Edge AI Foundation at Sensors Converge 2026 (with Pete Bernard)

I had a great time hanging out with Pete Bernard, CEO of EDGE AI FOUNDATION at Sensors Converge 2026 which took place at the Santa Clara Convention Center this week. We had a chance to chat about the key trends that we see shaping what I've been calling the Perception Edge. What is it?

Simple. Sensors + Edge AI. Check out our brief conversation where we touch on the implications of this convergence of sensor technology with edge AI, the possibilities created, and the critical mindset shifts we need to make to appreciate and capitalize on the change and opportunity that is coming at the Perception Edge.

Transcript

Copy Transcript Copied

Leonard Lee 0:00

Hey, everyone. This is Leonard Lee, executive analyst at Next Curve, and I am here at Sensors Converge- Yes in Santa Clara, California. This is, like, I think the 41st- 41st edition of this event. It started

Peter Bernard 0:13

in, 1893, I think.

Leonard Lee 0:15

See, I think so. I think so. And I'm here with my good buddy-

Peter Bernard 0:19

There you go

Leonard Lee 0:20

Pete Bernard, who is the CEO of the biggest-

Peter Bernard 0:24

Hmm

Leonard Lee 0:25

biggest tech theme and potential hype that has- Hype ever cruised through- the Santa Clara Convention Center, and that is...

Peter Bernard 0:35

Edge AI.

Leonard Lee 0:36

Edge AI. AI, AI, AI. And he... Yeah. And yeah, yeah, yeah, yeah, yeah. And, uh, Pete is the CEO of the Edge AI, AI Foundation- Yes which is, like, going gangbusters, right?

Peter Bernard 0:47

Yeah, it's been great.

Leonard Lee 0:48

Yeah.

Peter Bernard 0:48

It's been

Leonard Lee 0:48

great. It's pretty amazing.

Peter Bernard 0:49

It's been a great show. Yeah. It's my first time actually at the show. We were here last year as a community, and this year we have lots of new partners here and, a lot of good traffic. We got Murata here, TDK- Yep STMicro, Microchip, plus we've got this pavilion of partners. And yeah, Edge AI is obviously a theme, yeah. It's a theme. A

Leonard Lee 1:08

theme.

Peter Bernard 1:08

Sensors, the s- the sensor edge-

Leonard Lee 1:10

Yeah, yeah

Peter Bernard 1:11

is a big deal, and what you do with, applying AI and Edge AI to the sensor data- Exactly people are doing some pretty cool things.

Leonard Lee 1:17

Yeah, and stuff above it as well, right? Sure. It's not just- Sure it's, I think- When you talk about edge AI, which, from my observation at this conference and other conferences is still a squishy thing. Mm-hmm. It's kinda like edge computing. People after all these years- Yeah, it's, uh- are still wondering what edge computing is. It's an

Peter Bernard 1:33

inclusive term. We call it an inclusive term. Yeah,

Leonard Lee 1:35

inclusive. But y- to your point, it's a concept. Yes. It's an idea. It's an approach. And, I think it's just fundamentally a counterpoint to, that prevailing n- narrative of cloud computing, everything going to the cloud, right? Now we're talking about things actually moving much closer to the sensor itself. And then the sensor itself becoming more cognitive than- Right, right it was in the past.

Peter Bernard 2:00

Yeah, the idea is,, put the AI workload as close to where the data's created as possible- Right so that you get the lowest latency, lowest cost, maximum impact-

Leonard Lee 2:08

Yeah

Peter Bernard 2:09

and easier deployments. Yeah. And so we're seeing, like, uh, like, machine learning being applied to sensors to extract more signal out of the noise, which is pretty cool. Yeah. So they're getting- Yeah a lot more sophisticated. And then, uh, more applied almost like, small language models being applied to that sensor data to get more context-

Leonard Lee 2:27

Yeah, exactly about what's going

Peter Bernard 2:28

on. Exactly. So you're seeing, levels of AI being applied-

Leonard Lee 2:31

Yeah

Peter Bernard 2:31

to the sensor edge. And like you said, it all ladders up to some pretty cool-

Leonard Lee 2:35

Right, right

Peter Bernard 2:35

solutions. Yeah.

