for something, used your phone, you've used a data center. If you watched the canes just dominate Vegas the other day, you most likely used a hyperscale data center somewhere. The point is that every day we're using these uh cell phones, we're using uh information, we're providing services, all of these consume and use data centers. One of the big expansions started with cloud computing. So, what is cloud computing? Cloud computing is basically just computers that are somebody else's. It's where you're leasing computer use or space in a data center. The reason that cloud computing and data centers kind of came to be is that by consolidating all of these things into one place, we can share the care the cooling, we can share the backup protection, and we can share the facility instead of us all building our own and consuming more resources. And so cloud computing was kind of that first foray into that market of consolidation of those resources. Digital services, I love to say, you know, there's an app for that. And there is. Poor, that's not what we're seeing today. What we're seeing now in this new AI world is we're seeing very directed cyber attacks that look very, very real. And so in the cybersecurity space, we need to make sure that we're staying up with this. Autonomous systems, these are these agents that are assisting and doing like daily tasks and automate uh providing automation around things that may have done uh been done by humans in the past but now are being done by computers. But I would say the primary driver for data center expansion and the reason that it's kind of all the talk is around artificial intelligence. Over the last two years, the development in artificial intelligence has just exponentially increased with generative AI, and now we're seeing these autonomous systems coming up, and you're just going to see this continue to grow exponentially. These systems require a lot of compute power. So different types of data center. I'm not gonna go through this all of this information, but I just wanted to give you a little brief overview that all data centers are not created equal. And I'm gonna start with the edge data center. And so these are the smallest size data centers, so think of something that's like a shipping container or smaller. And these are typically located near the service that they're providing.

SUMMARIZED BY OPENPUBLICA AI
TRANSCRIPT VIA PUBLIC VIDEO
openpublica.com