Advisory Committee on Data Centers - See below for additional instructions on how to attend the meeting and provide public testimony.

Dane County, WI · · More Dane County, WI meetings · More Wisconsin meetings

Transcript

Download: Text · SRT
AI TRANSCRIPT

This transcript was generated automatically from audio using AI and hasn't been reviewed by a person — it can contain mistakes, including plausible-sounding sentences that were never actually said. Treat it as a starting point, not a verbatim record.

[0:00] And so with that, I'm going to call the advisory committee for data centers to order and I thought I did my agenda up and I don't, it's not confirmed.
[0:18] Yeah, so next under agenda would be consideration of the minutes from our April 14 meeting.
[0:27] you do the call the order quickly. What's it all the order? Oh, I'm sorry. Call the roll. Thank you. Sorry.
[0:34] Richardson, present. Yoke here. What for first? Here. Helm Nelson. Here. Single Terry.
[0:43] About to present. Yellow girl. Bradshaw.
[0:49] Skindora. Here. Shanna Han. Here. Possible. Also here. Maran.
[0:55] here. And miles, miles here.
[1:00] All right, so a corner being present. Now we will move on to
[1:03] consideration of minutes from our April 14th meeting. Very motion. Motion to approve.
[1:09] Move by. We'll put first. Is there a second?
[1:11] 10 second. Second. My written to approve the minutes. Any questions? Corrections anything to
[1:17] the minutes. Hey, hearing none. I'll unfavorable. Say aye.
[1:21] I. Oh, sorry. I. I'm so sorry to be in a room. I must be a member.
[1:28] Stop. I would.
[1:29] Post, I know. That motion carries minutes are approved.
[1:35] Moving on down to status on prior follow-up items and requests.
[1:42] Way I see this portion of our agenda working is at three of our meetings.
[1:47] We'll be accumulating things that we want some follow-up.
[1:51] It will be sort of our to-do list or staff to-do list to close the loop on some of these
[1:57] topic areas, but in the meantime, there might be accumulation of items here.
[2:03] So from our last meeting, we talked about community benefit agreements, and there's some
[2:11] comments and questions around that, and I think, and we need to give some consideration
[2:19] to, I think, you know, Laura, particularly I mentioned, you know, need for some criteria and measures for those.
[2:32] Now, I think we're going to be seeing as we move through the other subject areas that
[2:37] that this CBA may be a tool that can be employed
[2:43] and to address concerns in a number of these areas.
[2:47] But as time was in the context of land use compatibility
[2:54] and environmental impacts,
[2:58] ecological corridors and so forth.
[3:01] So the context there was CBAs
[3:04] is to maybe leverage some community benefit around agricultural or ecological preservation of
[3:18] ecological corridors or analysis of those and so forth.
[3:22] More to come on that, I think, but it's certainly something I'll stay on the list for
[3:28] But let's see.
[3:33] Oh, this last item on here, I'm not sure what.
[3:38] I'm coming meeting topics and timeline, do we have a slide on that?
[3:44] Just obviously subject to change, but one of the requests from the committee last time was to
[3:48] get a kind of a sense of what we were going to talk about in the future, potentially, meeting
[3:53] meeting, so this meeting obviously is about noise lighting, aesthetics, and community character.
[3:59] tentatively we're looking at our June meeting to talk about economic benefits and the
[4:03] lifestyle, life span of a facility and the decommissioning that would come with that.
[4:09] Potentially July 14th, we're talking about water usage and storm water,
[4:14] and we would still have four more topic areas, energy consumption and emissions,
[4:19] like all transmissions emergency services and intergovernmental relations.
[4:25] So those are the one, two, three, four, eight more topics that we would cover.
[4:32] It's going to be depending on speakers and who we can get,
[4:36] but that, again, is just a potential tentative timeline.
[4:39] I would also add that those, those number bullet points there from the resolution,
[4:43] so if you're kind of wondering where those numbers are coming from,
[4:46] from the authorized resolution for this body, so that's a, that's what does referentive.
[4:54] So, Cynthia had mentioned that the next meeting might be nice to see in the very least,
[4:59] a ton of layout of what we'll be tackling next. Of course, as I said, subject to change,
[5:10] something on opportunity and so forth for people to present and be panelists here.
[5:19] Also in working with staff and laying this out, I'm fully cognizant of the fact that this
[5:27] is really optimistic for a timeline, and I really do suspect that we may need to be extending
[5:36] the life of this committee by a few months, perhaps.
[5:40] So,
[5:44] that leads me in a way to the next item on all of the items, and that's this
[5:50] Manituak moratorium near the end of April, Manituak County adopted a moratorium and consideration
[6:01] and approval of development for what I should say of permits and so forth for the purpose of the data center.
[6:19] So, at the outset of looking at developing the resolution to create this committee,
[6:27] it also looked at into whether or not the county could do a moratorium and was my assessment that
[6:35] there's state statute that's as county is explicitly prohibited from doing a development moratorium
[6:42] Manitwak has my view used in semantics to get around that definition of development moratorium
[6:52] And since they did it, I'm going to do it too.
[6:58] So it's all going to be putting in a,
[7:01] well, Man of Talk did it by Resolution, one fell swoop.
[7:06] Oh, and we got it up here on the screen here.
[7:08] A first paragraph actually is a, I think it's not alone.
[7:13] Yeah, it's actually an ordinance moment to their,
[7:18] And what I'm assuming is chapter A is their zoning code for us will have to be two pieces
[7:25] basically of legislation one will be an ordinance amendment to create a definition of
[7:30] data center or most likely be talking with corporate counsel and zoning division staff
[7:39] about an appropriate definition and then the resolution would be to play a more
[7:49] and I'm going to sync it up with basically the work of this committee, so I think if we extend the life of the committee for a few months, I would just pretty much be a year from now for that.
[8:06] So, um, so I just want to give that update and, um, let's see how that goes.
[8:15] Okay, any, any thoughts, comments on that, I mean, the more time things sort of a, sort of a big deal.
[8:25] So, go ahead something.
[8:27] I just say thank you very, very much.
[8:30] That is really excellent news.
[8:33] Really appreciate it.
[8:34] Yeah, I think it's the right thing to do.
[8:36] And the other thing I just want to compliment staff and the committee on the reading background materials that were sent out like a week ago, I found a number of them, I mean, I've been doing a lot of research, but I still keep learning things and so the one on noise pollution and the one out of Virginia were particularly informative, especially the data center industry cash campaign contributions and all these machinations that we can kind of feel and see hints of behind scenes, but, you know, it's kind of documented in some of these.
[9:05] So I really felt that was useful.
[9:08] So thank you.
[9:09] Thank you.
[9:11] All right.
[9:11] Anything else?
[9:13] And we will, if not, we'll move on.
[9:17] We don't have any reports that can make, right?
[9:20] So, one, let's see.
[9:27] And I jumped ahead, sorry.
[9:30] Presentations and discussion.
[9:31] And today we have, in the, to talk to us about community impacts, but I also want to, and I think
[9:40] Mjid or somebody had sent an email out, indicating that our guests today in the work they've
[9:47] done have, have looked into a number of areas that we're looking at.
[9:53] So we've invited Mark Griffin from the State of Virginia, the Virginia Joint Legislative
[10:01] audit review commission to discuss the work that they've done on the subject and we've
[10:12] invited him to feel free to go beyond just talking about the community impacts, but some
[10:17] these other related issues that we're wrestling with ourselves, so welcome Mark. I'm assuming
[10:25] he's online and ready to go. It's okay. Good. I will stop sharing Mark and give you the keys to
[10:31] share your slides. Mark, thank you so much for joining us today and sharing the experience and
[10:39] perspective of your work. So I'll have me appreciate it. Can you guys hear me all right? Yes. All right,
[10:46] Let me pull up my slides here.
[10:49] Well, I'm doing that.
[10:50] I'll let you know that I'm making my first trip
[10:52] to Wisconsin next month.
[10:54] I'm going on a paddling expedition around the Apostle Island.
[10:57] So I'm looking forward to that.
[10:58] Oh, that's wonderful.
[11:02] I can't go on the Colorado this year
[11:04] because it's too dry.
[11:09] All right, let's see.
[11:11] Is that, can you see it?
[11:13] Is that, all right?
[11:23] Okay, is that full screen for you guys?
[11:25] It is now, yep.
[11:26] All right, terrific.
[11:28] So, yes, good morning, Mark Reuben from Virginia.
[11:31] I work for a small state agency called J-Lark, and we are a legislative branch agency, and we do non-partisan objective research on subjects as directed by our general assembly.
[11:43] Today, I'm talking about our 2024 study of the state state center industry, and even though it's kind of specific to Virginia in many ways, there's lots of lessons that are applicable to y'all in Wisconsin.
[11:54] And so skip to the beginning.
[11:57] What are data centers that's just over here in guys talk
[12:00] sounds like you're pretty up to speed on this?
[12:03] So data centers are key hubs of the world's digital infrastructure
[12:06] and they support our model or digital economy and society.
[12:10] The industry has been growing very rapidly in the past couple of years,
[12:13] not just in Virginia, but everywhere else as I'm sure you've noticed.
[12:17] That's happening to one sport, the ever growing use of cloud
[12:21] digital services and two of course with the growth of AI and a lot of the big bets that are being placed on the AI by AI by IT companies.
[12:31] In Virginia and really in the rest of the U.S. there's really four companies that are driving most of that growth.
