Edge computing

If I'm, you know, ballpark, there's two major different types of data centers.

There's data centers that are used for storage, memory, training, which can be in the middle of nowhere, but they're data centers that need to be near where we operate for inference and other sort of edge, edge compute, whatever they call it, edge compute data centers.

So, and obviously the one, the smaller ones that need that lower latency need to be closer to, you know, cities and, I don't know, whatever.

So maybe, Jamie, you can sort of shed some light on those claims.

And we're starting to see the same thing.

I'm already seeing developers, technologists in some of these areas realizing they don't have access to the compute, the chips, the data centers, the internet bandwidth, that they've moved to edge computing.

And so they're trying to make it that the intelligence lives locally on the device.

You connect from time to time to get updates.

And it hallucinates less.

So much so that there's a company I know in Hollywood started by a group of Hollywood producers and directors doing digital twins.

They shifted their entire AI stack into edge computing and they cut their costs by a little over 70%.

performance radically improved.

Optimize the use of the available capacity, make best use of it.

We see AI applications like Edge Compute where you have in, you know, AI edge inference, where you, you process your data on the edge.

You don't need all that bandwidth.

You basically just get the da- data you're really interested in.

So normally, if you're, say, an AI system, you're pinging back and forth, moving your data.

So would this act almost like an edge computing device? Like, would it be kind of edge computing capabilities is kind of where my brain went when I was thinking about it

for

your wearable or whatever that ultimate application was?

That's exactly right.

And to do edge computing, in other words, the computing directly on your wearable, it needs to be very low power consumption because your wearable is powered by a small battery.

So today, if I wanted to do the same function, so let's say you're monitoring your heartbeat on your wearable and you want to use sophisticated AI to detect the arrhythmia.

What you'd have to do is continuously send your data through the cloud to an energy-hungry data center where it would successfully detect the arrhythmia.

And people will just be managing agents that are connected to physical environments, that are connected to various IoTs.

And that will be various internet things like edge computing and so on.

That will be the future for facilities managers, what they will be doing.

They will be focusing much more of their work on the technology, much less manual physical work.

And we already are seeing a number of robots and co-bots.

Co-bots, of course, are cooperative robots in the facilities management space, in addition to all the signaling software and so on, Internet of Things, Edge Computing.

That will be absolutely connected by various AI tools, and it already is increasingly so that I'm in conversations with a number of startups that are very powerful in this space, and I can see the future is heading in that direction.

It certainly is an interesting time, and I'm excited to see what the future holds.

And so you have to be able to run these networks and these things.

And so we got really, really good at edge compute.

So building these really highly capable, highly sensitive for these mission critical environments, AI on the edge.

without internet connectivity.

providing centimeter level accuracy for reliably at a level they have never seen before.

We've done that through a lot of different things.

Edge Compute, which is another one of our, no one is better than Nordian at Edge Compute and just being able to build these Edge AI capabilities.

for a lot of the reasons I had mentioned.

It was down in Atlanta in the late spring, and we cover everything from what excites them about intralogistics, as well as trends such as why so many logistics companies are focused on standardization.

We talk about everything that's going on inside the four walls of a warehouse, robotics, edge computing, and more in today's episode.

So let's meet up with Joel and Markus.

[rock music] All right.

So that's really the charming thing about standardization is to be repeatable and to not only do it one time, but also overall it's really saving you operational costs in the end.

Yeah.

Well, let's, uh, dive into another area, and that's around edge computing in, in a logistics operation.

We've talked about robotics.