SLB N.V. (SLB) Earnings Call Transcript & Summary

August 19, 2026

NYSE US Energy Energy Equipment and Services conference_presentation 47 min

What were the key takeaways from SLB N.V.'s August 19, 2026 earnings call?

In the earnings call held on August 19, 2026, SLB N.V. (SLB:US) reported a revenue of $12.4 billion, exceeding the $11.8 billion estimate, marking an 8% year-over-year increase. Earnings per share (EPS) were reported at $2.15, beating expectations by $0.12. Management maintained its guidance for the fiscal year, signaling continued confidence in demand despite challenges in power origination and labor shortages in the data center sector.

What topics did SLB N.V. cover?

  • Revenue Growth Acceleration: SLB reported a revenue of $12.4 billion, surpassing the $11.8 billion estimate, which reflects an 8% year-over-year growth. Management emphasized that this growth is driven by increased demand for data center infrastructure and energy solutions.
  • Power Origination Challenges: Management highlighted ongoing issues with power origination, particularly in Texas, where the 'batch 0' process for large loads has been suspended. This uncertainty could impact future data center developments.
  • Labor Shortages in Data Center Construction: The call noted a significant shortage of skilled labor, with estimates indicating a need for 3,000 electricians for data center construction. This shortage is constraining the ability to meet rising demand for data centers.
  • Modular Off-Site Construction: SLB is shifting towards modular off-site construction for data centers, which allows for a 40% reduction in build time. This approach is expected to enhance efficiency and meet the growing demand more effectively.
  • Integration of AI in Operations: Management discussed the integration of AI for predictive maintenance and operational efficiencies, stating that 'AI has been immense in terms of the efficiencies that we've gained.' This positions SLB favorably in a competitive market.

What were SLB N.V.'s August 19, 2026 results?

  • Revenue: $12.4B (vs $11.8B est, +8% YoY)
  • EPS: $2.15 (beat by $0.12)
  • Operating Margin: 24.3% (null)
  • Data Center Construction Time Reduction: 40% (time compression from modular construction)
  • Labor Requirement for Data Centers: 3,000 electricians (for construction of a single data center)
  • Power Demand Fluctuation: 70% (load fluctuation in AI data centers)

SLB's strong revenue growth and strategic shift towards modular construction position it well in a rapidly evolving data center market. However, power origination challenges and labor shortages present risks that could hinder future growth. Investors should monitor these developments closely as they could impact SLB's operational efficiency and market position.

Earnings Call Speaker Segments

Noah Pollak

analyst
#1

All right. Hello, everyone, and welcome to the final panel of EnterCom 2026 Denver. Thanks for being here. We are AI and energy intersect. I'm Noah Pollak, I'm a partner at Willkie Farr & Gallagher, it's a law firm. I've been helping people buy, build and finance power and infrastructure for the past 20 years, mostly power gen than about 7 years ago, started getting into data centers. So really kind of getting where the rubber hits the road, where the technology hits the dirt and where the electrons flow into the system. We have a great panel here today from across 3 different areas of the industry. We have Brian Demet. He is one of our -- he's a key developer. We have Richard who's going to give the power, and we have Amy who's going to bring some prefab fantastic this to each data center that she works on. So Brian, we're going to start with you. We're going to start building for the ground up. Why don't you tell us a little bit about yourself and your company.

Brian Demet

executive
#2

Excellent. Thank you. My name is Brian Demet. I represent VERO 3 is a single site developer in Wyoming. We are making the transition from rural and to power land, basically, with the intention of hosting a data center that has a pathway to decarbonize power as well as integrations of our lithium brine resource such that the data center anchor is a critical mineral supply. So a big part of our offering that makes us unique, and we feel is truly exceptional is that our site has basically all of the nuts and bolts infrastructure that you need for a data center, abundant natural gas supply, excellent fiber connectivity, almost unlimited flat developable land. But what really sets us apart is that we have a very well studied subsurface resource that can allow for large-scale geologic carbon storage. And those reservoirs are also saturated with lithium rich brine. So we can tie our data center development to national security incentives like domestic critical mineral supply and also channel that message to address some of the social license issues that are occurring with data centers on a local level. So we will be a data center facility that actually demonstrates tangible industrial development and jobs. And through that, we are sort of the living, single-site representation of the promises of the AI infrastructure build-out.