[12:37] So you might have a dance or billet being built by QTS or cloud HQ or any of these other companies.
[12:45] A lot of times they're doing that under contract with one of the big four, so they're building a facility, they have a lease in hand and they're building that to rent out the whole facility to one of the big four.
[12:59] A lot of data centers are large industrial scale buildings.
[13:02] You've never seen one.
[13:03] They look like massive distribution centers.
[13:05] They can have footprints that are bigger than the basketball arena or even football stadium.
[13:09] They're filled with computer servers.
[13:11] generate a lot of heat to employ these massive cooling systems, powering all that computing and
[13:16] cooling, of course, acquires an enormous amount of energy, so they often have their own electrical
[13:20] substation on properties served by a nearby transmission line.
[13:26] A modern data center uses substantially
[13:27] more energy than most other commercial or industrial operations. Back in the olden days of 10 years ago,
[13:34] a 20-megawatt facility was the biggest you'd see, and that's about equivalent to a mid-sized automobile
[13:39] plant. However, the hyperscale centers that are built now are much larger drawing from 100 to
[13:45] over 200 megawatts of power, some approaching 300 megawatts. And I think that some of those
[13:50] discussing the trend is the ability to use multifacility campuses and depending on the size
[13:55] that campus, the total draw, could be well north of 1,000 megawatts. By comparison, we have two nuclear
[14:01] plants and four nuclear reactors in Virginia. Each one of those reactors is a little shy of 1,000
[14:06] So, you know, one day center campus is a half the output of a nuclear power plant in Virginia.
[14:15] Our Gives in America is concentrated in northern Virginia in the suburbs of Washington, D.C.,
[14:19] but now we're actually seeing that expand southward towards the middle of the state.
[14:27] Gives in America is the largest in the world, kind of about 13% of global operational capacity
[14:32] and 25% of capacity in America.
[14:35] because that's us on the left.
[14:37] We became the sort of market leader for a number of reasons,
[14:40] including being located near the birthplace of the internet,
[14:43] being close to the federal government, it's contractors,
[14:46] strong fiber network, reliable and cheap energy,
[14:49] and at the state sales tax exemption,
[14:51] all those served to sort of help that grow this industry.
[14:56] The point here, and bringing that up with you guys,
[14:58] is that we've got the forefront of this...
[15:00] So we've probably encountered a lot of you friction points earlier than others have. But now we're starting to see those friction points emerging pretty much all across the country. So hopefully I have some good lessons to share with you.
[15:14] So first I'll talk about those local community impacts. And if you have any questions any point, feel free to shout-in-all. I can't see you right now because of what slides pulled up. But I am listening.
[15:26] All right, thanks.
[15:29] So first we found that a growing number of data centers are being built close to residential areas, impacting nearby residents, and some localities have taken steps to try and minimize those impacts.
[15:41] Again, because modern data centers are large industrial facilities, they are largely incompatible with residential use.
[15:47] In addition to the building size of parents, there are often surrounded by security fences and lighting.
[15:52] They regularly test these diesel generators and they have massive cooling systems that have potential to create noise.
[15:59] They often, as I mentioned earlier, have an electrical substation that's located onside or close to the property.
[16:04] In some cases, a new transmission might get built to serve them.
[16:08] That could either be a quite extension line from the main transmission line,
[16:13] or what you see in Northern Virginia where you have these high concentrations, they've had to run several additional transmission line loops to provide additional power and redundancy to serve the high concentration of data centers.
[16:26] It's always to say that this is a matter of data centers, not like they were in the old days or were not like an office building where you can kind of innocently put that in next to the neighborhood. These are industrial facilities.
[16:39] So, even so, we found that about one third operational data center properties in Virginia were within 200 feet of residential zone properties.
[16:48] Part of this is our fact of where these were built.
[16:50] They were built in these growing exorbit neighborhoods that are now our exorbit counties.
[16:57] They're now pretty much fully suburban counties.
[16:59] And so we have this data center growth happening alongside some of the fastest growing residential areas in the state.
[17:06] So even though these are both in 200 feet, that doesn't mean they're exactly 200 feet apart,
[17:11] but they're still pretty close.
[17:13] Those property lines are right about in each other.
[17:17] And you can see a couple of examples of this of neighborhoods.
[17:20] Either actually that exist or that were planned and how the data center properties of butt to those neighborhoods.
[17:27] There's also many others in addition to the ones we found that were under construction or planned that had not yet been built.
[17:34] So we have a couple examples of those plans as well.
[17:38] And again, we could have pulled a number of other examples.
[17:41] There's a lot of this going on in the counties where both of the industry is located.
[17:50] So facilities, the question is why is this happening?
[17:53] Localities allow data centers to be built next to residential for two reasons.
[17:57] The first appear to be an adequate planning zoning.
[18:00] So some localities were still thinking of data centers like office buildings,
[18:04] which is kind of how they were before, and so they have them as a non-industrial use, which
[18:10] allowed the now industrial facilities to go into areas that were zone for commercial or mixed
[18:15] use, so essentially a mismatch between what they were thinking of a data center was and where
[18:20] what it really was and where it could go. In other cases there were some zoning changes over time
[18:25] and the counties had an updated their zoning maps, so for example somebody wanted to build a neighborhood
[18:29] in an agricultural zone, so they rezone that agricultural to residential, they built a neighborhood
[18:34] there, well that agricultural zone was next to it, industrial zone, that never got changed,
[18:39] then the data center comes in by right and then industrial zone and now you have data center
[18:44] right next to the neighborhood. So that was one way it worked the other the other way that worked
[18:49] and we see this a lot, I have seen this a lot, it's that the elected boards were proving
[18:54] special permits or rezoning for specific data center projects even though they were
[18:59] clearly close to residential areas, it's pretty notable that sometimes this was happening
[19:03] against staff recommendations or against a planning commission recommendations, get down
[19:08] on how you guys are structured but in Virginia at the county level of a planning commission
[19:12] and you'll have a board of supervisors and a board of supervisors has that final authority.
[19:18] So while the board is prerogative to approve those projects as elected officials,
[19:22] it's not necessarily good policy.
[19:26] ever since I really during the report and even more so since then we've seen localities taking
[19:32] steps to better manage future development for example allowed in Fairfax counties which are two
[19:36] of our big northern genial counties, a lot of the data centers. They've reduced the number of
[19:41] zones that allow by-right data center development and they've also made zoning changes to address
[19:47] a number of developmental concerns, so pretty strict restrictions on heights,
[19:53] setbacks, all those
[19:54] kinds of design considerations.
[19:57] And so what looks like O'Calley is our responding to resident concerns, the success of
[20:01] those efforts ultimately depends on the elected boards, because again, they have the authority
[20:05] to grant exceptions to whatever new rules are put in place.
[20:11] I think somebody mentioned a noise being one of the special concerns that's definitely a
[20:16] special concern when it comes to data centers.
[20:19] So, I would say that every data center is different.
[20:23] So, some data centers, you can stand right outside of it, and you won't hear a thing.
[20:27] All you'll hear is the traffic on the street.
[20:30] Others, you can hear that data center when you're driving up the road to go see it.
[20:35] The noise that they make is like a sort of constant low frequency drone or home,
[20:40] so similar to a commercial air handling unit, or maybe a window AC,
[20:44] but not a much, much more magnified scale.
[20:47] While that noise isn't loud enough to damage hearing, in a few cases, it has been
[20:52] significant enough to affect the well-being of nearby residential areas.
[20:57] So, for example, there's one terrible story we talked about in the report where a family's
[21:01] house was positioned just wrong on a small hill of a data center and the way that the noise
[21:08] came off that data center in the way the house was positioned, essentially the house was kind of
[21:12] vibrating with the hum and the family had to sleep in the basement because they couldn't sleep in their veterans.
[21:19] That's really the worst case scenario we found, but there are plenty of other cases where that noise has been cost and in a neighborhood and it's been bothersome for residents.
[21:28] Resolving noise problems is really hard to do.
[21:31] That's in part because, again speaking to Virginia, local or local or noise ordinances are designed to deal with that type of situation.
[21:40] our ordinances are typically set up to deal with the party issue or the parking dog issue, not a
[21:46] constant low frequency home from an industrial facility. So the way it's set up is, you know,
[21:53] if you violate the noise ordinance that's so many decibels, your parties too loud,
[21:57] I'm going to give you 150 out of fine. That really doesn't work in this situation because one
[22:03] that the noise might not be long enough, or sorry, loud enough to trigger the violation,
[22:10] because the hum is a different sort of decibel system. I can get into a little bit if you are
[22:16] interested, but a low frequency decibels can be measured differently from, you know, when a
[22:24] party or a broken dog is, and also $150, civil fun, that's not really going to diswake anybody
[22:29] for doing business.
[22:32] So one of the things to think about is the best way to resolve and noise problems to never
[22:35] have it, and that can be handled by not allowing data centers close to residential areas
[22:40] or to other sensitive areas like parks or schools.
[22:47] Again, the one thing that we're concerned about in Virginia, because our state motto is based
[22:52] on our history, we have a lot of historic resources such as battlefields and archeological
[22:57] sites.
[22:57] a lot of these are up in the areas where a lot of the data center development is happening.
