Catalyst with Shayle Kann - The state of play of data center development

Episode Date: June 12, 2025

The future of the grid increasingly hinges on where and how data centers get built. To forecast the kind of power infrastructure we need to meet AI’s growing appetite, we first need to understand a laundry list of variables: data center size, workload type, latency, reliability — even the variety of a data center’s coolant system.  So what’s the state of play in data center development today — and how are the trends shaping grid needs? In this episode, Shayle talks to Chris Sharp, chief technology officer of Digital Realty, a developer, owner and operator of data centers. They cover topics like: How AI inference workloads are clustering in existing regions, driven by latency and throughput requirements “Data gravity” and “data oceans”: how large concentrations of data attract more compute infrastructure What’s driving longer lead times: interconnection delays, equipment bottlenecks, or both? Large-scale builds vs. incremental additions and densification of existing infrastructure “Braggawatts” vs. real demand: separating hype from reality The diverging power needs of training vs. inference, and whether any workloads work with intermittent power The evolving role of “bridge power” and why diesel and gas are still in the mix Resources: Latitude Media: Google’s new data center model signals a massive market shift Latitude Media: The future of energy-first data centers takes shape Latitude Media: Can a new coalition turn data centers into grid assets? Latitude Media: Do microgrids make sense for data centers?  The New York Times: Wall St. Is All In on A.I. Data Centers. But Are They the Next Bubble? Catalyst: The case for colocating data centers and generation Credits: Hosted by Shayle Kann. Produced and edited by Daniel Woldorff. Original music and engineering by Sean Marquand. Stephen Lacey is executive editor. Catalyst is brought to you by Anza, a platform enabling solar and storage developers and buyers to save time, reduce risk, and increase profits in their equipment selection process. Anza gives clients access to pricing, technical, and risk data plus tools that they’ve never had access to before. Learn more at go.anzarenewables.com/latitude. Catalyst is brought to you by EnergyHub. EnergyHub helps utilities build next-generation virtual power plants that unlock reliable flexibility at every level of the grid. See how EnergyHub helps unlock the power of flexibility at scale, and deliver more value through cross-DER dispatch with their leading Edge DERMS platform, by visiting energyhub.com.

Latitude Media covering the new frontiers of the energy transition. I'm Shell Khan, and this is Catalyst. There is a lot of noise. One of the things we've been joking about is a lot of braggawatts. Oh, I have a gigawatt. I have a gigawatt. Coming up, what's more insatiable? Power demand from data centers or my appetite to talk about it? When utilities need flexible capacity they can count on, they turn to Energy Hub.

Energy Hub works with more than 170 utilities, coordinating over 2.5 million devices to manage 3.4 gigawatts of flexibility built for the moments when utilities can't afford uncertainty. Energy Hub builds and operates virtual power plants that utilities actually stake their grid planning on, coordinating EVs, batteries, thermostats, and more through a single platform built for utility scale. Predictive, verifiable, and designed to perform when it counts. Learn more at energy hub.com. Trillions of dollars are flowing into clean and critical infrastructure, but those investments aren't driven by technology alone. They're shaped by markets, by policy, by capital, and by the institutions that connect them. I'm Alfred Johnson, CEO of Crux, and host of a brand

new podcast, Critical Capital. Each episode, I talk with people deploying capital, shaping policy and building the clean economy. Tune in as we unpack how progress is actually made. Listen to critical capital on Spotify, Apple, or wherever you get your podcasts. Catalyst is supported by Fish Tank PR, an award-winning PR firm focused on climate and energy tech, renewables, and sustainability. Fish Tank is known for generating prominent and effective media coverage for the brands they work with. If you want a PR partner that's thoughtful, shoots straight, and gets results, you'll like Fish Tank PR. To learn more about Fish Tank's approach, visit fish tankpr.com.