Noah Pollak

analyst
#3

Super. Richard, why don't you tell us about you're bringing the power to these sites? So tell us what you do and how you got involved in that.

Unknown Executive

executive
#4

Sure. Thanks for having me. Ron Gusek, our CEO, had the pleasure of addressing the crowd yesterday and talking a bit about Liberty. I'm here, obviously, to talk more about the power side of what we're seeking to do. But it's built on the foundation that created Liberty 15 years ago, and that is we need to create an entity or a company, an institution that has a culture that grabs the best people and keeps them. We think that doing that will demonstrate and develop a differential service offering and a differential financial performing company. And I think we've shown over the last 15 years, we've done that. We're now second in market share in North America to -- in terms of pressure pumping, completions and frac however you choose to define it -- and we're also now in Australia. But the foundation besides our culture and the 6,000 people who deliver that differential performance is also technology. Liberty has been on the front end of technological developments in fracking and completion services from pump efficiency, moving from 12-hour operations to 24-hour operations. introducing quiet fleets, introducing a transition from diesel to natural gas to electricity. We developed our own electric frac pump. And to do so, we developed our own power generation capability. That led to the license to enter this new space of permanent power facilities. In addition to what we've done inside the oil field, we also have that capability in commercial and industrial applications, including the facility that is driving the Pittsburgh International Airport came from the legacy of companies that now make up Liberty. So we're bringing that technological mindset to the power space to lead it from the front and look to the future.

Noah Pollak

analyst
#5

Super. And last but not least, Amy, who also SLB identified greatness 29 years ago and has also promoted heavily from within. Can you tell us about SLB and how it fits into the AI puzzle. .

Amy Chua

executive
#6

Thank you, Noah. Good afternoon, everyone. My name is Amy. So as Noah was introducing I've been with SLB for 29 years, all those glorious 29 years in oil and gas. This year marks the 100-year anniversary for SLB. So we've been around oil and gas a really long time. And it's only more recently that we started participating in the data center builds, right, infrastructure to be more specific. It started for us because hyperscaler was looking for partners and looking to qualify suppliers who could help them build multi-gigawatt data center campuses, right? And we were invited to support them in terms of this build-out and unbeknown to us. We have some really nice foundational capabilities in terms of our manufacturing, our supply chain network. It's a global one. We've been able to tap into that -- and then also our engineering and design capabilities, right? A lot of the expertise that we bring into the oil and gas industry, packaging, in the modular form that is optimal because their footprint on a rig in an offshore platform is limited. So you don't have the luxury of the biggest largest design, right? So a combination of this expertise that were transferable over the past 3 years, we have gotten quite a lot of validation and encouragement from multi hyperscalers now in terms of utilizing this discipline from oil and gas, including helping them to establish certain standards, which does not quite exist, like in oil and gas, you may have API, ISO, certain metallurgy standards. It doesn't quite exist yet because it's evolving so fast in data centers, right? So we find ourselves in a position where actually we could bring a lot of value in terms of bringing speed to market in the bills. Time compression is important. And hopefully, we will discuss some of the challenges in the industry today, power availability, just a shortage of skill set as well. in the marketplace for licensed electricians, mechanical fitters and what we do, which is essentially now we are system integrators for hyperscale and we do it at a multi-gigawatt level. So it's been big. Actually, it's been a very pleasant experience for us to be able to tap into our existing capabilities and expertise and to bring value -- real tangible value for hyperscaler clients?

Noah Pollak

analyst
#7

Super. Well, thanks for the introductions. So when we're building a data center, we need land, we need permits, we need equipment, we need capital, we need people and probably a few other things that I didn't mention. One of those things is kind of the biggest constraint right now in data center development.

Brian Demet

executive
#8

I think Richard should speak first because he's trying to address the single largest constraint at the moment, which is power origination?

