Why AI Needs to Leave the Data Center for the Real World

Added: 2026-05-05

[00:00:00]be there to benefit the real world.

[00:00:03]And the physical world does not live inside of a data center.

[00:00:08]Yeah.

[00:00:08]>> Right? It actually lives out at the edge.

[00:00:11]So, knowing that, you know, knowing that this value exists out there, knowing that the people exist out there, there the people's devices, the autonomous vehicles that need to be orchestrated, all those real world things uh need AI inference.

[00:00:27]Now you can make smarter decisions around where that inference lies. Is this something needs to be on device because maybe the network isn't reliable and it and and I need it to be have that capability at all times.

[00:00:37]>> Yep. Is there value of having it at an edge location because autonomous vehicles is a good example because it needs to aggregate data at a metro level instead of at a global level. I don't want a a single data center that is doing that. And that also gives me a capability to do a couple of things uh you know, in this new world of inference, which is be able to react in near real time, you know, in in in less than 10 milliseconds rather than hundreds of milliseconds, which which can mean a lot. Yeah.

[00:01:05]Um and it also gives us a capability to um to uh not just react in real time, but to also reinforce the learning and hyper localize what I'm learning in real time as well. So, using things like federated learning capabilities, which, you know, has been around for years, but really hasn't been implemented at scale. Um but it's going to become more and more important over time as we think about where data lives, where data is created, and how we quickly get value from that data.