That's F-I-S-C-H-Fish-Tankpr.com. I'm Shail. I invest in early stage companies at energy impact partners. Welcome. So we've, of course, spent a lot of time on this podcast talking about the energy data center nexus. Too much time? Who's to say? Objectively, it's the biggest thing happening right now. So buzz off haters. Anyway, one thing we haven't done amidst all of that discussion is talking to somebody who's actually building data centers and has been for a long time, for that matter. So that seems dumb. Fortunately, there's Chris Sharp. Chris is the CTO of Digital Realty,

which has been around for 20 years, developing, owning, and operating, co-located data centers all over the world. Chris is just very insightful about what's going on in the space and what's coming next. So I brought him on to talk about what the hell's happening in data center world

and a fair bit about what's happening at the data center power Nexus. Here's Chris. Chris, welcome. Thank you. Thanks for having me. Looking forward to it. Excited to talk about the state of data centers

in the world, I think particularly in the United States. Let's restrain ourselves to something reasonable to talk about because there's a lot to talk about even here. You've been in Data Center World, how long now? Like 15 years or more? Yeah, 15 plus years. Longer than I could to admit, believe me, when I started,

it wasn't cool and it wasn't at the forefront of every headline. So lots change. Congrats on finally being cool. Thank you. I wouldn't go that far, but okay. Yeah, sure. I want to start by talking about geography. a little bit. Again, maybe we'll focus primarily on the U.S. I mean, my perception of how this

world has evolved is that historically, the data center development activity and operation activity was very concentrated in a pretty small number of regions, northern Virginia being the one everybody probably knows the most about, but then there were like some other tier two regions behind that. And that part of what has happened in this new wave of excitement and AI and hyperscale data center is getting planned everywhere, is that there's been a big geographic dispersion. And so I guess one question for you is, is that true or is it still really regions

that drive the majority of the growth? Yeah, so I think you have to take a step back and look at the problem from two lenses, right? The first lenses is, what are the workloads come into market? And I think that lens is interesting, right, where the most simplistic terminology, people hear about training and inference, I think training has forced a broader regional deployment,

but that's for training these kind of frontier models. I would say that there's been a lot of growth in that, but we see that kind of leveling out, where we really see the consumption of AI or inference, that's driving that regional-specific growth going forward. And I think that's where it's more embedded in a lot of the existing, if you will, follow the clouds with availability zones.

That's where it's really starting to be that investment growing and evolving quite quickly. And I think you brought it up with Northern Virginia, that Nova market has been one of the critical availability zones, which is now represented as a critical kind of AI growth sector going forward as well. When you say availability zone, what is the promise? What's the availability promise that's being made? Because this is what's driving its regions for this reason, right?

100%. It is a promise of a certain level of availability. Yeah. And so I think, you know, I always go one step further on that training inference, but it's monetization, those availability zones were foundational and set up gravity, if you will, of what was driving that as SLAs and consumption to the enterprise, right? And I think that's where you see a lot of these capability and infrastructure being invested in these zones all around the globe, where there's a major city center.

It's usually closer to the CBD because there's proximate requirements with throughput and latency associated with it. But those availability zones are what has built that kind of, if you will, first wave of cloud infrastructure coming to market. And, you know, availability zones are slowly evolved. They're not everywhere. They're not in these tier two, tier three markets. But we're seeing a lot of AI applications being embedded inside of those availability zones. And the last piece I'd leave you with is that AI is an and and not an ore to cloud.

Right. I want people to go really comfortable with that is that a lot of these AI capabilities. capabilities are being embedded in the cloud services you're consuming today, like co-pilot in some of the early capabilities coming to market. But that's how we really see a lot of these markets maturing over time. You mentioned latency there. So my layperson's understanding here of what you were describing is, okay, training a model, you can kind of do anywhere, but the models are getting bigger and bigger and bigger.