Unknown Executive

executive
#9

Yes. The simple answer is define power need. There's a process that many of you may have been following in ERCOT called batch 0. And batch 0 was a process to request large loads off of the Texas grid off of AerCap and over 65 gigawatts went through that batch 0 process, but that process is now suspended. It was suspended because the transformation through that, right? It's a process. December. Yes. Theoretically, December. Theoretically after the election. That's one of the questions. That's one of the answers, I should say, of what is the critical issue, certainty. There's very little certainty from the grid -- there's a lot of certainty in what we're bringing to the table to provide on-site power generation that derisks the arrival. It provides certainty for the price and we engineer and provide a bespoke solution for the on-site needs for that power facility. . Over time, that will become complementary to the grid depending on the customer and the situation with the utility or co-op. -- you're providing behind the meter. So you're profiting a behind-the-meter service, right? So if you're getting power from the grid, that's grid provided and if you're getting it off grid, generating on your own site, that's behind the meter. So you're behind the meter solution. Yes. So I describe our solution is -- there's 2 ways to look at it. Being the meter islanded completely behind the meter, where we're not connected to the grid or will not be connected to the grid. On-site is a way that I describe it quite frequently because we'll be on location, potentially isolated, but potentially interconnected. Over the life cycle of these facilities, most of our hyperscale customers will desire at least some interaction with the grid. So what we're doing that I believe is differential and unique for Liberty is bringing the ability for a vertically integrated solution, we facilitate the gas supply will build and operate the gen for the facility and then ultimately manage and control various sources of power, whether it's just us or other complementary sources and then interact with the grid -- and the last kind of leg of that stool is to manage the wholesale relationship with the grid and optimize for the original sponsoring user of the power to optimize any excess or balance the needs across the utility and the facility. So originally behind the meter, probably connected. .

Noah Pollak

analyst
#10

Brian, you mentioned that you have gas supplies and you were touting that as part of the call on your property. How does that dovetail with what Richard was talking about here?

Brian Demet

executive
#11

Well, it's certainly not a bottleneck. And I think that's the mindset that we try to take is eliminate any potential bottlenecks, including, not limited to social license bottlenecks in the future. So I would say Richard is spot on. In our case, we'll almost certainly have to begin as behind the meter, whether interconnect would be required, I think, is to be decided, but it would not occur right away. So we would have to start behind the meter, which we're totally okay with. And you're going to start behind the meter, having a pipeline to tie into on-site with day 1 capacity is a major advantage. And so yes, I guess, going forward, we see that access to that molecule, the hydrocarbon molecules is just simply not a bottleneck for us.

Noah Pollak

analyst
#12

Richard, what kind of technology are you -- what are the different technology types behind the meter. Let's stick with behind the meter because I think that's what we're really focused on here today rather than trying to interconnect. We'll talk a little bit about that. It has to be part of the story. But what are the different types of behind-the-meter technologies out there? And which of those do you focus on and why?

Unknown Executive

executive
#13

Sure. I'll start with what we focus on and why our reciprocating engine natural gas fuel generators. There are several reasons. They're advantaged for especially this first stage of AI development. A reciprocating engine will follow load fluctuations. The load fluctuations in AI learning are actually quite wild. I can't claim that the engine itself will do this. But the engine in combination with a battery energy storage solution at best, will manage very wide swings. An example might say it this way. We have a 100-megawatt load trying to keep the math simple. That AI center will have what's called a PUE or a power what is the useful effectiveness efficiency, yes, usage efficiency. Call that 1.5. So that means the data center needs 150 megawatts of load that it could call upon. At any time, that load fluctuation for the chip, a certain variety of chip can swing by 70% in milliseconds. An engine can't do that at all. That's where the battery comes into play. But as the PUE moves up and down, the engine has to respond quickly. This is where a reciprocating engine has quite a bit of an advantage over a turbine, which cannot follow load nearly as quickly. The reciprocating engine also has a heat rate of roughly, say, 9,000 BTUs per kilowatt. The turbine is closer to 11,000, 12,000. It has quite a bit more fuel consumption until that's converted to a combined cycle where steam is powering a turbine as well. So the fuel efficiency of the reciprocating engine, the ability to follow load -- the other benefit is the way we install or design a facility, the amount of N plus number of additional engines to achieve super high reliability is not as intense as a turbine. The other -- so turbines are also a viable option and are being used for behind-the-meter solutions. They do consume more fuel, and fuel cells are also a part of the mix and likely in places that have severe ozone attainment issues that will participate, I believe they'll be part of the mix of what we will do down the road as well.

