Datacentres with Richard O'Brien and Eddie Beer

Yeah, I think there's probably a lot of different angles you could take here, Stuart.

But I think initially what's probably most striking to us is the fact that if you take a typical construction project, let's say commercial office, a hotel, a block of apartments, the IT, the MEP, the technology serves the function of the building, right? whereas data centers is quite rare in that really the building is just really just there to serve the actual function of the IT, the MEP, the technology.

So I think with that brings its own kind of inherent complexities.

So there's a lot of systems, a lot of tech, whether it's the kind of IT side in terms of the servers, the CPUs, the GPUs, whether it's on the MEP side in terms of your cooling, your power, your resilience.

So there's just a lot more potential to go wrong ultimately if you were to compare it to kind of a more traditional construction project.

That's probably the first thing I'd say.

The second thing, which I think is quite interesting, is the scale of change in this industry.

It's unlike anything that we've seen.

And I suppose specifically what I mean by that is if you were to take a typical project lifecycle, let's say four, five years of a typical project, Within that four or five years from, let's say, concept to handover, you might have gone through two to three iterations of the same technology within that time frame.

And that can be on the IT side in terms of moving from CPUs to kind of more power hungry GPUs.

But it can also be things like the how we actually kill these buildings, whether it's cooling with air, whether it's cooling with liquid.

So the technology has changed, does change so quickly.

And I think for us as designers, that means we kind of have to design with one eye on the current, but also an eye on the future.

So having that built-in flexibility within the design is very important.

So that's probably what I'd say in terms of the difference between a typical project versus a data center.

And I think the last point I'd say is really the consequence for failure.

Particularly at the hyperscale end of the market, like 100 megawatts plus, You know, these things are kind of turning over more than 10 million pounds per day, tens of millions of pounds per day in revenue.

So you can't really afford downtime.

You can't afford delay.

And within that then brings its own complexities, not only in terms of the technology that we're providing, in terms of redundancy and resiliency, But things like program pressure, like we've seen this where, you know, ultimately the contractor should be procuring based on an agreed design.

But if you've got program pressure, sometimes, you know, there's a bit of a nuance there where maybe the contractor is procuring based off a 90% design.

So there can have a 10% risk there that needs to be factored in.

And that can be quite a challenge and quite dangerous.

So I think it's just, yeah, I think it's the scale of change and I think the consequence of failure.

They'd be the kind of the big things that I would kind of