[23:03] And so just like any other development during construction and the onsite or nearby historical
[23:08] resource can be destroyed or started by development. After construction, if you have a data center
[23:14] for example, as was proposed and no mayor may not happen, there's a giant data center facility campus
[23:21] proposed to go next to the Manassus Battlefield National Park. The part of where the data center
[23:26] going to be built was part of the actual historical battlefield, so the concerns about not only
[23:31] disturbing historical artifacts from the construction, but then it would also, the campus would
[23:37] be visible from the national park. So to help understand what those potential impacts are,
[23:43] you can do these things called a great development studies. In Virginia, we get help call them phase
[23:49] one studies, which is basically looking to see if there's anything of historical significance
[23:53] that could be affected by development. If you find something, you can do additional studies,
[23:59] phase 2 and 3 studies to determine the significance of what's there and determine ways to mitigate
[24:04] those negative impacts such as, you know, this is how we could preserve this historic structure
[24:08] or this out could, you know, unearth and relocate these artifacts or at the very least, we can
[24:13] document what's there to allow for future research. By the other type, I think a test-on was
[24:19] the viewshed study, which examines, if we build a data center over here, what's the viewshed?
[24:25] What's it going to look like and how visited will be from, for example, in the NASA's
[24:29] National Battlefield, which was a viewshed study that did as part of that project.
[24:38] This slide summarizes several of the practices we found localized should implement to minimize
[24:42] community impacts from data center development.
[24:44] For example, it's important to classify data centers as an industrial user, whatever your equivalent
[24:50] is make sure your zoning maps have been revised to prevent bi-right projects next to residential
[24:55] or other sensitive areas. Make sure you have the minimum requirements in place like setbacks
[25:01] and building heights that are appropriate to keep development from intruding on residential or
[25:05] commercial neighbors. That would also include, I think, one of the things in your old mandate about
[25:09] the lighting. Again, these are secure facilities, so that lighting is on all the time. Another
[25:14] reason why you don't want that right next to the neighborhood. You can also consider a static building
[25:20] standards and landscaping requirements for facilities that are highly visible to the public.
[25:25] Again, the idea there is that if you build this data center in an industrial area,
[25:28] you don't really care that much what it looks like, but if you've got something next commercial
[25:33] or that people are going to be driving by every day or close to residential area,
[25:37] then the appearance of that facility comes a lot more important.
[25:43] It's also important to deal with
[25:44] that noise issue up front, which I think I already mentioned. So if a data center is going to be close
[25:49] residential or other sensitive areas, you should probably require sound modeling studies to see
[25:53] what the noise impacts might be. There's a lot of variables that go into, if and how a data
[25:59] center can be heard and from where, so one is going to be that data centers designed, especially
[26:03] the cooling system that it's using, that's what generates the noise and then there are topography
[26:08] considerations. If that's in a valley and you have a residential area in a hill, that sounds
[26:13] going to go straight up the hill to the residential area. So sound studies can help get a handle on
[26:18] If a house sound might get me generated from that facility and how traveled in the surrounding area, you can also revise your noise, your ordinances to make noise violations and force both through zoning requirements.
[26:31] So instead of having that as a sort of a civil or criminal violation, where there's a penalty, a civil fine, if you have an zoning, you can give it a lot more enforcement teeth because now you're in violation of how you should be doing business.
[26:47] And the product is an area that's close to a historical site, it's probably a good idea
[26:51] to at least do a phase one study to figure out if there's anything there that should be preserved
[26:57] as well as if there's a historical site nearby.
[27:00] What would it, you know, how's that, how's that development going to impinge on the viewshet
[27:05] from that historical site?
[27:07] Mark, if we're moving on and quick question in regards to accountability for violations
[27:16] You mentioned, of course, you know, you know, fines aren't, especially when you're working
[27:24] with an industry where there's these deep pockets, what, what, what have local governments
[27:37] or even a state done in the way of creating real teeth for accountability?
[27:47] And for noise in particular?
[27:49] Well, since that's what we're talking about.
[27:52] Here's sure, but for any type of zoning or conditional use of violation.
[28:03] As far as the noise in particular, I think one of the issues we found was that nobody had any really strong enforcement mechanisms so that that was one of the things they were looking for was a way to better bring facilities into compliance with
[28:20] uh, uh, uh, noise standards, or not necessarily compliance with noise standards, but
[28:25] eliminate the conflict that they were having with the neighbors. Um, and so the, I think, uh,
[28:32] the suggestion is by putting that into your zoning, you can then have actual teeth in terms of,
[28:38] well, your evaluation of of how you should be operating as a business. So now things like
[28:43] your business operating permit are on the table, which is obviously a pretty nuclear option, but it
[28:48] one with a lot of teeth as opposed to basically having no leverage at all. I'm just saying
[28:54] I wish you guys would stop making this noise. Here's $150 fine but I don't even know who I'm supposed
[29:00] to find. I'm supposed to find the guy who's working at the desk or do I send this to Amazon or
[29:07] or Google headquarters. Because I think that's from the noise perspective. That's why the more you
[29:18] to see license. All those things can be used as ways to bring that facility into compliance
[29:25] in theory. Again, we haven't seen it played out in fact. In terms of other things, we just rely
[29:30] on a regulatory regime. So, for example, during construction, there's a problem with stormwater
[29:38] management. And then that just goes to a regular regulatory regime for stormwater management.
[29:48] Thank you.
[29:48] Can I quick ask a question, please?
[29:51] About two slides ago.
[29:54] It was a word I'd never seen before, so I'd appreciate.
[29:57] You tell me what it means, and I'll sign the second.
[30:00] So it was like examples. And if you could explain roof tiering and then the word was fenestration.
[30:08] Roof tiering is why I've kind of made up.
[30:11] So that's just sort of, instead of having a big monolith, you know, like a big giant box, you can have a tiered roof.
[30:18] You can require that in zoning. I'm not sure if that's exact term that is owning personalities.
[30:23] I tried to look that up to figure out what their exact term was. I knew we heard it somewhere.
[30:27] Anyway, so instead of having, you know, here's the curb, here's a parking lot, here's a 60 foot edifice that could be broken up by having a tiered roof, so it's 30 feet, then it goes back a little bit and then it goes up to the full 60 feet.
[30:41] Does that help with noise? Also when the hill example you gave the parking space?
[30:45] And it's really just aesthetic, I think, and I have to defer to a sound expert, but for the
[30:53] most part, all the equipment is on the roof, so I might push it back a little further, but
[30:58] I don't think it would really have a big effect on noise.
[31:01] Finestration isn't actual word, that's what they call, so these things, they're like distribution
[31:07] centers, so there's no windows, it's just all they need is a big concrete wall, but that
[31:13] looks ugly. So some places they've required finistration, fake windows, basically.
[31:20] Are you kidding me? It looks more like an office building.
[31:24] No, my goodness. Thank you.
[31:27] Cause, keep in mind, in northern Virginia, so a lot of these are industrial areas, but a lot of them are
[31:30] right by roads, and that was one of the things that that's come up in a couple of places.
[31:36] I've heard a roof tiering called stepbacks. That might, that's what I should have put on the slide.
[31:47] Next, so that was mostly the items that were
[31:50] pertent to your agenda today.
[31:52] So I'm happy to talk about those more.
[31:53] If you guys want to cover us on the other stuff
[31:56] from our study, which really is some of your other areas.
[31:58] Can I ask one more question?
[32:00] Yeah, it's a lot.
[32:02] So I'm a town elected.
[32:05] And this has been a whole different context,
[32:07] but we had a noise pre-development complaint
[32:11] or wanted to have a noise a sound study.
[32:14] It was very hard to find somebody.
[32:15] We finally did, but they operated in one to two states over.
[32:20] So, you know, it was kind of expensive.
[32:22] The good news is we had insurance, but did that very plentiful?
[32:27] I mean, who is qualified to really do proper modeling and proper both before and during and after
[32:34] if you have issues?
[32:37] I mean, it'd be great to have like a place that not recommending them, but just saying
[32:42] And here's a list of people who are qualified because it's very hard to find them.
[32:46] It was kind of connected through a university thing, word of mouth is how I found it.
[32:51] But it's just curious.
[32:53] Yeah, that's a good point.
[32:54] I think that's come up in Virginia as well.
[32:56] In some cases where they've had a require these in the past, the company had found somebody.
[33:02] I mean, of course, the residents don't trust what the company's house study said.
[33:08] But in some other cases,
[33:12] there's not a big industry out there for that, so yeah, I couldn't really give you any good insights into how to find folks.
[33:21] I know one of these neighborhoods where they had a lot of problems.
[33:24] They had a former NASA slash submarine engineer and he designed his own site.
[33:32] So if you have one of those, that could help.
[33:35] but yeah, I really can't help you out much there.
[33:39] Unfortunately, no, this is Laura Skender.
[33:42] My husband was an acoustical engineer.
[33:44] He worked in Washington, DC, and did a lot of ship silencing
[33:47] and submarine silencing as well.
[33:49] And that firm could help to people on the house
[33:51] and the hill too to figure out that residence.
[33:55] So they are out there.
[33:56] Is he still working?
[33:58] Yeah, but something completely different.
[34:00] But the firm is still around if you want,
[34:01] I can give it to you.
[34:02] Yeah, I'm curious.
[34:06] Mark, I'd like to ask a question.
[34:09] So, the most egregious sort of noise offenders, I guess, were in the range.
[34:15] You said, like, 40 to 60 decibels like a low drone, was that right?
[34:21] That sounds right.
[34:22] Yeah.
[34:22] I think, yeah.