And so we need bigger and bigger data centers, but they're not as geographically constrained, and thus you can put a training-focused data center maybe in the middle of nowhere, assuming you have all the other things that you need, you have power, you have labor, you have water, et cetera, et cetera. But then inference, latency matters more, and thus you want high, not only latency, but I guess availability as well

because these are time-sensitive requests. And so that's why you want to be clustered in a region and so on. I've heard some people, this will get to the energy data center nexus a little bit, speculating that you could bifurcate even the inference workloads into things that are latency sensitive and things that are not. And the ones that are not, maybe you go take advantage of cheap, clean power, maybe even intermittent power out in the middle of nowhere, right?

Which definitely exists, but is not where the rest of the data centers are getting cited. Do you view that as a viable approach given the actual workloads? It is, it is. And, you know, it's great that we're going through the workload. right? Like that workload is what depicts the infrastructure required to making it successful. And I think latency and throughput remain many different things. And so I always try to double click on a little bit. The amount of throughput required is what's challenging, right? And latency, as long as it's consistent for a lot of the workloads we see, they can operate fine. But it's that throughput.

The amount of data that's required for delivering kind of an inference type of solution is something that is, again, proximate, not only to the consumer, but profit. to an ecosystem. So I'll hit on your second point where, yeah, we see a lot of text-to-text scenarios where that workload can be deployed in two or three markets throughout North America and service the entire market. And so that's a very simplistic kind of scenario where we're in the early innings of AI and the complexities hasn't really come to fruition for the broader market. But as you see bimodal, some of these more advanced reasoning models where a token is just generated against a prompt and then you consume it and it's done, a token may be generated

inside of an AI world and go through multiple models to ensure that it's not hallucinating or that it has a mixture of experts or a depth expertise in that outcome of that token. So that's where these ecosystems of other AI infrastructure being proximate to itself, not only just the data, but I would be remiss not to hit that AI is only as smart as the data sets, you feed it. And so those training, you were able to feed that monolithic set of data in the middle of nowhere. But now we're more real-time, micro-learning, inference. It's starting to become more approximate to where the data oceans and that data gravity, which we've produced a report a long time ago, is happening around these availability zones and these epicenters of

these tier one markets. So it sounds like what you're saying, if I'm interpreting it right, is that this notion of, let's go where this, just all else equal. It's. just go where the stranded power is, it probably has some validity because there are some workloads, text-to-text, for example, as you said, that can handle that, where the throughput requirements are not so high and the latency requirements are not so high. But it also sounds like you're saying the direction of travel is in the opposite direction, because actually the workloads are becoming more sophisticated, leveraging multiple models, and the throughput requirements are getting higher. And so that set of opportunities, just go wherever the cheap, the available power is,

probably dries up, or at least the relative share of, like, how much you can build in that use case versus how much you could build if you actually have a cluster and it's all regional and it's near all the other models. Like, that's going to be a much bigger opportunity. Do I have that about right? Yeah, no, you're spot on. And there's a confluence of events that are happening there. It's not just power. The expense to stand this infrastructure up is these chips are not cheap, right? And so being able to utilize that over a longer horizon of workload and driving that utilization is also a form factor of if I haven't installed for training and training can be

very a spiky workload that I can embed inference in that capability to get higher utilization out of that investment because the Roik is real on this. And so you see a lot driving that direction as well. But no, as we see these higher value kind of actually. aggregators, if you will, of multiple models, right? Multiple capabilities. That's really starting to become more proximate to one another. And again, data is everything to a lot of these environments, be it hypers, or be it enterprise, which we focus on both, having the ability to embed algorithms or this accelerated compute infrastructure in close proximity to their existing

data oceans or constant data creation models is everything. to our customers. Okay, so assuming that the majority of the growth will continue to occur in regions, maybe not all in today's Tier 1 region, but that it's still going to be sort of a regionally driven market. I guess the question is, how quickly do we tap out these regions? From a power perspective in particular, because unless you tell me otherwise, I think that tends to be the thing that maxes out first, unless you tell me maybe it's labor.