Leonard Lee 2:37

So yeah. Yeah, so coming out of this conference, I coined this thing. Actually, we've been talking about it on IoT Coffee Talk for quite some time- Mm which is the perception

Peter Bernard 2:44

edge. The Turing test? Oh,

Leonard Lee 2:45

the perception edge. Oh, no, that too, the Turing test. We did invent a new Turing test. There you go. So if you guys are wondering what that is- Go look

Peter Bernard 2:51

on YouTube

Leonard Lee 2:52

yeah, it has something to do with methane. Um, human-generated methane.

Peter Bernard 2:56

There you go.

Leonard Lee 2:57

And, yeah, and intelligent, cognitive sensor- Yes detection of-

Peter Bernard 3:03

All that good stuff

Leonard Lee 3:04

such... Yeah. Um, human-generated-

Peter Bernard 3:05

But yeah, perception edge-

Leonard Lee 3:06

Yeah

Peter Bernard 3:06

is a good way to describe- Yeah what's happening in this space.

Leonard Lee 3:10

Right.

Peter Bernard 3:10

Right?

Leonard Lee 3:10

Right. Yeah.

Peter Bernard 3:11

It's not just about measuring- You know, so traditional sensors are like, you detect something and you- Yeah when you reach a limit it goes bing or whatever.

Leonard Lee 3:18

Yeah, exactly.

Peter Bernard 3:19

And but now it's really how do you perceive what's happening in the real world?

Leonard Lee 3:23

Right, right. Right. Right. And-

Peter Bernard 3:24

Which is tough

Leonard Lee 3:24

and so, Karthik, from Apple- Yeah his, uh, last name escapes me, so Karthik, don't kill me for forgetting your last name, but he's one of the board members of, Sensors Converge. I think he put it really well. It's taking things from just sensing-

Peter Bernard 3:41

Mm

Leonard Lee 3:42

toward understanding. Right. So understanding what's happening at the periphery of your intelligence system.

Peter Bernard 3:48

Yeah,

Leonard Lee 3:49

yeah, yeah. Right? And, and, and so going from sensing to perception, right, and then infusing that cognitive element to it. And so what you guy- uh, like a lot of your partners here- Yes and we have, your friends from EnaTerra- EnaTerra's here and Brain Chip- Brain Chip and EMASS and- Seba, EMASS, all these guys are working on, technologies, in particular silicon- Right and, devices that Enable that, cognitive evolution of the sensor- Right, understanding the sensory system.

Peter Bernard 4:19

Yeah, and that means you can take action on that understanding faster.

Leonard Lee 4:22

Yeah, yeah.

Peter Bernard 4:22

Right? Because you're at the edge, you're not sending it to the cloud, and then the cloud's like, "Oh, hey, there's gonna be an explosion," you know? Like- Yeah now you can actually, take action a little sooner.

Leonard Lee 4:32

Exactly. Uh- Exactly, and so this is where all the connectivity folks need to really pay attention. Mm. Because there's this assumption that the AI is gonna cause an explosion of traffic, and I know that there's gonna be a lot of people who don't like what I say right now- Right but that may not be the case. The whole point of edge AI may be actually to reduce the amount of data that's being- Yeah raw data, really, what it is- Raw

4:58

data

Leonard Lee 4:58

boiled down

Peter Bernard 4:58

to. Definitely raw data. Right. You're not sending, video feeds up to the cloud to analyze them, for sure. Yeah. But I would say it's like mitigation, mitigating communication risk-

Leonard Lee 5:07

Yeah

Peter Bernard 5:07

in deployment is what people are interested in.

Leonard Lee 5:09

Yeah.

Peter Bernard 5:09

And that simplifies deployment, it reduces the cost- Yeah and therefore you proliferate solutions- Yeah a little faster-

Leonard Lee 5:17

Yeah

Peter Bernard 5:17

uh, if you don't have as much complexity in communication. But yeah- Yeah I mean, it's... As you can process more data, it's more like using comms for the metadata-

Leonard Lee 5:26

Yeah, yeah

Peter Bernard 5:26

as per the actual data, which is, you know-

Leonard Lee 5:28

Right, yeah it's

Peter Bernard 5:29

good. Which means you can actually do, use, like, LoRaWAN, you can use all kinds of, low bandwidth communication- Mm as opposed to high bandwidth communication.