[34:24] Yeah.
[34:25] I'm sure there's some sort of fundamental difference.
[34:28] But I know Texas has become kind of the hotspot for crypto mining.
[34:33] And a lot of the same aspects go into those facilities when we're talking about cooling and, you know, trying to minimize noise.
[34:42] I know some of the most egregious examples of crypto mines are up to 80 and the even 100 decibels.
[34:50] And they're having trouble. I'm not sure how sustained those 100 decibels are, but there are consistent violations on some of those facilities up to like 85 decibels sustained for hours on end.
[35:03] and I'm just wondering what is the difference between the data center and the crypto mine and what's
[35:11] what's kind of hold. And I don't expect you to know this mark. I'm just kind of thinking out loud.
[35:16] Actually, if I could just, yeah, the ATC sub station in the town of Middleton on Highway 14,
[35:23] I don't know from a technical point of view, but go there. I mean, you can't, there's no trespassing
[35:29] size as you can drive in a little way. I mean, there's actually a study from the UK where you
[35:34] can run a drill without having to be plugged in under one of those. That's what's, and you
[35:40] can kind of see it on really damp days and you can shear it. And it's this very, it's almost
[35:46] like a buzz. I mean, it's, it's, you know, it's intermittent, but it's in so many seconds that
[35:52] it's like it's constant. Yeah. And so I don't know, that's why it'd be good to find some
[35:57] objective true experts to be able to. That would be a great session. Yeah. Find such people
[36:05] because they're rare and hard to find. You responded surprising that Texas is dealing with such
[36:13] noise issues with crypto minds and there's such a drastic difference in the kind of complaints
[36:20] and levels that are that you're seeing and in Virginia versus what they're seeing in Texas.
[36:25] And again, I know there's, there are different types of facilities, but you would think it's somewhat similar.
[36:32] They're basically doing the same thing. My guess would be that.
[36:37] So what did it, the, the mountain noise, you, it most of that, if not all that's coming from the cooling system.
[36:45] And depending on, and so depending on the cooling system you're using, you may be generating almost no noise to a fair amount of noise.
[36:53] The second part is how cheaply constructed is your facility.
[36:58] And so these data center facilities are at least building them for, you know, to last 50 years.
[37:03] And so there's, there are expensive facilities, they're putting some good investment into them.
[37:07] I would guess the quality of the mechanical equipment is going to be higher than you would
[37:13] find in a crypto mining facility, because there are a lot more fly by night.
[37:17] They don't necessarily care if there's going to be there in 50 years.
[37:21] They might only be building them for, you know, five years or 10 years.
[37:25] So my guess would be the quality of construction of the quality of equipment.
[37:30] Would probably account for that difference.
[37:33] That's that even within the data center industry you can see.
[37:36] There's some facilities built by some companies which tend to be the ones that have the noise problem.
[37:43] And facilities built by other companies which tend to not be a problem at all.
[37:47] And I'll go a lot of that's designed, but also something that is is cost and equipment.
[37:55] So I'll leave that there.
[37:57] If you want to ask me, I could tell you too.
[38:00] Yeah.
[38:01] Can I just ask one of the quick question on this one article that was sent to us as background on noise?
[38:07] There's a quote near from an Amazon spokesperson and they're talking about redesigning their products to significantly reduce sound or noise levels.
[38:16] received that properties of joining data center facilities. How could one get more information on what's a, you know, a good sound noise mitigating design versus, I mean, you mentioned the quality of construction and how long its life expectancy is expected.
[38:31] But do you have any insights on that?
[38:34] I think that would not get pretty technical pretty quick.
[38:38] One is you could look at, you know, it's just model off in other facility and then see if there's a facility that, that it, you know, that it scores up to.
[38:48] But yeah, I think, because like even if we're talking to Amazon, so Amazon was the one that was a problem in where the family was sleeping in the basement by the Great Oaks neighborhood.
[39:00] They were trying to figure out why it was so loud and how to reduce the sound and one
[39:05] thing they did was they realized that they're, I guess the intake fans are probably getting
[39:09] that wrong were two narrows and they made the bigger and that reduced the amount of noise
[39:15] coming off that facility. There's also something to do with, you know, they didn't have some sort
[39:21] venting that was then creating some sort of weird noise
[39:26] reverberations.
[39:28] So it's a really technical area out.
[39:31] I don't know if there's any,
[39:32] I don't know how you address that excited from getting
[39:34] some of who is an expert in a sound study to try and look
[39:38] at the design, compare to other designs.
[39:40] But as a lay person, I don't know if there's
[39:42] much that I could do to try and figure that out.
[39:46] I'm looking at the University of Virginia,
[39:47] or some other esteemed and UW-Madison with their new computer center to try to get a program
[39:54] to train people. That would be a great idea. That would be a recommendation possibility.
[40:01] Yeah, and somebody, it would, because of the variance and data center construction, it would
[40:09] have to figure out like each data center is going to, unless they're at some of some companies
[40:15] replicate them. A lot of them are, it's a unique design. So, you know, one day the center, what you would
[40:22] expect from it might be different from the next one just because the different equipment, different building
[40:26] height, where things topography is located. So, yeah, there will be a lot of variables to consider.
[40:35] You know, we had one day the center here, we had emailed you about it. It was in an agricultural area.
[40:46] I'm
[40:46] probably more urban, any impacts on noise to livestock or wildlife, particularly because I think
[40:54] we would see that more here, where that's where you get more bang for your buck for larger data
[41:00] centers to be placed in a agricultural area. A lot of the development that's happening in Virginia
[41:06] now is pushing into agricultural areas. So even down in Canada, which is days in a couple of the world,
[41:13] that kind of is historically half developed
[41:16] and then half being preserved for as rural.
[41:19] Prince William County, kind of similar,
[41:22] and so you're seeing a lot of that pushing in.
[41:24] In terms of actual effects on other than elimination
[41:29] of agricultural land, which is a pretty clear effect,
[41:33] it's not really clear, or we can find a lot of information
[41:36] on how that might affect, for example, livestock.
[41:39] We heard some anecdotal concerns,
[41:41] but nothing, nothing concrete, but yeah, that's definitely pushing on the rural areas in Virginia as well.
[41:47] I don't think we have quite as much a livestock as you have in the Wisconsin, but we do, you know, we do have
[41:51] cow farms and I guess we don't have a sheep farm, I think. Yeah, I don't think I have a good answer for it, I don't know.
[42:01] So Mark, I'm curious from an environmental standpoint, you've mentioned the lighting around the data centers.
[42:08] Is there an issue then for migratory birds with all of that light at night?
[42:15] Did you look at that?
[42:16] That's not something we looked at.
[42:18] Again, there's another there's a number of potential impacts on the surrounding environment.
[42:23] Those lights tend to be pointed down.
[42:27] So I don't know if and how that would affect migratory animals.
[42:32] Sort of bigger picture on that environmental piece.
[42:34] Because most of these are, you know, pushing out into more agricultural areas, this is in northern Virginia, which is a pretty heavily populated area.
[42:43] It's not like we're building these in, again, talking with wildlife experts and environmental experts.
[42:51] We're building these in southwest Virginia or in the Appalachian Mountains where we have our biggest, most diverse ecosystems.
[42:58] that would have a much bigger ecological impact on the environment than it would
[43:03] building these in, you know, basically an expanding suburban area.
[43:12] I should say ecological
[43:12] impact. But that, even in that case, there are having cases where, you know, projects that
[43:17] where they found some ecological niche, you know, that we're relying on existing environmental
[43:23] protections to keep that stuff from being destroyed. So I think in this case it was a
[43:29] So, I can't remember what a Westman was a rare, very rarely occurring top ecological feature involving
[43:36] mosses and salamanders that needed to be protected.
[43:42] So even in a suburban area, there's still environmental concerns.
[43:49] Mark, you talked about the development on the stark battlefields and that data centers
[43:56] may not be appropriate, wouldn't it be true that maybe not any type of development would be
[44:01] not appropriate on those battlefield sites?
[44:05] I think that that's certainly an argument. Again, I have my own personal views as an
[44:09] organizational or agnostic on that, but this particular site or particular area, at one point
[44:14] they tried to develop a, I think it was going to be an location for Disneyland and that got shot down
[44:20] then there was going to be a Walmart Supercenter and that got shot down and then there was
[44:25] centers where the latest attempt to develop next that battlefield. And that court case was decided
[44:31] pretty recently, so it looks like it got shut down again. Okay. Then my other question is in regards
[44:38] to noise when you look at, if you're living near an airport, obviously there's a lot of noise,
[44:43] the issue from what I'm hearing is the constant hum of the noise, not necessarily the
[44:48] decibel level. Is that true? Yes, that's that's what it is because again it's not allowed
[44:54] to damage hearing, but it's a constant thing and it's because it's a little
[45:01] If you're a student from a speaker with playing bass, it's a kind of sound that if it's hitting just right, you can you can feel it. And that's where the example of this house where, you know, this this homeowner, again, this is the, this is the worst case scenario that we heard in a tire study.
[45:20] But essentially he couldn't stand the homes, so he changed his windows thinking that the new windows would do a better job of keeping out the sound, but then the windows the new windows started vibrating from the sound that actually got worse.
[45:39] But again, not all, I just want to, I just want to be clear that not all data centers create sound.
[45:45] There are definitely some problem children that we came across, but that doesn't mean that all of them were were in noise problems.