But like, let's take, we talked about Nova, right? Northern Virginia. how close to tapped out are we there? How much more can we possibly build in that region? Yeah, so you bring up great points, right? Where tapping out is the right word, where in a lot of these markets, power has been tapped out. I mean, it is a phenomenal market.

I mean, some of the most recent stats, it has 0.5% vacancy rate, which is phenomenal. I mean, it's a multi-gigawatt market. I think, you know, one of the things that differentiates how we view these markets is coming in and master planning, not only with like the entitlements and making sure you have access and the rights to the land, but that master planning arc is sometimes five plus years. And a critical element of that is working with a utility operator so that they know that, you know,

when we say we're going to need a gigawatt like with what we're building right now, right next to the Dulles Airport, that they have an understanding of that power requirement. And in a lot of the cases, they're able to meet that. But in certain cases, particularly in northern Virginia, which has been wildly, you know, publicized, is that some of the, not necessarily generation, but the distribution of the grid has been challenged. And so we're always working with different solutions to overcome those shortcomings in the short term,

but then ultimately working with that utility operator so that they get an understanding of the future growth associated with these markets. Because, again, this infrastructure, and by saying this, this AI kind of secondary wave to cloud, wants to be proximate to existing. infrastructure. So being able to tie those things together and the power has been challenging, right? And I think it's challenging not only throughout the globe, but in a lot of these markets where you need to be

working with utility operators, which is why I love talking to the market and educating not only the in consumer around what's happening in AI and the workload, but ultimately the broader infrastructure like the utility operators and some of the other technology coming to market in solving for that power constraint. Yeah, you mentioned timelines. I wanted to ask you about that. So it's obviously location-specific, but can you talk to me, particularly relative to your history in this sector, from, I don't know, from the beginning of development of a new site to operations of that site, what does that timeline look like? What's the range of timelines that that looks like today? And how does that compare to history?

Yeah. So it is a challenging scenario where all things being equal, It takes about two years from concept to delivery to build out what I would say versatile data center. And by versatility, I mean comprehensive portfolio of solving for the hyperscaler needs, but also solving for the enterprise customers. So that's a 24-month window. But with the backdrop that we're experiencing, particularly with the power and the grid and just the overall equipment bottlenecks, I mean, utilities are requesting an aggressive kind of four-year rant projection that when you start to take that power down, you need to utilize. at which we've been very good stewards in a lot of these markets, that when we do that master planning, we project that we are going to need 500 megawatts, we take down that 500 megawatts and operate that over a longer period of time. But some of these other interconnects are definitely,

no two markets are alike, but they're elongating even beyond the 24 months that it would take us to pull that together. So there's a lot of challenges there. And I referenced that at a high level, but some of the broader infrastructure constraints are transformer lead times are 50 plus weeks right now. Well, I was going to ask you about that. Right, because transformer lead times and switchgear and stuff like that, that has been a challenge in all sorts of areas of the power sector. But I wonder whether because you have the added constraint of really long interconnection lead times, does it just mean that the interconnection is the long pole in the tent? And so you sort of, you have enough time that the transformer thing doesn't actually, or switch gear or whatever, doesn't actually delay projects because you happen to have another thing that takes even longer, or is it its own constraint?

Yeah, there's two high levels. It is a constraint. Don't get me wrong. There's two high level elements that we've been doing at digital for, I've been here 10 years, companies have around 20 years, is vendor managed inventory. So not only understanding, hey, here's our portfolio in a single market, but really operating at a point where we're buying that, which gear buying that infrastructure ahead of time where we can alleviate some of the bottlenecks. But yeah, that secondary constraint, and this is why I referenced earlier, the master planning, showing and signaling to the utility operators and being a good steward of having, you know, top tier customers and credit worthy customers in our portfolio, which want to operate with us 10 plus years in that asset, balancing that together is everything. And so that interconnect from the utility has become constrained. and some markets were always investigating different types of solutions,

via gas turbines, and even those are backlogged plus 29, right? Like, that's an extensive background as well. Yeah, we're definitely going to talk about bridge power because that is super interesting. Before we do, though, one of the question I have for you about sort of how the market has developed is about scale of data centers. And we sort of alluded to this when we talked about, okay, the training models need really big scale data centers. But, you know, over history, right?

