Speed or certainty? On this AI factory build, they refused to choose

0:00 Speaker 1: Projects of this size and this scale, about 95% of them finish over budget and behind schedule. The fact that we've been able to stick to our schedule, deliver on time, and up to this point stay on budget, that puts us in the top 5%. The biggest cost to uncertainty is our inability to make timely and accurate decisions…

0:24 Speaker 2: …and this technology is helping us do just that.

0:34 Speaker 3: The future of AI demands infrastructure built at a scale the world has never seen before. Applied Digital is helping define that future, transforming ambitious vision into the next generation of high-performance data center infrastructure. At Polaris Forge One in Ellendale, North Dakota, that ambition takes physical form: 2.2 million square feet, engineered to deliver 570 megawatts of power.

1:04 Speaker 1: This was Applied Digital's first hyperscale site, so it was very important for us. AI is growing so much right now; they just can't seem to build them fast enough.

1:14 Speaker 3: But delivering at this scale takes more than vision; it takes foresight, precision, and the ability to bring the very best in the business together around a common goal.

1:24 Speaker 1: The biggest risk to delivery is coordination and communication.

1:29 Speaker 3: From the outset, Applied Digital brings together proven expertise across the construction ecosystem, building a team equipped to deliver against the most demanding ambitions.

1:40 Speaker 4: There are design changes that happen because of modifications to the building itself. A lot of times, the information on those changes does not get communicated out into the field.

1:50 Speaker 5: So that would cause rework, because someone's building off old drawings or an older version instead of building from the coordinated model. You notice there are mistakes after they happen, when you're doing the job walkthrough, so that's already too late.

2:03 Speaker 3: Working alongside McGough Construction, that shared ambition becomes a commitment to build excellence, aligning expertise, solving challenges, and driving delivery forward.

2:15 Speaker 1: The project is moving so quickly, and we really need to have information available right now. We're looking for updates constantly. You can talk to 10 different people, and you can get 10 different answers on something. It makes it really hard to validate what the truth is.

2:31 Speaker 4: In the past, it relied on just field verification. It relied a lot on human observations to walk into a room that's 10,000 square feet, 50,000 square feet. Being able to look at and identify anything on the ground, anything on the wall, anything up in the air that's installed incorrectly – it's almost impossible. And the need to field verify thousands of components, tens of thousands of components in this building, would have detrimental impacts to the schedule, right? And it would result in delays that far exceed what the needs are of the building owner.

3:06 Speaker 3: Through collaboration, trust, and a relentless focus on execution, complex challenges become opportunities to push the boundaries of what's possible. With a little help from some cutting-edge technology.

3:24 Speaker 1: The biggest thing I saw, from our very first build on this property, where we were not using this, to going into the next project, where we partnered with XYZ, is that it has already shown significant improvements on schedule and delivery. I saw a lot more information coming out early on, letting us know that we were either behind schedule or we had areas that were lagging, where we were able to shift resources and make up for that.

3:56 Speaker 4: On a traditional project, without being able to leverage this technology, we would be relying primarily on a slower schedule, right, with a longer completion time frame. But by having this technology, we're able to compress the schedule and still keep installation quality where it needs to be.

4:16 Speaker 2: Well, when we can rely on accurate data, it helps us to be more confident in our decisions. We can make decisions on a faster, larger scale and replicate those as we go forward to more projects in the portfolio, which makes us better, faster, safer, with a higher level of quality.

4:36 Speaker 4: Having the precision that we need to identify better than what the human eye can identify, right? This precision allows us to really know that we can rely on the installation that we have. We're able to be more honest with our schedule. There could be any number of people that could walk into a room and say, "We think we're approximately 60% complete." Somebody else might come in and say, "No, we're only 25% complete." XYZ's dashboard allowed us to really understand, at any stage of the project, down to a tenth of a percentage point, how well we were tracking. Not only that, but it allowed our owner, our building owner, to understand where we were at in real time, and then we were also able to convey that to our trade partners. Being able to adjust manpower, being able to adjust material deliveries – it really just created this cohesive environment of everybody being on the same page, without that human element of guessing.