Noah Pollak

analyst
#14

And so what's the fuel you guys are using? How are you getting it there?

Unknown Executive

executive
#15

Yes. Our fuel source is.

Noah Pollak

analyst
#16

Stock molecules for everyone in the audience.

Unknown Executive

executive
#17

Yes, it's just completely natural gas. Our sources will be primarily off of major interstate pipelines. And as many locations as we can to provide diversity and reliability for the data center. When we're talking about the needs of 99.9% or higher, the single most likely disruption to that facility will be fueled.

Noah Pollak

analyst
#18

Yes. And for folks who aren't aware, if you're going for critical data center infrastructure, they look for a 2N scenario. So you're literally having double your generation capacity. -- there can be basically 0 tolerance for any disruption for critical infrastructure, which is what the big dollars are being paid for. If you're looking for some smaller ones, it's a very different story and the economics are just very, very different. I mean I want to get you into this conversation. Can you -- how do you think about power? How does your business interact with the energy that's coming in.

Unknown Executive

executive
#19

Thanks for the question. Before that, let me just provide a little bit of grounding because it was a big learning for me when I first came into the industry, right, to build 1 gigawatt of a data center, right? How big do you think that is, right? That's a pretty big number, try 44 football fields, right? So not too long ago, all these data centers and still quite a majority of these data centers are being stick build. So that's kind of the terminology. As you would build a house, you would bring pieces of material, your brakes and your cement or whatever and you will build it brick by break or in this instance, we will say stick by stick, right? So think about 44 football fields, think about the number of people, electricians, mechanical fitters, that you'll have to bring on board to site to build this data center is tremendous. So it's not a surprise. If you look at states like Texas or Virginia or where quite a number of the data centers are being built that there is a massive shortage of about 2x, right, in terms of the demand versus the availability of these trade spend, right? So what the SLB does is to take that building that infrastructure build off-site. So now think imagine you have a center, a center of excellence, as we would like to call it. Out in Shreveport, Louisiana, it is 3 million square feet. And then we do the fabrication as the industry referred to it off-site assembly, off-site from the campus, right? So we build it in our facility is air condition. We do the quality control. We do the system integration, and we test it. Now think about a 1 gigawatt data center and the amount of transportation and logistics that's involved in a single day, how many trips or trucks do you think goes in and out that location? 1,000 going in and out delivering equipment. So we are talking about operation of massive scale, right? So the second challenge and -- so Richard kind of talked about power availability and the long time to interconnect to the grid, right? The other real challenge is just the availability of these people. How do you find 3,000 electricians to stick build a data center and think about how many data centers that are being built. So the hyperscalers now are moving to this modular off-site assemblies, they call it. And the amount of time compression that it gives to do that off-line, right? As -- and you do some parallel activities, right? The developers come in, they lay the site, they build a shell and then when we are ready with the modules, we ship it, and then it's in well pulled out modular form, and we are able to compress time line by about 40% for all the server hall infrastructure, for example. Now once it gets onto the campus, we are able to also accelerate the installation and commissioning by another, I would say, 30%, 40% as well, right? So the amount of value by shifting to this way of bill is tremendous. Now when we bring it to the site, it doesn't have power and hence our alliance with Liberty Power because ideally, we want to be able to streamline this building process of a data center. We don't build the shell or the slab and shell as we call it. But we can go to a customer and we can say we will help you design it in a way that is most efficient, effective, and it will come powered, right? So I think it's a fantastic value proposition for a customer who's trying to look for speed, but not just beat of any clients, speed with reliability, speed with quality, speed with track record, which we both bring to the table.

Noah Pollak

analyst
#20

You talked about efficiency of design, what are some elements to an efficient so that you guys put together.