[45:54] But all data centers do emit a number of different wireless frequencies. Is that right?
[46:02] That I'm not aware of. We did not get into that at all?
[46:06] No.
[46:08] Really?
[46:14] I just want to check with Supervisor Marin.
[46:16] I, you know, since I can't see you on the screen,
[46:18] I do have a chat window up, but I just want to check if you had any questions.
[46:23] I'm all good. Thanks for checking, though.
[46:25] Sure.
[46:28] So I have another maybe a couple of questions first.
[46:34] Should I just turn off the presentation?
[46:37] Sure.
[46:38] If that's your last slide, that'd be.
[46:41] Well, I have additional ones on some of your normal stuff.
[46:44] No, no, no.
[46:45] No, leave it up then.
[46:46] Leave it up.
[46:48] So, you know, you had some consideration
[46:53] and in one of the other slides of some design standards and so forth to address aesthetics and whatnot,
[47:03] but also it sounds like there could be some construction standards employed to mitigate sound and so forth.
[47:14] just curious if localities in Virginia or the state has done any work in creating standards
[47:28] that are codified in building code. We do have a uniformed state building code,
[47:39] so
[47:40] that but the in terms of putting special requirements on data centers that's really up more of a
[47:47] local issue because the building code is just you know what's safe and structurally sound.
[47:53] In terms of you know we want something to be less uptrusive or for aesthetic reasons
[48:01] that then is something that comes down to what the locality wants and and how they're making
[48:07] there, there are plan project plan approvals, facility approvals, development approvals.
[48:14] And so you have seen, I said a couple of these to, to Majeed. There's a couple of counties
[48:20] in northern Virginia, especially that have put in sort of the data center ordinances that have a lot of
[48:37] kind of things. So that's a lighting. I think it was one of the things that came up. So those
[48:43] specific things would be at the county level. And again, Marjita, as in a couple of weeks that
[48:48] would be helpful for that. Okay. And another thing that could contribute to noise among other things
[49:01] would be, some of that's intermittent, but it'd be the exercise of the generators for these
[49:07] facilities. Is there any, what, what consideration was that in your report?
[49:15] Yeah, so there's really three things when it comes to noise. One is what we already talked about,
[49:19] which is the operational thing. Operational noise that one is by far the worst, or it has caused
[49:26] the most conflict, I should say, because that is something, as you said, it's constant.
[49:34] The generator testing, again, if you're living near a facility, there's a lot of complaints
[49:40] from folks nearby about the generator testing that I wanted to do, and don't want to
[49:45] do in the morning, don't want to do in the evening.
[49:48] So that has definitely been a source spot for some residents, but again, we didn't see that
[49:53] had seemed to have the same level of concern as the operation went back because that is intermittent
[50:00] it does go off. I live in a neighborhood where on any given day somebody's moving along or running
[50:06] leapblowers kind of the same as somebody testing a generator and at least in terms of the sound
[50:10] it it makes. The last one is construction noise. So a day of center is a big project takes a year
[50:19] have to build one and that's pretty darn fast, but you know, it's still a long time.
[50:23] It really dates our campus. What happens when they finish building that first day of
[50:27] center? They move next door to the next day of the center. So some of these construction sites,
[50:33] you could have, or some of these campuses, you could have construction, and I'm going for years.
[50:37] And obviously that does create a lot of noise, and that's another source of noise concern to neighbors.
[50:43] And another reason why not why you should not have these right next to a neighborhood.
[50:50] Anybody else with any questions around community impacts?
[50:57] Actually, I have one.
[50:58] Yeah, go ahead.
[50:59] So Mark, with the diesel backup generators,
[51:02] is there, are there, have you looked at any pollution standards?
[51:06] I know I've read a lot about people saying
[51:08] they gave in smell of fumes in their homes.
[51:12] Well, that is my very next topic here.
[51:14] We're going to need to jump onto that.
[51:18] Well, actually before you delve into that too much detail,
[51:23] I guess it's sort of a transition to that, but my next question in that is, one of the things
[51:29] will be subjectaries we'll be looking at is, I guess, the impact or demands for emergency management.
[51:42] And so I'm curious about what your insights are in particular, again, looking at community
[51:53] impacts what types of, you know, unique materials might be used both in construction and operation
[52:04] of a facility that might pose risks, say, if there's a fire that maybe normal fires oppression
[52:18] materials would not work so. I don't think we got into that in specific. I know that there
[52:26] is a concern that for example the main thing in here that would be different from another building
[52:33] would be the volume of computers. So you have all the rare earth minerals, all the stuff that goes
[52:43] in the computing chips, all that is in there.
[52:47] So if there was a big fire, what's that emitting?
[52:51] I guess it'll be the, to speak the angle you're talking about.
[52:54] I think the concern that we had heard more
[52:56] was about what happens to those materials
[52:58] because the computer server life spans
[53:00] about three to five years.
[53:01] So what happens to those materials after?
[53:04] So we looked into that a little bit
[53:05] basically they get recycled.
[53:09] So but yeah, I don't know about in terms
[53:12] there's a fire, there's extra risks. Again, these are, you know, these facilities have to meet
[53:17] whatever they were fire standards are, and so they should have pre-effective fire suppression
[53:20] systems and understand those don't always work right. But, you know, if you're spending
[53:24] billion dollars in a facility, hopefully you're putting it at these in fire suppression system.
[53:30] Something I hadn't considered was the useful life of equipment. That's a good thing to think about.
[53:37] So thanks. We do have an error report and appendix on that. It's been a while since I looked at that one, but we have a one or one half page or on what happens to that equipment, what else is most of.
[53:50] Okay, well, we all have a link to that so thank you.
[53:54] And then.
[53:56] Like I just said, we have a link to that report, but I think it'd be helpful if you could forward to us to staff a PDF of your slides too.
[54:10] Sure. Thanks.
[54:11] I know, probably over my time. I'm happy to give you this.
[54:14] Let's keep going.
[54:15] Okay.
[54:16] I think there's interest, yes.
[54:18] So, yeah, we appreciate your time.
[54:21] So, absolutely.
[54:23] Go ahead.
[54:25] So, today, the backup generator point, so data center backup generators do emit pollutants,
[54:32] but their use is minimal and existing regulations can help curb those adverse impacts.
[54:39] So, to that first point, data centers rely on a large number of diesel generators for backup power.
[54:44] They have enormous energy use, they need to stay online 24-7, and so they have these diesel backup generators, which is the industry standard.
[54:51] These are, you know, these are the, the backup generator I have in my shed, these are tractor trailer size, industrial generators, there's easily north of 50 of these for one data center.
[55:03] Those generators, because they're burning diesel, they emit several harmful pollutants that contribute to the development of small and can have other health effects.
[55:10] for that reason in Virginia, they're regulated by using state and national air quality standards,
[55:16] so in Virginia, all these generators must be permitted by the Department of Environmental
[55:20] Quality. The permits set limits on their use and the amount of emissions they can emit over the
[55:25] course of a year and then the emissions are monitored for compliance. In practice, the backup generators
[55:32] are rarely run for prolonged periods. I think it's where you talk about the mostly run for maintenance
[55:37] testing that happens about once a month for 10 to 30 minutes per generator. So if you have
[55:42] 50, you can imagine that you're probably running one or two of those every day for maintenance testing.
[55:48] So if you're a neighbor, that might not be something that you want to hear. They do run for longer
[55:53] periods during outages. So if the power goes out, those things are all going to click on, they're
[55:57] going to run until the power comes back on or they hit their emissions limit. However, data center
[56:03] companies report the experience few outages. Again, this is because, you know, whereas my house,
[56:09] your house, the building you're in is served by the overhead distribution lines. That's not
[56:14] the case for a data center. They are served by a transmission line, which goes to a substation
[56:18] that substations powering that data center. So any outages are usually due to a localized
[56:23] problem, so the a squirrel in a transformer at the substations, one of the examples we heard or somebody
[56:30] It does a fact-fingers move and cuts the cable that's running from the transmission or from
[56:36] the substation to the days center, and the power goes out, and then the generators come
[56:40] up.
[56:43] So, don't mean to, so for those reasons, because they're not really being used that much,
[56:49] we found they're unlikely to harm the regional air quality in northern Virginia, which
[56:52] is a non-attainment zone, which means there are some air quality concerns in that area.
[56:56] They're basically a small part of the emissions over all emissions and overall emissions in the region are going down.
[57:03] All that said, there are two real risks to consider here.
[57:08] One is the worst case for long large-scale regional power outage, so we get a hurricane and a transmission line
[57:13] is getting knocked out. And that event, all those backup generators are going to kick on.
[57:18] And if that outage lasts long enough, those emissions are going to contribute to air quality concerns.
[57:22] So there's going to be a lot of diesel fuel going out all at once for a couple of days maybe.
[57:30] Fortunately those algae are rare and once the event is over the air quality would return to normal.
[57:36] The other risk which I think one of you all pointed out a minute ago, excuse me,
[57:42] is that there could be localized concerns, so if you're in a neighborhood in data center
[57:51] alley and loud county, you are in the vicinity of a hundred facilities, each one
[57:58] has 50 backup generators and they're testing two or three of those every day, that's
[58:04] a many diesel fuels, fumes that on a regional scale are not that significant, but that might
[58:08] have localized effects. Because again, there are, it's particular matter, and there that can be harmful nitrate nitricks.
[58:16] I'm going to get all the chemicals wrong, so they're that, I guess.