Like, you guys probably were developing 20 megawatt data centers 10 years ago, right? And now it's hundreds of megawatts or you mentioned a gigawatt. What does the demand picture look like for you? Does everybody want the biggest possible data center that you can build them? Or, like, what is the nuance to that? Yeah, it's a good piece to dig into, right? Where there's a lot of noise. Like, one of the things we've been joking about is a lot of braggawatts.

Oh, I have a gigawatt. I have a gigawatt. And there's just so much noise out there that you really want to get underneath the workload and the durability of the company behind the workload. And that's where, you know, being a publicly traded operator, we're constantly watching that. And not everybody needs a 100 megawatt data hall. And there are certain use cases where a contiguous set of GPU infrastructure, which requires a very discreet capability, which we have some of the strongest heritage of engineering

talent within digital that have been solving this for the clouds. And now it has grown, but they want to contiguous 100 megawatt GPU array. So it's not just about the total capacity block, but then it's the densification of that capacity within the asset. And so we're always watching that. But what we're really seeing is inference can come in in like five-ish megawatt blocks and you can solve for it a bit differently. Now the densification is still there. And then the private AI piece is there's hotspots where it can be, you know, a couple of megawatts as well, but they want to be embedded in their existing portfolio of assets and balancing those two things is something we're always eyes wide open.

Virtual power plants are becoming a reliable way for utilities to manage capacity, but enrolling devices is just the start. What really matters is confidence, knowing those resources will perform when dispatched and being able to prove it, from the control room to the living room. Energy Hub's platform handles the full picture, from near real-time forecasting, locational dispatch, and the kind of rigorous verification that holds up when regulators,

grid operators, or leadership ask, did it deliver? Easy enrollment creates momentum, proven performance builds trust. That's why more than 170 utilities rely on Energy Hub to manage over 2.5 million devices delivering 3.4 gigawatts

of flexible capacity. See what that looks like at Energyhub.com. We're living through a profound economic shift. and energy sits at the center of all of it. Trillions of dollars are flowing into power plants, transmission lines, battery factories, data centers, but the future of energy isn't shaped by technology alone.

It's shaped by markets, by policy, by capital, and by the institutions that connect them. I'm Alfred Johnson, CEO of Crux, the capital platform for the clean economy. Join me for my brand new show, Critical Capital, as I talk with people deploying capital, shaping policy and building projects. Together, we unpack how risk is priced, how incentives are structured, and how progress is actually made.

Listen to Critical Capital on Spotify, Apple, or wherever you get your podcasts. Are you tired of overpaying for big-name PR firms, but not really knowing what they're delivering? Is your comms team wasting time reviewing lengthy messaging briefs and decks, instead of engaging journalists or producing content? Are you wondering why your competitors are getting press and you aren't? Fish Tank PR is an award-winning climate and energy tech, renewables, and sustainability. focus PR firm dedicated to elevating the work of both early stage and established companies. Whether you need to position yourself as a thought leader in between project announcements or translate complex ideas and technologies into tangible, compelling stories that resonate

with the media, Fish Tank Can Help. Check out fish tankpr.com. That's F-I-S-C-H-Fish-Tankpr.com. So this goes to an interesting question I've been wondering about, right, which is mostly what you hear about these days is... Right. 100 plus megawatt data centers getting built. Mostly hundreds of megawatts, if not gigawatts, right? And that's what all the news is about. But if

the individual inference workloads need to be, they can be 5 megawatt chunks, albeit you want some degree of densification, could you employ a strategy where you go build 120 megawatt data centers all in a region? Is that a viable approach? Because from a power perspective,