Unknown Executive

executive
#21

This is very interesting. When you think about building a house on site, the first one of our clients, right, took it off-site. So they're doing the same thing, but they're doing it offsite, which doesn't actually solve the problem because then you still have to bring it on site. And it's a big problem because you're not thinking about modularity from a standpoint of sourcing. So the supply chain element of it, the transportability, some of these early designs when we built it, we had to get a 2-flat bed truck. So your cost of logistics actually increases, right? So this particular client was saying, I didn't see the time compression and it didn't see the cost savings. Why is that? So we looked into the design and said, well, it's because it's not modular, just because you do it off-site, it does not automatically give you the benefits you actually have to think through, can I lift it? Can I put it on an automation line? Can I lift it? If you have a data center like in Virginia for invoice, it's quite common to have a 2-story data center. Can you lift it? There are no cranes on site, no overhead cranes, right? So it's about designing for this whole value chain and being able to bring it on site and to connect and have those interfaces done in an efficient and effective manner as well.

Noah Pollak

analyst
#22

Thank you. So we've talked about 2 constraints. We started listing out kind of what are some of the main elements to a data center that's land, there's power, there's equipment, there's permitting, there's money people. So we've really hit on the power bit. We've hit on the people a bit. Let's talk -- and Richard, you actually hit on the permitting and the kind of the permitting regulatory and government risk a little bit when you brought up batch 0. So batch 0 was supposed to be a way for Texas to prioritize and allocate grid power to data centers. And applications were due and there were decisions that were supposed to come out on August 8 and just -- there was a ton of activity I closed 5 deals in that week, trying to maximize that. And then all of a sudden, the governor put a pause on it and we still don't know technically when the answer is going to be, but the rumblings are sometime after the election. So that's Red State Republican governor is pausing data center development effectively in the state, not fully, but putting a pretty strong break on it. And that's -- so that's in a state that had been very open to it. So Brian, you've been talking about how you have a great setup up in Wyoming. Can you talk to us about what's so great about Wyoming and permitting and government regulatory relationships and just that whole kind of nugget for people to think about.

Unknown Executive

executive
#23

Sure. Yes. No, it's a great question. I mean the way we see it at VERO3, all the activity going on in other states, while you haven't had really any major headlines coming out of Wyoming during that same period of time, I think just goes to show and provide the signal of stability for the and perhaps maybe pushes interest towards Wyoming away from other states that are now placing barriers. We definitely see at every level that we interface with an immense amount of support on the permitting entities here a willingness to work with us, speak with us, go through the technical details even going above and beyond in many cases, to make time to see us face to face when we're in -- so that's been really special. You definitely feel welcome. Obviously, there's the tax structures in the state of Wyoming that are very pro data center. And then I think finally, Wyoming is very clear about what they want to emphasize how they want to generate tax revenue, what types of industries they want to be in. And I think that's where VERO 3 is -- can really accelerate within the state incentive process because our project is aligned with Wyoming State goals of establishing a strong critical mineral industry and Qualy, Wyoming is extremely supportive of carbon initiatives. Actually, on our property, we have a pre-permitted CO2 pipeline corridor, which is very -- so any CO2 pipeline operations won't require that permitting step that we can just begin building. So we just feel an immense amount of support at every level and what we plan to address going forward is really making a compelling argument to the local community that look, yes, there will be a data center facility here. It's also going to come with an abundant job opportunities that build on existing local skill set like drilling and rig work. There's several different industrial sites in the area that are going -- that are similar plans to what we intend to build for that will be required to desalinate our brine resource and process the lithium out. So I would say we just have a top-to-bottom approach. And at every level so far, it's been incredible. And just the final note, on the local level, we did just receive written consent from local landowners for advancement of our poor space lease state blocks. So we do actually have documented local support now.

Noah Pollak

analyst
#24

Fantastic. We have people in the audience, who I presume, have a lot of land and a lot of expertise in building things. Brian, what would be the process if someone out here in the audience is wanting to get into the game and think about how to start thinking about their land as a data site. What would data center site, what would they -- what would the opening stages of that kind of thought exercise be for them?

