The AI Industrial Revolution: Securing the Factories of the Future

Hey, everyone. Welcome back here to Text Drunk tv. I'm happy to introduce you to my next guest on today's show.

His name is Samuel er, and I, I had a good, and I spoke and forgot it. Samuel, pronounce your name for us again, so we get it right. Samuel Pesky.

Pesky. I wish I could say it as well as you could. I just, I'll, I'll never be able to, I'll never be able to.

Rolling the r the whole thing. It just doesn't work for me. But Samuel, it's great to have you here on Text Drug tv.

Samuel is the, by the way, is the VP of product Management for Industrial IOT networking over at Cisco. And we're going to dive into what exactly industrial IOT networking means. But first, let's, let's find out a little bit more about Samuel.

Samuel. Give us kind of your career path, if you don't mind. Give people a little sense of who they're talk, who they're listening to.

Sure. So thank you so much. Thank you for having me here.

Uh, so I'm Samuel Pasky and my past is, you know, I grew up, uh, I was born in US, believe it or not, but I grew up in France, so that's a French accents. Mm-hmm. And, um, I, after my study in France, I studied, uh, software engineering and I moved to the Bay Area in California to work as a software engineer.

And I was lucky enough to work on Catalyst 6,500 for mm-hmm. The audience that have been in networking for a few years. Catalyst 6,500 was a, a big, uh, workhorse.

Yeah, exactly. So I worked for many years on this platform and then I made the move between engineering to product management. You know, you never want to change too many variable.

So I kept the product, but I change role. I became the product manager for this product, 6,500, which was amazing because I knew a lot about the technical aspect of the product. So it helped me to build trust with my audience, with the customer.

And then I would say the rest is this story. You know, I went from 6,500, became a manager, led the launch of the 6,800, then moved to a Catalyst 2K, three K, you know, the access, uh, switching platform at Cisco. And 10 years ago I moved to, uh, this thing called industrial networking, which I always thought was really interesting.

And you see it is been 10 years. Uh, I'm here and, uh, in initial networking and I love it, Love, love, love, love it. What a great story.

Yeah. The 6,500, you know, brings back a lot of memories. It was a, it was a, it was the backbone in many ways of a lot of networks.

Exactly. A lot of going on Samuel, when we talk about industrial IO ot, and you even say that so much nicer than me with your accent, but industrial iot networking, what exactly are we talking about? Yeah, so maybe let, let's, uh, you know, let's go through a story lesson.

You know, I know you've been familiar with Cisco, like a lot of the audience, Cisco has been building network for 40 years. What people may not know is Cisco has been building industrial network for 20 years. So it's pretty much building, uh, networking equipment to go into a manufacturing environment, uh, can go into a, uh, utility network substation, connecting the grid, or even all the way to, uh, I would say more heavy industry, oil and gas mining, those kind of things.

So what we do in industrial networking is really build the connectivity and security that you need to connect those meshing that are outside the office space. You know, outside of the nice climatized, uh, uh, room that we are, it's outside. It's where it's dirty, where it's, uh, and where you have different type of constraints in terms of networking, uh, need.

And that's what we do. And we do that cling the Cisco, uh, uh, DNAI will say, and the Cisco, uh, uh, knowledge and knowhow, uh, to bring that to the initial network. Very cool.

Um, you know, for those of you out there who maybe have not worked or dealt a lot with the, uh, industrial IO OT world, you know, you focused on the norm, the regular IT world. It's a big world iot. There's, you know, not to sound like Carl Sagan, but there's billions and billions of IOT devices that are out there in the world.

The more coming on every day. And, and every single one of them has a unique ip. They need a unique certificate for security or a unique id.

Um, and they're doing beyond mission critical life or death kind of functions. And, and, uh, you know, this is, it's not, I remember a friend of mine was the IT director for a hospital, and he said the unique thing about running technology at a hospital is you had three networks. You had the normal IT network that we all work on, right?

With computers and laptops and printers and and so forth, servers. Then you had the medical device network, the in, uh, the intravenous pumps, the heart monitors, the, the dialysis, the CAT scans, the MRI, the stuff that makes modern medicine, right? It's all networked, it's all connected.

And it's, uh, beyond mission critical, it's life and death. And then you had a network for the doctors. 'cause they don't follow any rules and their security risk anyway.

And we had to keep them on their kind of, their separate networks so they don't mess up everything else. But, you know, in a lot of places, industrial I OT is, as I say, life and death mission critical and even bigger in some ways than, than just, you know, traditional it. So, And you know what I, what I would say to echo to what you're saying is when you look into the industrial world, you look into factory, uh, you look at utilities, you, you look at all those space, there's more and more, uh, connected device connected to an IP network like you mentioned.

And those device needs to totally need to exchange with an increased amount of data. And the truth is, a lot of those device have been connecting for many years, but it's been done, uh, in a very organic way. And now with security being top of mind for a lot of people, you, you mentioned hospital, but the same in a factory.