[58:21] But there's some, some form of nitrogen, particularly matter, and then a third one that I can't remember off top of my head.
[58:27] Anyways, that's, that's not things that you want to be breathing. So that's the concern that if, if you're living close that area,
[58:33] you're how much of that is going up and going away that you never breathe versus how much of that is
[58:38] lingering at the service level. And so we would have had some recommendations to look at this
[58:44] for our department of environmental quality has already launched the study to monitor those
[58:49] emissions and see what the potential health risk might be. Mark, let's go quick question.
[58:55] Yeah. Two different things. I've been reading about how there's a Elon Musk and Memphis
[59:03] Tennessee big problems with the diesel generators. The people can't breathe and nobody's
[59:10] going to help them do anything from these data center turbines. I'm reading about in Texas,
[59:15] how to sidestep all these types of good intention regulations. They're just doing everything
[59:20] private on a piece of property just wondering if you have any knowledge of anything like that
[59:25] in Virginia and give us some flavor around it. And then just thinking out loud, what happens
[59:30] to the property values of any homes that happen to be in your data center? I can't imagine it
[59:35] helps the value. On the first question you're talking about on site generation like gas turbines,
[59:44] is that kind of thing?
[59:45] Yeah, the diesel gas turbines.
[59:47] So there's like, it's private and they're not using
[59:49] all the things you're talking about.
[59:53] So it would be gas not.
[59:56] So these are diesel generators
[59:57] since like the diesel that you put in a truck.
[1:00:00] Which is different from a gas turbine or behind the meter generation, which would typically be a gas turbine?
[1:00:06] Are you talking natural gas? Natural gas generators.
[1:00:08] Yeah, no. They were saying it was a diesel fumes, but I don't know. It was just news reports.
[1:00:14] Well, we know just in terms of the backup generators, diesel is sort of the standard.
[1:00:20] There is this big move to behind the meter generation.
[1:00:23] We had not seen much of that in Virginia at the time of our study, there's only one data center.
[1:00:29] We found that had on-site behind the meter generations that had a gas turbine on-site.
[1:00:34] And that was because there used to be a factory or something there and was left over from that.
[1:00:41] So, and some have been proposed in Virginia, but if you think it's hard to get a data center,
[1:00:47] try building a data center with a gas turbine attached to it.
[1:00:52] At least of Virginia, that's a local use decision, so we haven't seen a lot about happening.
[1:01:02] What was your other question, I'm sorry.
[1:01:03] About property value impacts.
[1:01:05] It's been an easy study for residents near a data center.
[1:01:09] We did look at that.
[1:01:11] It's really, and I've done a lot of studies trying to look at the effects of things on property values.
[1:01:16] Figure out the effects on property values is very hard to do.
[1:01:19] So for example, all the neighborhoods around all these data centers in Northern Virginia, the
[1:01:23] property values have been going out.
[1:01:26] And that doesn't mean that having a data center nearby isn't somehow affecting a value of
[1:01:31] a particular property.
[1:01:34] But if you're, if you're on the, if you're backing up to the data center, I'm pretty sure
[1:01:39] that property value is effective from being backing up directly to the data center.
[1:01:42] But if you're in that same neighborhood, but you're two streets over, there might be zero
[1:01:46] effect from that, especially if it's not a data center, it doesn't make any noise.
[1:01:51] So we were not really able to tweak out how property values are being affected because
[1:01:58] generally they're going up.
[1:02:00] There is also this strange effect where property values as a whole are going up because of data
[1:02:06] centers because a data center company is willing to spend a lot of money to go where they
[1:02:13] want to go. And so you're seeing pretty insane with a pricing per acre around places where
[1:02:24] data centers want to go. We're talking millions of dollars in acre to buy plots to build these
[1:02:30] facilities on. And the effect there then is that one is it has priced out other development.
[1:02:38] So, if I want to develop single-family homes hard enough to do that already in Virginia,
[1:02:46] but now that you want to high density residential development, I might be able to,
[1:02:50] I might be priced out of that because the data center company is on that land or that big tract.
[1:02:57] That's one thing we've heard concerned from residential developers.
[1:03:02] The other interesting thing is you're seeing that property owners will be like,
[1:03:07] Well, I don't want to live next this data center.
[1:03:10] Hey, everybody in the neighborhood, why don't we all get together and we can sell all of our land
[1:03:14] to a data center developer and they'll pay us out the news for it and then we can move somewhere else.
[1:03:20] And we've seen several instances where that has happened where groups of
[1:03:24] neighborhoods, resident groups of residents have sort of banded together and agreed to sell
[1:03:30] their property to a data center developer.
[1:03:33] Wow. Thank you. You just mentioned real quick a data center that makes no noise. Is there really one of those?
[1:03:41] And would that be, could you give us an example? Maybe not now or in the future that maybe would be a good case study,
[1:03:47] that we could see how it is they're making no noise.
[1:03:50] Well, the exception of generator units and there are things you can do.
[1:03:54] So, but yeah, a lot of it, I would say I can't remember how many, but
[1:03:59] And all the ones we visited, and we purposely visited ones that were problems, I would say most
[1:04:06] of the ones we visited, you can stand outside and you don't hear anything.
[1:04:09] You don't hear the cooling systems.
[1:04:13] And a lot of times, some of the ones that we were told were problems, we showed up, and
[1:04:17] I've got a well, it's not making noise today, because the cooling system wasn't running.
[1:04:23] So it really depends on the cooling system, so some can make noise all the time, some only
[1:04:28] noise when the cooling systems running and at full capacity, it doesn't matter what day you
[1:04:36] show up, it's just not, you're not going to hear anything from the cooling system. Now you may
[1:04:40] spill here when you're testing the generators, but there's other things that you can actually
[1:04:44] do to help reduce that noise, like requiring generators to be housed in sort of generator
[1:04:51] sheds, so there's other things that you could do from a zoning standpoint to try and address
[1:04:56] that generator noise. Thank you.
[1:05:05] What are you? So we found that this underwater uses currently
[1:05:08] sustainable and the state helps ensure sustainability through its regulations. So again,
[1:05:14] we talked a lot about cooling systems and noise, cooling system also has a major impact on
[1:05:20] how much water a facility uses. So you either have two flavors, you have your water dependent
[1:05:26] cooling processes, basically, evaporative cooling.
[1:05:29] We're using water to facilitate cooling.
[1:05:34] The great thing about these things
[1:05:35] is they're more energy efficient.
[1:05:37] The bad thing about them is that they obviously
[1:05:39] use more water.
[1:05:41] The alternative is a dry cooling system.
[1:05:44] It uses more electricity, but doesn't use any water.
[1:05:47] So, for example, your residential air conditioning system,
[1:05:49] that's going to be a dry cooling system.
[1:05:51] There's no water being consumed by that system.
[1:05:56] Those data centers, at least in Virginia, get their water through our water utility, we've collected data from those utilities and found that well some days that are used a lot of water.
[1:06:05] Most use about the same amount or less than an average large office building.
[1:06:10] So in terms of how that scales up as an industry as you can see that on this graphic here.
[1:06:15] So, for example, those ones at the bottom, those are ones that are, you know, it's a data center,
[1:06:20] but because it's using a dry cooling system, the only water being consumed is in the bathroom
[1:06:25] or if there's a kitchen in there for the staff, that's the only water that the facility is using.
[1:06:31] So that is, in some cases, about the same as an average home.
[1:06:35] Other cases, up to the top, they're using the evaporative cooling system, it's probably a large facility,
[1:06:40] and that's consuming a fair amount of water, but only a few cases was more than what you'd expect
[1:06:45] from a large office building.
[1:06:50] Again, in terms of how that scales up, so there's the facility, it's self-question, but
[1:06:54] then you have to scale that up to the industry use as a whole.
[1:06:57] We found that data center customers accounted for between 0.2% up to 21% of a water utilities
[1:07:06] total consumer use, and so those, that's a pretty big range, and 21% obviously a lot
[1:07:13] the utility, use 0.2% very little bit. Those were largely a function of how many data centers
[1:07:20] that utility served and how big the utility was. I'm not 100% sure but I think that 21% was
[1:07:27] there's a really large campus in a very small county and I think that's why that was so big
[1:07:34] but it could also be loud in county. I can't remember. But again, to put this further into perspective,
[1:07:41] of the industry as a whole uses less than half a percent of the state's total what it
[1:07:46] withdrawals. So yes, the facility does use water, but it's not on the scale of say something
[1:07:52] like paper mills in state, which are enormous water user.
[1:08:00] All that said, the data in the
[1:08:01] industry is growing so we can expect that water used to increase over time. And that raises concerns
[1:08:06] about sustainability, especially in Virginia. We're still in a drought right now. We had a winner
[1:08:13] out and to a spring drought hopefully will be a summer drought.
[1:08:17] So those changing weather patterns kind of compound those sustainability concern.
[1:08:22] We found that at least in Virginia our state regulation appears to sufficiently address
[1:08:27] that.
[1:08:28] It doesn't mean it's not concerned, it just means that we do have good regulatory processes
[1:08:32] in place.
[1:08:33] Our state regulator regulates water withdrawals and requires permits for large-scale
[1:08:37] draws, it models the impact of any changing withdrawals on water availability and wildlife
[1:08:44] that happens when the permits issued and also happens when the permits renewed.
[1:08:48] So if utility gets some big gas in our customer, they have to go get a permit to add more water.