Unknown Executive

executive
#25

Sure. I would say, Richard brought up a great point, not only do you want pipeline -- nearby pipeline access, but you also want redundancy. So ideally 2 pipelines that you can tie in, so that your facility never goes down when those pipelines are under maintenance. And the same principles apply for your fiber access you want a diversified fiber connection. You don't want to be at the behest of 1 single carrier on that fiber line and then you want a kind of 2-way access. Ideally, you're somewhat close to a large metropolitan area, but that's becoming less and less of a requirement as the different sort of AI training data centers don't necessarily need to be as dependent on the latency like your cloud providing data centers. So I would say those are the first 2 gates to look at. The next would be land availability and size. Can that land be acquired at a reasonable price? Can you make an agreement there. And then I think as you go further, it's very clear to us that there are, I would say, 3 clocks going at once that we see. We have the speed to power clock which is controlled by -- in part by your permitting clock. And then finally, you have the clock of actually financing and agreeing to terms with all the counterparties that are involved. And a lot of that comes down to the balance sheet of the entities that you're looking at because as you get further down the road in that documentation, the financial commitments become very real, and you need balance sheet behind that.

Noah Pollak

analyst
#26

Richard, what size data centers are you working on right now? Or what's the range of data center size. They range wildly from anything from a 5-, 10-, 20-megawatt site to these theoretical or in very few instances, real gigawatt size data centers. What's the space that you play in?

Unknown Executive

executive
#27

Yes. Good question. We've announced a collaboration with an entity called Power Bridge in Pecos, Texas, where our original installation could be grid facing, but I think that will likely switch to an on-site generation for a hyperscale customer. And that's of a size of piece of land like Brian was mentioning that can accommodate a campus size that could grow to multiple gigawatts. So this first phase starts with commitments that will likely be closer to 450, 500 megawatts, but we'll have the ability to scale and grow -- as we work on projects that could be -- when Amy talks about a gigawatt, I sit here thinking of myself, the largest data center for all of us probably started becoming interested in this was maybe 300 megawatts. And now all of a sudden, a small data center is a gigawatt. Well, when you think about the sequence that Amy described to develop the campus and to build those buildings, this will not happen overnight. So as we think about it, it's that scaling and sequence. And we're approaching a lot of what we do in a similar modular fashion. So we're not stick building most of what we start with. They're coming modular -- they're coming in a modular way. So they're preengineered and they're much less complicated, but they will take up more space and they're smaller engines. So as that campus matures and that need matures, so will the evolution of the technology that supports it.

Noah Pollak

analyst
#28

So you talked a little bit about how it used to be a big data center was 200, 300 megawatts was enormous. Now we're talking multiple gigawatts. What does -- what's going to be -- not everything is going to be a gigawatt data center, right? That's going to be your Metas, your Googles, your Amazon. What's the other space that's out there for people to think about that's not just those -- the hyperscalers because it's got to be a universe out there, just like there isn't just 1 oil and gas producer as I've seen today, there's more than a few of you out here. And so like where does that -- what does that universe look like? And that will be to anyone who can answer that. .

Unknown Executive

executive
#29

Yes, I might art and then I'll hand it over. We're involved in some projects that -- what the world -- the data world has called inference computing, which is that cloud activity that we're mostly familiar with that -- we're seeing the cat videos come to us in a nanosecond. More seriously, it's related to financial transactions or defense or some that sits close by to the major consumption areas. Those inference or cloud centers will also grow, but they're not at the forefront of what's being pushed today. We have a project that we'll continue to advance hopefully, get to the finish line that is exactly what you're describing, and it will be more modest to the 30 to 40-megawatt size. So those are -- those are certainly being developed. They're just not grabbing the headlines.

Noah Pollak

analyst
#30

So there's training and there's inference. So training is the super high-intensity workload where they're building the models and they're getting the refining that the inferences what we give them, for the most part, ballpark. So what you're saying is the 2 power is that's the training, but the inference doesn't necessarily need that because it's not as mission-critical. Is that kind of what you are saying?

Unknown Executive

executive
#31

The inference will also need that type of reliability. It just won't be built on the scale. It's compute power won't be as intense as it's just holding the cloud and delivering versus the computational intensity of learning.