5:34 Speaker 1: Having XYZ on site has allowed us to have a lot of extra eyes in areas where we wouldn't necessarily be able to keep track of the project progress with our team, the size that it is out here. Managing 2,000 contractors out there, we can't go out there and get all the information firsthand. So we need other ways to be able to stay on top of the information and know what's happening and when it's happening.

6:01 Speaker 2: This level of visibility is just such a stark contrast. This allows us to see in real time how installation is going in, and we are able to take corrective action if we need to, to stay on course, to stay on schedule. And when we have visibility with our owners, they can see that we are actually doing what we promised.

6:21 Speaker 4: Being able to have a uniform system that can clearly communicate all of those design changes in real time allows our teams to be more proactive.

6:31 Speaker 5: And having a partner like XYZ, getting all the issues discovered before they're too late, helps us to avoid costly rework and schedule delay.

6:42 Speaker 1: And having verifiable information that's been proven just makes us able to make decisions that much faster.

6:54 Speaker 6: McGough came to us and asked if we could inspect a series of conduit in the generator rooms prior to the delivery of some of their equipment that was to be installed, to make sure that the delivery and the installation was seamless and didn't cause any delays on site.

7:09 Speaker 7: Upon inspection, XYZ noted that they were two feet out from the BIM model. This is immediately a huge problem.

7:17 Speaker 4: It was a critical item, with our subcontractor trade partner having no awareness that it was in the wrong spot.

7:22 Speaker 7: The current placement of the conduit would have clashed with the day tank. They came out into the center of the day tank when they were supposed to come out into the front.

7:31 Speaker 2: Having that design overlay with what we're installing is really critical to maintaining the quality, the precise locations of things, and staying on schedule…

7:41 Speaker 4: …especially when you have a lot of repetitive components installed in a room. Right, we have a generator room with 12 generators. We've got 96 generators throughout the building. That issue gets repeated all the way down the line. The impact of not being able to identify this incorrect installation would have been really detrimental.

7:59 Speaker 8: We would have completed it and then found out it was done wrong. We would have had to break concrete, remove conduit, you know, do recompaction, do more concrete. It would have pushed back everything that we had worked for.

8:15 Speaker 4: We knew that we had equipment that was being delivered for that generator, and being able to get in front of that prior to the placement of the housekeeping pad, prior to the placement of the equipment, we were able to get that moved and relocated to the correct location without missing the installation date of that day tank. It was standard practice from that point on to have XYZ in every room before the concrete pour.

8:42 Speaker 1: The biggest value I see is that as we continue to work together with XYZ, we are able to improve our communication and improve the process, and that helps put us in a position to deliver even faster and more accurately – sometimes also at a lower cost – in our future builds.

9:05 Speaker 3: At this scale, progress is never achieved alone. It is built through shared ambition, trusted partnerships, and the collective determination to deliver what comes next.

Polaris Forge 1 is Applied Digital’s AI factory campus in Ellendale, North Dakota, built by McGough. It covers 2.2 million square feet and 570 MW, with 2,000 workers on site, all working to a fixed ready-for-service date. This short documentary follows the owner and GC teams as they work from one shared, verified view of the build from XYZ Reality. Our engineers check each install against the coordinated model as work goes in, so the teams can spot risks to the schedule early enough to act.

By measuring installed work against the coordinated model rather than estimating it. On this campus, progress was tracked to a tenth of a percentage point, so a room reported at 60% couldn’t turn out to be 25%, and the date everyone was planning around could be trusted.

A walkdown shows what’s visible on the day, after the work is in. In rooms tens of thousands of square feet, with services on the ground, on the walls and overhead, checking each element as it’s installed is what brings risks to the surface while there’s still time to act.

A laser scan or BIM snapshot is reactive. It records what was installed after the work is in, so any deviation it finds has already been built over. XYZ Reality works proactively: our engineers check each element against the coordinated model with engineering-grade accuracy as it goes in, so risks are prevented before the next trade builds on them.

On Polaris Forge 1, XYZ Reality’s engineers found conduit 2 feet off the BIM model in a generator room, before the concrete was poured. Left in place, it would have meant breaking concrete, pulling the conduit, recompacting and pouring again, with the same layout repeated across 96 generators. It was moved before the equipment arrived, and McGough went on to have every room checked before the pour.