my suspicion is that in the regions with a lot of data center activity already, the scales might soon tip where it could actually be easier to build 120 megawatt things than a single two-gigawatt thing or a single two-one gigawatt things. Yeah. One of the most challenging things represented by AI is the ambiguity in the workload. There may be one workload to like, yep, okay, I can take five or 50, one megawatt and I don't care where they are. That's less than, that that's the outlier of what we're seeing is that to operationalize, they would really like it to be more of a contiguous scenario where they do logically think

about them as five megawatt chunks and they want a bit of resiliency. You had said it earlier, reliability becomes increasingly important with inference because that's the consumption of that capability. What we're seeing is having a 100 megawatt hall where a bunch of five megawatt deployments could come in and be represented as inference, that has a higher viability for a lot of the hypers plugging in, you know, a multitude of capabilities. Because it's not one AI workload. It could be a bit of, you know, I always am challenged and not to reference a specific workload that I'm working on with a specific customer, but just some of the stuff spoken about publicly, but you

look at like some of the most recent announcements from Google in like Gemini and the, in V-O-3, those are probably compromise, like coming to market as a composition of multiple, you know, inference capabilities coming to market to meet that customer demand. And you have to solve for the peak, right? Like, people always forget, like, we learn this in the web scale, hyperscale. It's like the grid. It's the exact situation as the grid, right? We build the grid for the peak. We also build data centers. You have to. You can't not deliver. Absolutely, absolutely. You mentioned the Braggawatt thing. I mean, that's the other thing that feels to me like we're

clearly in a moment, like two things can be true at the same time. There can be explosive actual demand growth for compute, leading to actual need for lots of gigawatts of new data centers. And also, it can be true that the volume of data centers, quote, in development, and certainly the volume of load interconnection requests going to utilities, is like an order of magnitude more than is actually going to happen. Like, both of those things can be true. But I want to under the degree to which that presents a challenge for folks like you, because you're, you know, you need to get stuff built. But on the other side of the table from you is a utility who's inundated with load interconnection requests and needs to figure out which things are real

and which things are not. And I imagine that sort of comes up to works a little bit. Yeah. No, you're spot on, right? And I view that as three elements, right? Where, you know, the power, the amount of power and even the amount of financing required to meet these up or in projections, it doesn't exist, right? And so you can't solve at all even if you wanted to, but then double-clicking on aligning to your customer, right? And not all customers are equal and really understanding what their goals are and what they're trying to achieve. That's the heritage of digital reality, right? And that's where we've been doing that in pretty much every theater on Earth over multiple cycles, right? So AI represents a new cycle in a new wave that's

bigger and faster than we've ever seen before. But it takes partnerships. to really pull that off correctly. And I think, you know, I couldn't say it better in that, you know, some of the works that we've been doing together collectively and also with utility operators, having a communication with them to show that, you know, they won't overbuild unless they have a level of comfort that you'll take and utilize that infrastructure they brought to market.

So that ramp is everything to them. Working with that customer to show them that we have the right customers, we have the understanding to support the workload is everything, because there's going to be some probably written about very big challenges and failures where they wanted a total capacity block, but they couldn't support the densification. Or they were just building for the spike. It was very spiky, but the longer term utilization is much lower than anybody had projected. That will have very negative impacts on them to operate in a longer term horizon.