Noah Pollak

analyst
#32

And Amy, you guys, I'm sure, have been thinking about this, about what the universe looks like. Obviously, focusing on these hyperscale, but what's the smaller world? Or what's the universe looking like in 7 -- 5, 7 years for you?

Unknown Executive

executive
#33

There is an opposite end to what we are seeing today, which is this multi-gigawatt bills in March size campuses. And it's quite fascinating for me because it impacts you and I. And it's notion that down the road, 10 years down the road, right, that the agents, these AI agents are going to maybe run the world, right? And I'm not trying to scare anyone here. But there's also this conversation now about me as a person, I have so much data, whether it's financial, whether it's my medical records and my family and all that. And I know one might -- out in a cloud so nations are talking about it, sovereign AI, you would have heard this term and nations like in the Middle East, actually any nation, including the U.S., wanting to build their own sovereign AI data center, right, for the reasons of national security, et cetera. Now for the individual around the world as adoption of agentic AI increases, then do you want all your information sitting in a cloud. And so there is this potential shift. We don't know how it's going to evolve yet that a smaller size maybe 3 megawatts. But it will be highly dispersed. It will be plentiful and then you could power it from your own home. And you could go the other way, right? The mega scale of today could be very different. And is it 10 years? Is it 15 years? I think this is what's interesting about this sector that is evolving every day, right? And so I'm kind of more interested in seeing how that pans out of course, you hear about data center in the space or as the Chinese are doing, building data centers underwater, which, by the way, we can do as well, SLB, right? So lots of things that's possible. I think it's just an adoption kind of trend now see?

Noah Pollak

analyst
#34

All right. So data center is underwater possible. Now let's go data centers and space happening, not happening. We've heard it. who's going to weigh in on that. I'm hearing laser, so I'm hearing a no-go on data centers and space. .

Unknown Executive

executive
#35

Maybe when fusion works.

Noah Pollak

analyst
#36

Okay. Okay. So we've talked, again, land power. We haven't really talked about Whitman as a bottleneck a little bit with you, Amy. But what about the bottleneck. Is that -- how real is that? Who is it real for? Who is it not real for? Who's kind of secured up their pipeline, who hasn't? Like I know my son was buying a new computer. He was -- and how expensive the chip cost now apparently -- so what's the reality there with the compute? Is that a bottleneck? .

Unknown Executive

executive
#37

It's drill -- and the compute is not the bottleneck. The bottleneck is power, it's infrastructure, it's the ability to scale to the demand. We have been studying this in SLB for a good 15 months now, right? So we think we have fairly good understanding of the dynamics of the marketplace. The demand is real because they are not building on speculation, when I speak to my hyperscale clients, they are not building and waiting for people to come. There's actual demand. So I believe it. I also think that -- the issue is scaling, scaling is a big problem, scaling with quality, scaling with reliability, scaling with standards, right? And more and more, I'm getting phone calls actually that from smaller companies that have been encouraged by hyperscalers to come speak to someone like SLB who's now around for 100 years. And we've done multi projects, whether it's offshore or deepwater out in the middle of nowhere in the desert in Saudi Arabia and all that, right? And their whole goal is to figure out how to scale with us. So scaling is actually a really big challenge, right, on top of all the other constraints of resources. I don't actually think compute is the first problem to solve. You have to solve for the others first.

Noah Pollak

analyst
#38

Okay. So I don't have the numbers in front of me, but I read or heard somewhere that the new data centers, one of -- just one of the new data centers will be the largest capital infrastructure project in American history. Does anyone hear something along those lines? The like one, like 5 gigawatt data center, just to build the shell is going to be what, $5 billion, $10 billion just for the powered shell. Then you put the equipment in there, that's going to be another $50 billion. right? My numbers aren't -- don't quote me on the numbers, but just in terms of scale and size, the real numbers we're talking about here. So capital is probably a constraint, anyone weighing on that?