Leonard Lee 5:37

Yeah. And I think it, it's the data traffic and volumes that, uh, don't become the key drivers of what is gonna be required in the future. It's going to be other aspects that- Yeah you have to really look at. But then that's where you r- have to follow what's happening with edge AI-

Peter Bernard 5:55

Yeah, yeah

Leonard Lee 5:55

and, uh, sensor technologies and how things are evolving at the- Right

Peter Bernard 6:00

perception edge. Yeah, and it's a different, Perception edge perception edge. TM. The, You heard it here first.

Leonard Lee 6:04

Yes.

Peter Bernard 6:05

But it's, you know- Literally it's interesting flip from IoT. So in the IoT days- Yeah we had sensors and other things sending data to the cloud.

Leonard Lee 6:12

Right.

Peter Bernard 6:13

Connectivity, blah, blah, blah. Yeah. And then pr- now we're like, actually- Yeah with edge AI you can actually do a lot of processing locally. Exactly. You can mitigate the risk, mitigate the cost.

Leonard Lee 6:21

Yeah.

Peter Bernard 6:21

Get faster, cheaper deployments. So I think that's gonna help- Yeah the whole IoT space, as it were.

Leonard Lee 6:26

Yeah, exactly.

Peter Bernard 6:27

The edge space.

Leonard Lee 6:27

Well, I think it's gonna flip it because what you're talking about is sending data up. Now when you look at what's happening with perception at the edge enabled by, multimodal, by the way, multimodal sensors-

Peter Bernard 6:42

Yes, multimodal sensors

Leonard Lee 6:42

or fused sensors, plus, multimodal, AI- you transcend data. You're not dealing with data anymore, you're dealing with higher orders of information. Right. So it's context. Yes. So metadata provides context, but then at some point it's just gonna be understanding.

Peter Bernard 7:02

Yes.

Leonard Lee 7:02

Right? So those signals are different. The way... The... What you're dealing with in terms of information is different. Yes. It's no longer... You're moving away from raw data- Right is really basically what it boils down to. And, that's still, I think, in tr- talk track around edge AI, that shift hasn't ma- Because everyone keeps talking about data, and- Yes You're not dealing with data anymore.

Peter Bernard 7:24

Not as much. Yes. Yeah. And actually, your point on multimodal sensors is interesting, 'cause we're getting closer to, the digital nose, right? Yeah. Where it's like people are using AI on the sensor signals to extract, what's actually being sensed.

Leonard Lee 7:37

Yeah.

Peter Bernard 7:37

So that's pretty cool. Going back to our methane comment,

Leonard Lee 7:39

that's a cool thing. Wow, yeah, the Turing, Turing

Peter Bernard 7:41

test. Turing test, yeah.

Leonard Lee 7:42

Turing 2.0.

Peter Bernard 7:43

Uh, and also we're seeing, like- For physical AI doing, vision without cameras, audio without microphones, right? Yeah. So you're analyzing radio waves and ultra-wideband and things to detect presence in a room and how many people are in there and stuff like that. Yeah. No cameras, so but you can use these... You're doing AI vision without cameras- Yeah and things like that. That's pretty cool, and we're seeing a lot of that here too.

Leonard Lee 8:02

Yeah. Yeah. In a way, these are becoming superpowers. They're like, like Marvel-

Peter Bernard 8:06

Like Superman, like the X-ray specs

Leonard Lee 8:07

Exactly, X-ray vision- Yes and all these other cr- crazy things. Maybe

Peter Bernard 8:10

that'd be good. It's, the smart glasses. Yeah. I would like them to be X-ray specs.

Leonard Lee 8:14

Yeah. Uh, I don't know.

Peter Bernard 8:16

Like, uh, we'll have the- I

Leonard Lee 8:17

think we would have a- the comic books ambient, as, Karthik calls it. An a- ambient privacy issue. So, hey, look that up. That is true. Ambient privacy.