You know, that puts some very interesting challenge, uh, for people who have to manage this kind of infrastructure. Absolutely. Absolutely.

com that they could go to to find more information there. But I want you to turn over to a recent report that Cisco put out. It's Cisco's 2026 state of industrial like AI report.

And, um, like, you know, like everywhere else, AI is moving from pilot into production, but as we said before this, these are potential life and death situations, right? And, you know, where we can't have cyber instances or downtime or safety failures. Talk to us maybe a little bit about the rationale behind the report, and then maybe we could dive into some of the key findings.

Yeah, sure. So to give some context, like I mentioned, we have been, uh, helping to build initial network for 20 years. And, you know, when we saw AI coming on the market and we, we were, first, we were wondering, you know, what's going on in initial network and what the trends.

So two years ago we did a report for state of initial network. And two years ago, uh, when we asked a thousand, uh, uh, people that are dealing with industry network, they told us that AI will be the biggest impacted trend for them in industrial in the next five years. So we are done two years.

We thought like, okay, let's focus on state of industrial ai. What are the use case? What are people doing?

How, what, how are they using it? Try to demystify that a little bit and really try to understand when we talk about impact on industrial network, what does it mean? What are the change, what are people doing?

And so on. So that's a little bit why we decided to do this report, to really talk to ask practitioner people who do and build and operate in industry network every day. What do you see?

What, what are the use case? What, what's going on with ai? Where are you using ai?

What are the limiting factor for you to deploy AI? And so on. So that's really, uh, why we decided to do this report and try to get insight.

And being a product manager, obviously I want to make sure that we're investing in the right direction to build the future of industrial network. Right? So that was really the intent beyond this report, and we found a lot of very interesting information, and that's why we published it.

Right. Excellent, excellent. Um, and you know, I, I look just from my own experience with I ai, it is gotta be a very different picture this year from last year, right?

Because we've made so much, we've made so much progress. Well, some would say progress, some would say not, but we, you know, we've certainly learned a lot about how, how AI can help us and what the things that it could do now, what are the things we're learning to use it for, et cetera. Mm-hmm.

With these kinds of reports. Samuel, I always like to ask, you know, what are the gimme the top three things, the three key findings from this year's report? Yeah.

So maybe, you know, we should, um, try to put some color on what does it mean AI in initial, maybe first, and I can explain you what the finding, because when we talk about AI in initial, you know, we, people may have a view of what is AI in the IT world, or what is AI in your personal life. You know, we play with, we, we still, and so on initial world, we really see, I will say three things opening. One is, uh, the explosion of machine vision.

So, um, machine using camera to see the world, right? And that change a lot because you go from a world where you sensor to a world where you have machine eyes and you can do things. So that puts constraint on the network.

The second thing that we see is, uh, you know, everyone is talking about physical ai, and the reality is you have robots, human doing or not, doesn't really matter. That are getting smarter and smarter to do things and are moving. So there is a concept of, you know, the mobility of those machine, of those robots.

How do they connect? How do they get the information? And when you combine the two together, you have something that is even more powerful.

Now you have robots that can move, that can be smart. They have a view they can see. And now the question is, where do you process all this information?

So you need to be able to connect to a place where you have unlimited CPU resources and memory to be able to process all this information to give back to the machine or the robots what to do. So that's in a very, very simple way. What we see in the initial world and out of the report when we ask people what the big finding is really three things.

The number one is the infrastructure gap, right? Uh, meaning, uh, you want to have more and more camera. See, the beauty of that is you can save things to learn on the huge set of data, right?

So you can have a deployment of using a camera for quality inspection on let's say your car manufacturer on one station, on your car. That's really good. Where it become very powerful if you can do that to a thousand of car a day.

Can you cor correlate a little bit of learning? Can you do learning from the volume of data? If you want to do that, you need to have the bandwidth, you need to have the storage, you need to be able to capture all this information.

Can be quality, can be a quality of the world, can be general assembly. Did I put the screw? Or I need to put the screw, those kind of things.

The second things that comes to mind is, so you need to connect all of that. But like you said, you know, everything has an IP address now, which means your attack surface is increasing. And what you don't want is in, in short, it's not, I don't want to lose control of my factory.

I don't want anyone to come and stop my factory. I may lose some data, I don't like it, but I may survive that. Stopping the plant is a no, is a no-no.

Right? Or if you think about u utility substation, you don't want to lose power, right? So the idea of security in that environment is super top of mind.

And that has been for a few years and we just accelerating and with now the new thread that can come from ai, it's even more top of mind. And then the last thing is, okay, so you need to have people who understand the initial environment you need to connect to build capacity. That I would say it's, it's okay, but you need to secure all of that at the same time.