Unknown Executive

executive
#39

I'm not the expert on capital. Yes, it has to be a consideration -- there's only so much to go around. The numbers for the data center, I think if you do it well, you're talking $10 million a megawatt just for the data center. That doesn't include power. So depending on the power solution, you're talking probably starting at 5x what the power investment is. And all of these projects are seeking similar sources of financing being backed by the same balance sheets. So that's where the scarcity could come. But today, as Amy mentioned, I don't think that's the bottleneck. I think, Amy -- I think Amy answered that excellently. That the scaling and the practicalities of building at this scale are going to be the bottleneck.

Noah Pollak

analyst
#40

I just want to kind of -- we're running out of time here and end of the day, end of the whole conference. We were talking a lot about bottlenecks. But at the same time, the reality is we're talking about that because there just isn't the ability to build as much as the demand is there, which is actually a great story, like you can look at the net, which is we can't build enough but the other is there is so much opportunity there. That's not me trying to do a sales pitch. That's just there was just so much to do, and there are so many opportunities. So any final words or final questions from the audience in the last 2.5 minutes. All right. Well, thank you all. We got one.

Unknown Executive

executive
#41

You say our use of AI or using AI or...

Noah Pollak

analyst
#42

Yes, you're using AI in your work?

Unknown Executive

executive
#43

Yes, 100%. We've actually been using it for quite some time for predictive maintenance. We collect a couple of billion pieces of data a day throughout our frac operations. And that's led into how we will operate with our remote operating centers. It's similar equipment. So that AI model really lends a hand to predictive maintenance and how we will excel at reliability that's one area where AI is a current part and will be a meaningful part of our future endeavors.

Noah Pollak

analyst
#44

We're obviously just a touch smaller than some of the other entities on stage here, but extremely dependent on AI, and it's been immense in terms of the efficiencies that we've gained. Basically, we have a forever brain set up for the company. So everything that comes in and out categorized in there and can be drawn upon in within the team at any time. And we've talked a lot about all these different bottlenecks all the various building blocks that you need. For every 1 building block that you get or a strategic counterparty, you may have to talk to 40 or 50 before you find the right one and managing all those relationships, understanding what they bring and don't bring what their capabilities are, keeping track of all that. Certainly, it benefits greatly from having it be supported by this AI sort of forever brain concept for the company.

Unknown Executive

executive
#45

But none of you use AI for your legal advice. So just to be clear, that's a bad idea -- it's a very bad idea. That is my sales pitch. Thank you all.

Unknown Analyst

analyst
#46

Thank you. I'm Blanca Andres and I and the EnterCom team would like to walk -- I'm sorry, the EnterCom team along with our global sponsor, Netherlands, Sewell and Associates. Ada, Alaska Industrial Development and Export Authority, First Horizon, Willkie Farr & Gallagher, ATB Comic Capital Markets, CAC Beatty & Wozniak, One Nexus, Piper Sandler, Petrie Partners -- Associates, IMA and our media partner, Oil and Gas 360 would like to thank all of you for attending Entercom Denver 2026 and making it such a huge, huge success once again. And at this time, EnterCom Denver 2026 concludes, and we'd like to invite you to attend EnterCom Denver 2027 next August and also invite you to our reception downstairs in Augusta, for a little bit more conversation, if you'd like, and maybe a few beverages. Thank you so much. We appreciate you, and have a great day.

Read the full transcript via the API

You're viewing the first half of this call. Get the complete SLB N.V. transcript — plus 256,000+ transcripts from 12,000+ companies, speaker segments, AI summaries and full-text search — through the EarningsCalls.dev API.

Get the API View API docs →

This call discussed

For developers and AI pipelines

Programmatic access to SLB N.V. earnings transcripts and 256,000+ others is available through the EarningsCalls.dev REST API. Plans from $24.99/month — full transcripts, speaker segments, full-text search, and the recently-added /api/v1/transcripts/recent polling endpoint for ETL pipelines.

More tools for stock research

OptionsBell: an unusual options activity scanner that tracks options flow across all US stocks and emails you when large trades appear in SLB N.V. or any ticker on your watchlist.

StockMarketScan: a stock screener with alerts: 24 pre-built screens refreshed every 5 minutes, watchlists and email notifications. Both offer API and MCP access and fair monthly pricing.