[1:08:53] That's all going to be reviewed. And before that permits approved, they can't approve something that's
[1:08:59] not sustainable. You might know, utility might have to go build something like a reservoir or something
[1:09:04] to help make sure that they have the resources that are needed to have negative effects
[1:09:09] and that the water source can be sustained.
[1:09:12] All that said.
[1:09:13] Go ahead.
[1:09:15] And why don't you interrupt you?
[1:09:17] All of that said.
[1:09:18] Virginia is a pretty water rich state, even with those drought concerns.
[1:09:24] But that doesn't mean that all localities have equal access to water.
[1:09:29] So, a locality that doesn't have a lot of localities are located by rivers.
[1:09:33] I'm here in Richmond, we've got a river that runs right through town, but a lot of localities
[1:09:37] don't have a big water source like that.
[1:09:40] And so on a locality scale, a locality can be water restricted.
[1:09:45] And so anything that locality proves for a data center development might then be taking away from
[1:09:51] their ability to do some other type of development, especially in certain parts of the state
[1:09:55] where, you know, my county doesn't have a river and we're relying on groundwater or we're
[1:10:00] on piping water in from the neighboring locality. Those are all restrictions and localities
[1:10:06] really need to understand how much water is this project going to use. How does that affect
[1:10:11] our ability to meet future-resinced demand or to pursue other development opportunities,
[1:10:16] like I know, for a manufacturing center or something that might create more of a job.
[1:10:22] Okay, that's my question now.
[1:10:26] Just quick, here the Wisconsin farmers union in its current policy statement, you know, context
[1:10:33] much larger and not specific to data centers, but just they're recommending a groundwater
[1:10:38] mapping project that supports the development of a mapping tool that will lead to better
[1:10:43] understanding protection and utilization of groundwater and drinking water supplies in
[1:10:48] Wisconsin has Virginia done anything like that or do you how do you know what your water
[1:10:54] supply is other than there's a river running through the area kind of thing.
[1:11:01] We do we have can pushing the energy to my knowledge but we basically have I think three
[1:11:06] of we have we have one major aquifer which is on the coast and that has been a ground that's been the main water source for a lot
[1:11:14] of our coastal cities. And also is I mean water source for a couple of large industrial legacy
[1:11:20] industrial users. Basically that aquifer takes a long time to replenish and we've been
[1:11:27] sucking that aquifer dry. So it's been something that's well known policy issue and concern
[1:11:32] for several years. We have parts of our state that are sinking just as sea level is rising
[1:11:37] because of of taking water out of that aquifer. So that aquifer is very well understood and has
[1:11:43] researched a lot, just because it's so well used by urban areas and because the effects that we're
[1:11:51] seeing are draining that aquifer. The other aquifers are much smaller and regional, so we have
[1:11:57] this topography, carcesterography, where you might have a pocket here, a pocket aquifer here,
[1:12:04] another pocket one over there, but it's really only the eastern part of this data has that big aquifer
[1:12:11] that is sort of a giant thing that's sort of affects many areas. I don't know if that means for
[1:12:19] your question. No, I appreciate that. Thank you.
[1:12:28] It's the last finding here,
[1:12:29] data center impacts, work quality, native habitat and agricultural lands. Those impacts are
[1:12:34] pretty similar to other large-scale developments. So development of produce and undeveloped areas,
[1:12:40] as you all know, that creates environmental risks, new development can destroy natural habitat.
[1:12:44] can reduce the amount of land that's available for agriculture. So depending on how
[1:12:48] development it's done, it can also have other environmental repercussions such as
[1:12:52] increased flooding, erosion, or water pollution. These are all environmental risks that apply
[1:12:57] to any greenfield development. So that one could data centers just as would include any large
[1:13:03] commercial industrial development or even a high-density residential development. And a lot of the
[1:13:10] where we have data centers being built.
[1:13:12] We also see a lot of high residential,
[1:13:14] high-density residential development,
[1:13:16] like townhouses, picture of massive townhouses with
[1:13:20] big parking lots.
[1:13:22] In all these cases you have clearing,
[1:13:24] grading,
[1:13:25] an installation of a pervious surfaces like rooftops
[1:13:28] and parking lots that basically forever
[1:13:30] changes that like end.
[1:13:33] So our research found that at least on the
[1:13:35] the water side that the state and federal
[1:13:38] regulations for new construction largely mitigate the risks to things like stormwater run off at wetland impacts on the landside again
[1:13:47] You're you're for ever changing that land through development whether that's data center or some other are high-dense to be used
[1:13:53] And the loss of those lands is sort of trade off when you do development of all types
[1:13:58] So again, does apply for data centers, especially seeing as in some parts of the state that is the main development pressure
[1:14:04] But also, it's not unique today, so there's this kind of a thing that happens when you develop wind.
[1:14:11] And in Virginia, again, I'm not sure I will scan some words, but our localities have traditionally had, and it's still do have all the authority over land use decisions.
[1:14:25] I have a couple more slides on the economic tax and energy stuff, but I realize I will over my time here.
[1:14:36] Well, we got about 10 minutes.
[1:14:39] So, I can go on on your insights while we have you here, so go ahead.
[1:14:51] All right.
[1:14:53] So, on the economics side, in the tax side, in Virginia, the industry has had positive effect.
[1:15:00] On both economics and local tax revenues. So economically, mostly benefits occurred during construction. So again, these are massive construction projects. And they can have up to 1500 workers on site during peak construction. And those jobs are in well-playing blue-collar jobs, construction trades, so electricians, pipe fitters, steel workers. After a facility is built, a few jobs remain. So again, most of the impact is happening during construction, because that's when the most people are employed to work on that.
[1:15:30] on that site. We asked me that for a hyper scale facility, after it gets built, there's around
[1:15:35] 50 people per facility. Those jobs are pretty easily split along sort of your IT professionals
[1:15:41] and building management professionals, which is what you might expect. But there are also a lot
[1:15:45] still a lot of contract trade workers. So there's miles of pipes and electrical wiring in these buildings.
[1:15:52] So there's basically a full-time team of electricians and the pipe fitters who are working
[1:15:57] facilities after they've built. And of course I also have contract security which is probably the
[1:16:02] lowest wage of those different jobs that are created. Data centers can be a cash cow for
[1:16:09] a locality and that's because they can generate revenues but they're not sending a lot of kids
[1:16:15] to schools, they're not aside from the construction period, they're not putting a lot of cars on the
[1:16:21] So it's more or more on say a net positive when it comes to the revenue that really depends on the where
[1:16:27] locality decides to tax it. So in Virginia we've seen some of our localities get a really kind of a
[1:16:35] enormous amount of money from these facilities we're talking 10 to 30 percent of their budgets come from
[1:16:40] data center revenues. In other cases we've seen localities are trying to attract data centers and they're
[1:16:44] cutting their local tax rates. Well if you do that then you don't really get the tax revenue benefits
[1:16:51] You're just getting whatever the amount of jobs that it's going to create create
[1:16:56] an energy-side.
[1:16:58] The internet growth is driving an enormous increase in energy demand.
[1:17:01] This is more pronounced than Virginia than anywhere else, but you can really see this ticking up pretty much across the board.
[1:17:07] To look at this graphic, those top lines are showing the very top lines showing our unconstrained demand projection.
[1:17:15] The other, the solid line, kind of about halfway down that showing, you know, what if only half of that unconstrained demand shows up?
[1:17:21] and then the bottom line is showing what if there was no new data center demand.
[1:17:26] The takeaway here is that if unconstrained demand happens and we can meet it,
[1:17:30] we're talking about tripling the amount of energy that's consumed in the state.
[1:17:34] Only half that shows up, we're still doubling the amount of energy that's consumed in the state.
[1:17:39] By 2040, if we had no new data center demand, demand would not be going up or be fairly
[1:17:45] going up at all. So most of this this demand growth is coming from that industry.
[1:17:50] I want to reiterate this is what could happen because, you know, one that demand might not show up.
[1:17:56] It may be AI is a bit of a bubble and a lot of what is projected doesn't happen.
[1:18:02] To even if that demand is there, you till these still have to be able to meet it, right?
[1:18:09] To that effect, meeting this unconstrained demand would be also moonshot scenario in terms of building enough generation enough transmission
[1:18:18] and importing enough energy to keep up with demand.
[1:18:23] So that scenario one, the unconstrained demand scenario
[1:18:26] through that really highest line, that's going to be really, really hard
[1:18:32] if not possible to do.
[1:18:35] Maybe more realistically, that half of unconstrained demand,
[1:18:38] maybe we're able to meet half of that.
[1:18:40] We're still talking enormous investments in generation,
[1:18:43] transmission, and ported energy.
[1:18:44] So even that, keeping up with the man to go under that half-demand scenario is going to be really
[1:18:50] hard to do for our state. This obviously has some cost implications that the cost come from a
[1:18:58] couple of areas. One is the big one is obvious is that you're building a lot of new stuff.
[1:19:04] You're building a lot of new generation. You're building a lot of new transmission. So historically
[1:19:09] adding a new customer would actually reduce every way's rates but in how we're talking about building
[1:19:15] a bunch of stuff for a small group of customers, which would not otherwise be built.
[1:19:21] That then creates more costs in the system that would not otherwise have occurred.