Peter Bernard 8:26

That is true.

Leonard Lee 8:26

So there's a lot of things to think about here as we look forward to the future of, uh- Yeah the perception edge and the role that edge AI plays enabling that, and how all this, sensor innovation, i- is go- are... these things are gonna come together and- Mm-hmm Shape the new possibilities for intelligent systems, right? Yeah, for sure. More cognitive systems.

Peter Bernard 8:48

Sure. Um- And this place, even though it's 41 years in the making, this show is very relevant. There's a lot of new stuff here, so which is pretty cool.

Leonard Lee 8:54

Yeah. Yeah. No, I'm... You know what? I, I'm really glad, being a board member of, the, Sensors Converge, event here-

Peter Bernard 9:00

Mm

Leonard Lee 9:01

I've- I was excited to see that you guys were-

Peter Bernard 9:04

Yeah

Leonard Lee 9:04

making a presence here, and then, looking forward to-

Peter Bernard 9:07

Next year

Leonard Lee 9:07

seeing you guys deepen your presence, and then maybe do some of that converge stuff, right? Converge. The sensor converge with Edge AI to get to cognitive sensing and, uh-

Peter Bernard 9:20

There you go.

Leonard Lee 9:21

Yeah, cool stuff. So- Good anyways, hey, Pete. Good

Peter Bernard 9:24

to see you again.

Leonard Lee 9:25

Yeah. As

Peter Bernard 9:25

always.

Leonard Lee 9:25

Always, always. A pleasure. Yeah, and so, thanks for tuning in.

Peter Bernard 9:30

Yeah. Check out our YouTube channel for if you're really into the Edge AI videos. We have a new video every day there, the YouTube channel, Edge AI Foundation, or edgeaifoundation.org, of course, which is, the center of all Edge AI things-

Leonard Lee 9:42

Yes

Peter Bernard 9:43

to go to and learn about things, so-

Leonard Lee 9:44

Yes kinda cool. And then, you can also, get a taste of the metaverse and what it could have been, right? With all your content, you guys do a lot of the- Yes nano banana.

Peter Bernard 9:54

Nano bananas. Nano

Leonard Lee 9:55

banana stuff.

Peter Bernard 9:56

We're into nano

Leonard Lee 9:57

bananas. Yeah, yeah. No, it's really great. And then, remember, yeah, definitely, get involved into Edge AI Foundation. Great content, great focus, whether they know it or not, great focus on practitioners. They bring a lot of- Practitioners researchers in, but it's all about how do we now convert some of the-

Peter Bernard 10:15

Yes

Leonard Lee 10:15

research-level stuff, pre-commercialization stuff into solutions. Yeah. And so it's- I

Peter Bernard 10:20

would say also, too, speaking of practitioners, if you are in the European theater- Yeah as they say, in London, June 8th and 9th, we have some, we have Edge AI London 2026 going on, and we have workshop. We- we're world's first Ventuno Q workshop going on there- Ah, yes with this incredible NXP workshop with, their new A240 and, and talks from, the president of Avnet Europe is gonna do a keynote. Okay.

Leonard Lee 10:44

Now he's bragging.

Peter Bernard 10:45

I'm doing... Yeah, it's all this stuff. Max Versace- Yeah from Analog Devices. So- Yeah if you're in the London, Europe area and are into the Edge AI scene, and you wanna really soak in it for a couple of days- Yeah Edge AI London will be the place to be June 8th and 9th.

Leonard Lee 10:57

Yeah. And, next year, make sure if you haven't, ever attended Sensors Converge, make sure that you attend-

Peter Bernard 11:04

Yeah

Leonard Lee 11:04

next year. Physical AI obviously is gonna be a big topic, and the grounding is here. It starts with the sensors. And so- Yeah you wanna... you'll want to pay attention, and it's all about the perception edge going into 2027. So, uh- For sure and, these guys are at the epicenter of all that, Sensors Converge and, Edge AI Foundation. So we'll see you next year.

Peter Bernard 11:27

Sounds good. Bye-bye.