So we're always focused on right types of customers, right types of partnerships to meet that peak load demand. and the finances are required to hit it. Okay, so you mentioned bridge power. I want to hear how that is playing out in the market. We hear a little bit about it, right? There are folks who are saying, I mean, the famously, the GROC data center employed this substantially, where you just say like, okay, the grid interconnection,

the timeline is too long. And so I'm going to throw a bunch of generators on site and operate off of the generators as a bridge until the grid comes along for me. How common is that, actually? Yeah, I think there's some outliers. Very few of it gets covered in the press because nobody wants to really go on to the market where the grid can't meet the customer ramp demands today. I mean, full stop. And I think, you know, one of the things that we're always looking at, and I keep harping on this, is that customer ramp requirements are an absolute key driver. But bridge power is one tool among many that we're always looking at, right? And, you know, natural gas, which is what you were referencing earlier, I think it has a solution. a shorter term horizon, but we're always looking at what are the longer term power generation capabilities that potentially coming online and working with the utility operators on if it is

grid constrained and they couldn't get the resiliency in the grid or if it's a generation challenge. We're always looking at how do you hit that peak demand with some of the batteries and some of the other technology that we've been seeing come to market. But yeah, it is an outlier. I think a lot of the utility operators are starting to understand that, hey, this is real demand and I'll align to it, they're not chasing the noise because nobody wants to invest in a bubble, right? Like I'll go on the record of saying that. We're always looking at to ensure that we're not aligned to a bubble and that long-term durable workload is there.

But, yeah, we too investigate Nat gas turbines. We investigate all options within the grid to overcome some of those shortcomings so that we can service our customers because I think it's often missed. You know, I talk to the utility operators that if our data center, you know, goes dark, and is dormant, it doesn't allow our customers to grow and hit that next capability they need represented as AI or not or even cloud services, they have to be able to get revenue out of that very expensive infrastructure going forward. So they're looking for that long-term master planned alignment to the utility operators. My sense is that there's kind of two different things you can do.

If you think about bringing assets, beyond the assets you would normally bring, right? You're always going to have backup power or whatever. But if you're going to think about bringing anything beyond that alongside behind the meter at a data center, you can either do the pure bridge power thing, which is we will supply our own generation and operate the data center off grid or partially off grid until the interconnection arrives. Or there's this other thing which is, okay, we will proactively strike a deal with the utility wherein we will bring our own generation or we'll bring our own batteries or whatever it might be and we'll have an interruptible tariff or something like that. And in so doing, we will get faster time to power, but it's a negotiated sort of deal

with the utility as opposed to a bridge to the utility. By sense it that the latter version is more common, more prevalent than the former. Is that right? Absolutely. Yeah, absolutely. And it's because, like, we don't want to be all things to everyone because we won't be good at anything, right? Where you want to invest in the utility to allow them to do what they were good at and get over this challenge of the spike in demand. But that longer term, you know, environment should be with the utilities, and that's why we form long-term relationships with the utility operator and act as a good customer to them on behalf of our customers. It's that chain of value that we're always watching because, you know, doing behind the meter, you know, becoming a power generation

capability, I think, is a short-term gap. Nobody would be investigating that if we didn't have the constraint. And so that in itself tells you that a lot of individuals want to stay to their core stitching of, hey, what are we good at? What is our core capabilities we're bringing to market? But I think there's a short term, let's meet the demand, and the longer term, who would be better at managing and operating these things going forward? The way I think about the archetype of like what are the energy resources at a data center historically was you would have a UPS system and you would have backup power.

And those are basically, you had a diesel generator and a UPS system. And that like every data center had those things. Do you think that'll change? Is there a new archetype? I was hopeful early on, but I've never seen one built with, you know, the right type of SLA. So said a bit differently, yeah, I wanted a straight-in facility where the software would have resiliency to fail over, where I didn't have to invest in all the diesel generator or some backup system if the utility failed. I haven't seen one come to market, but we've always been.

been watching that because there is a lot of, I mean, believe me, I don't like to admit the secondary piece, which maybe a lot of the listeners already recognize, we operate almost three gigawatt of diesel generators today. And so finding that balance, and, you know, we do utilize some of those for peak loads and peak shaving and things like that. But I would love to build an environment where for certain workloads, we can build a different type of data center. But just because of the SLAs, because of the requirements associated with these chips where they're liquid cooled. Right? And that liquid cooling, you want almost three-in worth of reliability where if that pump goes down that those hot set of infrastructure, that accelerated compute, continually has the right type of liquid for a certain period of time where it doesn't damage that infrastructure. And we're talking billions and billions of dollars for 30, 35-megawatt build up to 50 megawatts. It's very expensive infrastructure that we're watching. So high hopes, but it never really came to fruition. That's a really interesting point, one that I hadn't fully appreciated, that liquid cooled actually possibly increases the need for reliability, if anything.