[1:19:25] And so everybody's piece of a cost-by could get a little bigger unless we change how those costs
[1:19:31] get allocated out. The other challenges and this will be built I think pretty much everywhere is that
[1:19:37] If you have a supply demand problem, if demand is higher than supply, what happens, prices
[1:19:43] go up, and that's about effect all energy users, and that's something that's really hard
[1:19:49] to deal with those costs, and only assign those to people who are increasing costs on
[1:19:54] the system, because they're, you know, how can you say, you'd have to price things differently,
[1:20:00] It just gets really complicated.
[1:20:05] And yeah, so I covered the energy piece a lot quicker than normally do, but how we take
[1:20:11] questions on that or anything else that you guys have, here's my contact information as well,
[1:20:16] and I'll send you guys a copy of the slides after we're off.
[1:20:21] Mark, I have one question. If I may, this is kind of a two part question.
[1:20:25] From our experience around here, we've seen data centers come in with proposals that are
[1:20:30] hundreds of thousands of acres, more than what they really need to have.
[1:20:36] Have you seen in Virginia where they're using any potential excess land for other types
[1:20:41] of development?
[1:20:43] And the second part of that question is, Oregon had you seen just the opposite where maybe
[1:20:47] it's stifling economic development where companies don't want to be next to a data center.
[1:20:55] So on to that second one first, I think that we've seen the effect on other development
[1:21:00] And it's really a competition for land.
[1:21:05] It's hard to outbid the deep pockets that a lot of these data center developers have.
[1:21:11] And so what we've seen is the competition for land.
[1:21:14] And again, this isn't something we've been able to prove empirically, but anecdotally what
[1:21:18] we've heard is that competition for land because I'm a developer, maybe I'm not going
[1:21:22] to build a commercial mixed use development or a residential development because I can't
[1:21:27] to pay however much that is per acre that is being offered by the data center industry.
[1:21:35] And on the first question, and I think that was, you know, an access land,
[1:21:42] we haven't really heard that we have heard about data centers putting in for more energy
[1:21:48] than they actually need or more water than they actually end up planning, end up using.
[1:21:56] So it maybe makes sense then I think that would be putting it from more land.
[1:22:01] One thing that I would question about that is that they might just be trying to buy all
[1:22:06] the land for a campus and they're only going to plan to develop one part of it at a time.
[1:22:11] So maybe I want to eventually say I'm a developer and I want to have a 2,000 acre
[1:22:18] data center campus. I might buy all that land, but then my first development, I'm only developing,
[1:22:24] you know, a quarter of it, so I've got three buildings going. And I might not have my development
[1:22:29] plans for the rest of it yet. So that would be my guess is they're more likely planning to build
[1:22:36] stuff in the future, but they just haven't put it forward. But again, I'm just speculating at this point.
[1:22:41] Okay, thank you.
[1:22:45] Yeah, they're questions.
[1:22:47] Something.
[1:22:48] Have you had any of these go bankrupt or be abandoned?
[1:22:51] Or not least as fast as they think they will.
[1:22:55] And so they have financial issues.
[1:22:57] And if so, how have you dealt with that?
[1:23:01] We have not seen that yet.
[1:23:03] I think I heard anecdotally that happened with a crypto
[1:23:06] day center somewhere in crypto. We don't have crypto in Virginia. I think those are a little more
[1:23:14] fly by night. I think that's a very valid concern. A lot of the industry is being bankrolled by
[1:23:21] the big four, so even though it's data center company, data center developer X, they may already
[1:23:28] have a contract and hand to build that facility for one of the big four. The financial
[1:23:36] back and should be there, but, you know, you never know, things are going to play out, maybe
[1:23:39] some of these AI bets don't play out, and some of the data centers that need it, right
[1:23:43] now, the demand though, far outstrips the supply, so if a data center failed, somebody would
[1:23:51] pick it up immediately, because if they have a contract to get, if they have a function facility
[1:23:56] that has power, some of it will snap it up right away, that's really been the main constraint
[1:24:02] on the industry is access to power, I think in Virginia especially but maybe other places as well.
[1:24:09] In Northern Virginia where you have this high concentration, we've had projects
[1:24:13] I've been waiting years to get full power because there's not enough transmission capacity to
[1:24:18] serve them. So that has been the main thing. It's more demand has exceeded supply. So that has
[1:24:25] you know, an empty facility has a minute of concern yet.
[1:24:29] Are you asking the data centers to help pay for the power that there, that was your question, go ahead.
[1:24:38] You take it.
[1:24:40] Is the data center covering the full cost of the increased cost for production for the energy,
[1:24:44] or are they passing on to the other consumers as well?
[1:24:49] The answer is right now those costs are being passed on to other consumers in a way.
[1:24:55] So, on the, there's three parts of the system, distribution, which is the local level, there's transmission and a new generation.
[1:25:04] On the distribution costs, at least in Virginia, they are fully covering those costs because they're, you know, there's a substation and then the either the, the electric co-op or big investor own utility or big monopoly is basically making pay for the cost of that because they're building it.
[1:25:20] It's still the utilities property, but they're willing to serve that facility, so they're paying the full cost of the substation.
[1:25:27] However, on the transmission side, that is something that's originally set past their costs.
[1:25:34] Generation, depending on your utility, is either being bought from a market or it's being paid for by the monopoly utility.
[1:25:43] And those costs have historically the way they are allocated gets spread across everyone.
[1:25:53] Right?
[1:25:53] So if you're building new generation or transmission, that costs that get spread across everyone.
[1:25:58] The issue here is in our building new generation and transmission that will not otherwise be built for these small number of customers.
[1:26:05] And so if those costs continue to be spread across everyone, all of our residential and commercial bills are all going to go up.
[1:26:11] So, one of the things we'd recommend in our report that our utilities and our utility regulator do is look at one create a new customer class that you can park these high load facilities in to look at how costs for transmission and generation are being allocated and make sure that that customer class is paying for the cost that is occurring.
[1:26:39] So both of those things are currently underway.
[1:26:42] Our utilities are mainly utility created a customer class last year, and in their next
[1:26:47] rate case, they are supposed to be looking at how to allocate costs to that customer class,
[1:26:53] to avoid having those go to residential customers.
[1:26:58] And on our electric co-ops, they're basically just passing through.
[1:27:01] So they're pretty well insulated their customers.
[1:27:03] because we have, for the big, for pay utility,
[1:27:06] and then we have these little electrical co-ops,
[1:27:07] which used to be real co-ops,
[1:27:09] but are now serving data centers
[1:27:11] and are actually pretty big themselves.
[1:27:15] The one thing that was that prices,
[1:27:17] as long as supply exceeds demand,
[1:27:18] you're going to have those price increases,
[1:27:20] there's really no way to insulate against that effectively.
[1:27:25] Hey, anyone else, questions?
[1:27:30] All right, Mark, we really appreciate you.
[1:27:35] taking time to walk us through through a sort of a summary of
[1:27:43] a real report and insights on this really big topic, so thank you very
[1:27:51] much. Thank you if anybody has a vacation property in Bayfield that you're
[1:27:56] You're not cheap for a night, let me know.
[1:28:01] Enjoy.
[1:28:02] It's beautiful up there.
[1:28:04] I myself haven't paddled around the dining and kayaking
[1:28:08] around the island, but I've camped out there.
[1:28:11] It's great.
[1:28:12] So I'll bring your foul weather gear.
[1:28:16] I'm going to tell the radio club net.
[1:28:18] Yeah, that's the next.
[1:28:20] Yeah.
[1:28:21] OK, I'll appreciate it.
[1:28:23] OK, thank you.
[1:28:23] Take care, Mario.
[1:28:24] Thank you.
[1:28:25] Thank you.
[1:28:26] All right, well, we're running right up against our time, and I'm actually due to another meeting.
[1:28:36] So I'd like to ask everybody if there's items from this presentation, questions that are raised recommendations,
[1:28:48] or suggested those would be things that, if you would maybe email staff, your thoughts or just itemize list of things that we can add to the list of follow-up items, appreciate that.
[1:29:04] So any conversation, discussion is a result of presentation.
[1:29:13] We want to run through very quickly, otherwise we can adjourn and save it for another week.
[1:29:20] But I just have one.
[1:29:21] I want to go back to your opening statement about proposing a moratorium in Dane County.
[1:29:28] Obviously, the city's and villages association will want to see what that resolution looks like.
[1:29:33] see if it impacts our members or not, I understand that you're reasoning for what I always worry
[1:29:41] about the unintended consequences that it may stifle economic development within Dain County itself,
[1:29:49] if companies looking to locate here are hearing about a moratorium, a fully different issue.
[1:29:56] So just keep that in mind if you would.
[1:30:00] As you're drafting the resolution that be very specific about what it's for.
[1:30:05] For sure.
[1:30:07] We'll have that as well as the length of time.
[1:30:13] Maybe hold on some of the opportunity.
[1:30:17] As far as the impact to cities and villages, this will only impact, we'll be limited to towns.
[1:30:28] I'm in corporate areas that are subject to county zoning.
[1:30:33] This morning, if I don't green counties,
[1:30:36] considering it's similar or torturing.
[1:30:39] So, spreading.
[1:30:42] All right.
[1:30:44] All right.
[1:30:45] Then if there's another discussion,
[1:30:50] again, email those thoughts for anybody.
[1:30:54] So moved.
[1:30:55] with my password.
[1:30:56] Second registered.
[1:30:57] So I can buy Richardson on favor of adjournment say aye.
[1:31:00] Aye.
[1:31:01] Those, for adjourn, thank you all.