And then come the problem of the skill gap. So I like to joke that, you know, you can always find the unicorn, someone that is very strong in industrial and understand the, the, the space and understand networking and understand security. But the truth is, uh, you need scale.

You need more people than you can really find with this skillset. So there's no other choice than having a collaboration between it, the, the people who know networking and security and more the operation team, the people who know the process, no industrial automation. No one needs to be done.

And, uh, what information to collect to be able to improve the quality. So that's really the, the third point is really this collaboration between the IT team and the IT team. So to summarize the issue, infrastructure readiness, do I have the, the pipe and the connectivity that I need?

Do I have the security that I'm worried about security? And last is skills. I need people, right?

So that's really the, the three big finding. And we hear, uh, uh, reading more and more, I would say not concerned, but we can see it top of line for, for people who reply to this, uh, report. Right?

I love it because it's a good way of looking at this report. You know, especially now with ai, so many things we see to kind of challenge our previous norm, right? What used to, what we used to take for granted is not grant taken for granted anymore.

And what we used to think of as, you know, dreams is reality. What about anything in the report this year, Samuel, that made you think, my god, I didn't see that coming, or I didn't see it coming that fast. I would say it is the other way around, if I may.

I, I, I, it is not, um, uh, you know, look in industry, uh, things move maybe a little bit slower than in the IT space, right? Which might be a good thing, you know? Mm-hmm.

Uh, you know, people go slowly, they want to make sure things are working, eh, let's not be crazy. Uh, so people are adopting technology, they are a little bit more risk drivers, but they move with a lot of, um, I will say inertia. You know, when they move, it moves they go, right?

There is no way down. The thing that was the most surprising for me is, look, I've been in 10 years in this world being an IT company, working in industrial network, and maybe I've been biased to think like this, IT OT collaboration is very opening and it's going on full swing. I was a little bit surprised that it is not as, uh, I would say prevalent as what I thought.

I thought it would be much higher in terms of percentage and what it is. I think people are moving. But when I step back, I think it's a lot about the culture change, right?

You need to have, and changing the the people is very odd, right? And, and I think that's the thing. So I, I've personally observed some customer who are very, uh, also leading edge.

I, I can think about, uh, uh, one utility customer where they build, uh, their team to handle their networking for the grid and for the team come from a service provider background. People who build very large scale R network and half for the team come from electrician. They put all of them together, they have 12 people.

Let's go and build the, the network for the grid of the future, which is very interesting. And you have some other customer who are a little bit more on the way back and kind of trying to fight it, but they cannot really fight it. I would say the writing is on the wall that you need to have those team working.

And I think that's where we are progressing. So I'm convinced, uh, this partnership will will increase significantly, but it's not at the level I thought it would be. So that's not your question.

Absolutely. Um, but again, I think the, the, there's more a front of us to come, right? And, and, and it so it will, but you're right, it does, you know, there's the hype and then there's the reality of these things, Samuel, right?

The hype is we're all gonna have robots in our house, and of course they're connected and everything else. I don't know if it's gonna happen in, you know, in three years, five years and who could look out beyond five years. And there's some of that in, in these, especially when we talk about things like IO ot, right?

Though we have billions, literally, I forgot what the latest number was. Was it six or 7 billion or was it 4 billion IOT devices connected in the world today? They estimate something like that.

Uh, and I'm not tracking, I'm getting confused of what IOT stands for, but, but it's, it's a big number more than I can Have. Yeah. It's a big number and, you know, and, and, and it's a hard number to kind of wrap your head around.

I mean, it's, it's just, it's incredible. But, you know, we'll, we'll have to see. So, and, and that, and it's not like, as you say, it depends what IO OT is.

It's not like all I OT moves at the same speed, right? Different areas of this different, different things. So, we'll, we'll have to wait and see.

Anyway, Samuel, for people who wanna download the report maybe and get their hands on it and dive in, what do you think is the best? Where should we send them? Well, so we, we have, uh, uh, maybe we'll pause the, uh, the link, uh, after the video.

Yes. com/go/uh, industrial, uh, networking. I think we have, uh, we have quite a few link over there.

And, uh, the report will be available and we'll post the link, uh, with the video so people can go and We'll put the video, we'll put the link in the show notes. But you know, there's still good old Google. And if you Google Cisco's 2026 state of industrial AI report, I bet you'll get it.

You Will find it faster as I'm trying to go and click on the link. You Remember that? Yeah.

Well sometimes, you know, you gotta do what the old school works. Best, Samuel, thank you so much for coming out here on Tech Trunk tv. We appreciate it.

Great report. Keep up the great work there at, uh, Cisco and come back and keep us informed. Okay.

Thank you. Thank you very much. Thank you for seeing me on.

I will be back. Thank you. My pleasure.

Thank you. Samuel PAs a vp, product management industrial IOT networking Cisco on the Cisco 2026 State of industrial AI report. Dear on text, drunk tv, we're gonna take a break.

We'll be right back.