That thermal doesn't go away. Even if you had a workload that didn't need it, right? Because the promise, the thing that people talk about sometimes, it feels like it's this like ethereal concept that never occurs really is like, oh, but there are some workloads that like don't need three-nine's reliability or five-nines reliability. It's like, okay, right? We're tagging, the cloud example used to be, like we're tagging photos for Google Images or whatever. You should be able to operate those in a different manner. You shouldn't need the diesel generator. You should be able to place it wherever you want it.

Like, we should relieve all these constraints. And I think there was this concept that, look, if the grid is as big a bottleneck as we think it will be, and it gets just harder and harder to find sites to build data centers at the scale that we need, then, like, naturally the market is going to start to separate out those. workloads and put them in the places that you can still build. But there are all these other constraints. It's not just about the workload. It's, though that is challenging for the reasons you said before, it's also like you don't want to fry the chips, basically, and you invest a lot of capax in those chips. Yeah, and you can double-click on the infrastructure. There's probably some

components that could go in and have a little bit more resiliency. But if you're liquid cooled, the thermal doesn't just displace itself. So you have to continually run liquid through that so you don't damage the chips. But to your other point, the availability zones have grown, right? Because of the constraint, lack of land, lack of entitlements, they've grown, but they're still proximate to that kind of CBD within these critical markets that we see evolving quite, you know, in the foreseeable future. We just see so much demand with, as more customers start to utilize AI and the complexity of these models come to market, we only see that increasing. All right. Final question for you, What are you most excited about?

What's the coolest thing that might be coming on the horizon in data center technology world? Yeah, so I think some of the newer designs we see and some of the efficiency factors, what's awesome is, you know, having two small children myself, we want to be good stewards of the power we take from the grid. So the PUE is increasing, shifting to liquid, you know, liquids 800 times denser than air, so you can get more efficiency factors out of that. I think that's going to be a net positive. And then, you know, I'm a technologist at heart. Some of the designs of the hardware coming to market are just phenomenal, right?

And working with our partners that, you know, across the broad spectrum, watching and working with Nvidia, Vladimir Troy, who runs R&D there, we spend a lot of time with them, not just on the current generation that the public gets to see, but what are the two and three generations out? The ability of the token production against the Watts associated with that is going to be phenomenal. And hopefully everybody, all of our listeners here today, they understand the value of not only the data center, but these tokens and AI. I'm a very pro-AI kind of individual. There's always going to be some negative associated with it, but what AI is going to be able to do for us, not only as individuals, but as a society. I mean, I'm pretty excited about some of the use cases and workloads that we've seen. One case I would leave you with is Geffian, a project we did out in Copenhagen.

It's one of the largest DGX pods for Novo Nordisk. just the amount of pharmaceutical work associated with that one deployment is just what it's going to be able to do for humanity is very exciting to me. All right, Chris, this is a really fun conversation. Appreciate the time, as always. I appreciate it. Thanks for the opportunity. Stay safe out there.

Chris Sharp is the CTO of Digital Realty. This show is a production of Latitude Media. You can head over to Latitude Media.com for links to today's topics. Latitude is supported by Prelude Ventures. Prelude Backs Visionaries, Accelerating Climate Innovation that will research. shape the global economy for the betterment of people and planet. Learn more at preludeventures.com. This episode is produced by Daniel Waldorf, mixing and theme song by Sean Marquand.

Stephen Lacey is our executive editor. I'm Shale Khan, and this is Catalyst.