Robert Ussher

speaker

856 appearances 4 recordings 1 series first heard Jan 2026 last heard 12 Aug

Robert Ussher’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

Recordings per month over the last 12 months — 4 in all, peaking in Aug 2026 with 1.

Appearances

newest first · ▶ plays the moment

Cooling Data Center Podcast · Clifford Federspiel, Vigilent: Self-Driving Cooling, Predictive Control & Why Fixed Set Points Are Costing You More Than You Think · 8 Jul 2026 podcast

Yeah.

And now instead of delivering the kind of efficiencies that you'd like to, uh, reap the reward of deploying liquid at, you're paying for twice because the big

equipment that needs to chill the water.

You've got to double up on more pumps, more fans, more chillers.

Yeah.

I guess that's an interesting one, because you said that there's so many, you said there's multiple stakeholders in the Vigilant system.

And it's a very interesting one that

I'm thinking of it just from the engineering point of view, because that's where my head is at.

But when you think about it from the cost point of view, and you think about it from the strategy point of view, you can then start to reap the reward of more simplified designs.

Because as we know, most of the systems that we see out in the plethora of data centres that have been built and are being built at the moment,

If you actually had 100% cooling from all of the devices in it, it's wildly, wildly over the top for what the load is currently doing.

And I guess that's because people are frantic in trying to come up with a design solution that will accommodate the next iteration or the next variations without having to go and do a rip and replace on the data hall.

Whereas what might be

worth considering is the vigilant system because that will actually give you the capacity breakthrough to be able to reap the reward of the savings when you can and yet move to max cooling if you do need to go down that route in the future or change to a more liquid rather than air cooled strategy in the future.

Yeah.

I'm aware that like we often say when people want to, if somebody wants to go from one X to two X in terms of a manufacturing throughput or output, we say design a different machine.

Don't rely on the software to come up with a two X throughput or whatever.

But we do see that these AI workloads are driving that data center energy demand through the roof.

And what I'm kind of saying is whilst you've said that maybe

Showing 381–400 of 856 · page 20 of 43