Sensors Converge 2024 - A 40th Anniversary Celebration (with Prakash Sangam)

Sensors Converge celebrated its 40th year as a fixture and iconic event that has tracked the history of sensor technology in the Silicon Valley. This year was a comeback year as the even has worked its way out of the pandemic nadir that has challenged the conference. This year, event drew over 5,000 from around the world attending one of the few sensor and MEMS conferences in the epicenter of tech innovation, the Silicon Valley.

Prakash and Leonard cover the following topics and highlights from Sensors Converge 2025:

• Prakash provides is impression of Sensors Converge 2025
• Is IoT back at Sensors Converge?
• The influence of wearables and the search for the next big device
• Edge AI takes center stage bringing the excitement of AI to the sensor edge
• Trends in sensor and perception system architectures with edge AI
• The digital twin and physical AI and its impact on the sensor discussion
• Wireless made a Wi-Fi HALO heavy presence
• Trends in wireless changing and energy harvesting
• Leonard invents "Long-thinking Sensor Data Fusion"

Hit both Leonard and Prakash up on LinkedIn and take part in their industry and tech insights. 

Check out Prakash Sangam and his research at Tantra Analyst at www.tantraanalyst.com. Also check out the Tantra's Mantra Podcast. 

Please subscribe to our podcast which will be featured on the neXt Curve YouTube Channel. Check out the audio version on BuzzSprout or find us on your favorite Podcast platform.  Also, subscribe to the neXt Curve research portal at www.next-curve.com for the tech and industry insights that matter.

Transcript

0:06

Next curve.

Leonard Lee: 0:10

Hey everyone. Welcome to Next Curve, rethink webcast slash podcast. I'm Leonard Lee, executive Analyst at Next Curve. Joined by my good friend and fellow industry and tech analyst, Prakash Sang. Tantra analyst. Hey, how's it going man?

Prakash Sangam: 0:27

Pretty good. Glad to be here. You are looking good. Looking back.

Leonard Lee: 0:30

Looking good, Prakash. And in this episode we're gonna be talking about and recapping, an event that is near and dear to our hearts, sensors converge. Which, celebrated its 40th anniversary. Can you believe that? That's amazing. So, congratulations to the Questex family. The folks at, fierce Electronics, as well as, the folks at Sensors converge, the team that puts this event on every year. Yeah, and have put this invent on for the last 40 years, which is quite incredible. Indeed. Yeah. Right? Yeah. So it's great to be part of that legacy, it was a great event this year, before we get started, please remember to like, share, react, and comment on this episode. And also subscribe here on YouTube and on Buzzsprout to listen to us on your favorite podcast and platform. Opinions and statements by my guests are their own and don't reflect mine or those of next curve. We're here to provide a. Open an open forum for discussion and debate on all things tech and sensors and edge ai. So, with that, hey, let's get started. What'd you think?

Prakash Sangam: 1:50

I think it was good event. This is my fourth, fifth time, attending this event, I was there last year and compared to last year. I think there are more people. I think it started a little bit slow and I got in very early. it was a little thin, but I know it grew pretty rapidly throughout the day. And I left, around four o'clock? On the last day?

Leonard Lee: 2:12

Yeah,

Prakash Sangam: 2:13

Usually on the last day. I mean, by that time, you know, people are packing and leaving. I saw a lot of people around the, boots. The boots were up. Exhibits were up. Demos were up. It was pretty attended, I think. I think it's like 5,000 or something. They said, the attendance.

Leonard Lee: 2:29

Yeah. It was up like 18% according to them, registration last year.

Prakash Sangam: 2:33

Yeah.

Leonard Lee: 2:34

And

Prakash Sangam: 2:34

also I thought. That it was more diverse in the previous years. It used to be in focused smoothly on sensors, but I thought I saw different things this year. Of course, tons of sensors, but also a little bit more on connectivity. Of course Ed ai, which is, but there was more talk of Edge AI this year than last year, for example. I saw more connectivity, which I was not expecting, of course. That's interesting. Two years ago there was more connectivity because five, 5G was going, G was talking 5G or the not showing. But this year I saw Halo, which is a wifi, long range wifi for sensors and IOT was there.

Leonard Lee: 3:20

Yeah.

Prakash Sangam: 3:20

And there was, for sure, more iot. Yeah. And last year I think iot word was suppressed, but this year. I saw a few iot and, people talking about mentioning iot.

Leonard Lee: 3:32

Yeah. It's funny because it all just ends up being IO ot. It's just iot is this really amorphous idea, right? As a concept. It's not the whole collection and like ai, right? Everyone's like, oh, do you have an AI strategy? It's well, what do you mean by that? You know? Yeah. And then, typically they can't explain themselves, but that's a different. Problem. But IOT has the same problem. If you ask, what if you ask them, what is iot? Then the Only proper response is something incredibly simple. And then it's just an idea, you know? but yeah, it all ends up being, I iot in some form and I agree with you, but it is curious that you saw so much connectivity. I didn't see as much connectivity as I saw in the past, but you rightly mentioned that. in years prior, there was probably more present due to, private wireless. Yeah. I think that about two, three years ago, that was a big thing. you had the 5G, and V two X These macro outdoor, themes that, was reflected in the, exhibition hall. I. Uh, I just, I thought there wasn't that much, but that's an

Prakash Sangam: 4:52

I saw. Couple of interesting, energy harvesting companies. I mean, it's all part of IOT as such.

Leonard Lee: 4:58

Yeah.

Prakash Sangam: 4:59

So those were interesting. And also since I don't cover sensors that much, so maybe those look more interesting to me than the same, everybody has their, gas sensor or some other kind of a sensor.

Leonard Lee: 5:11

Yeah. A lot of mention of, I amus. Right. so, I. You know that it, it seems like the wearables market is, we saw a little bit of it last year, but the wearables market, it seems to be, influencing a lot of the conversation. So when you sat in the sessions, the technical sessions, there was a lot of mention about. novel, personal devices, whether it's for medical devices or for personal computing. Right. Whether, it could be, well, a lot more talk about smart glasses.

Prakash Sangam: 5:54

Yeah. Not so many levels though. in the previous years there are more the talks. Off, but demos, like two years back, even last year, I think almost every other, exhibition had some sort of a glass as a demo, right? I did not see that. not even one. demonstration of, AR or XR glasses. So I was actually surprised, was expecting more. to your point edge devices used to be like, I saw some, I went on, talking to a couple of startups. Basically they had a variable, With, fall detection and such in an extremely small, very light, kind of a pendant. And it had some AI for detection and so on. So it was not just, based on a sensor, but more intelligent sensing rather than just Sense and so on.

Leonard Lee: 6:46

Yeah, I think the scale down, so, you know, like in ai, supercomputing, it's all about scaling up, right? Correct. And then, you know, there's the scale out aspect with networking, but Correct. In the sensor world for ai, there's scaling down, right?

Prakash Sangam: 7:01

Correct.

Leonard Lee: 7:01

Yeah. I think there's been a lot of traction there because a few years ago. ML was just being introduced. pretty close to the sensor. It was not quite on device. That was still exploratory, but now you're starting to see more of these companies actually embedding. AI functions And augmentation on the sensor device itself. Right. it's either co packaged or packaged with the device. it might just be a peripheral, but it's very, very close. To the actual device. So the architecture seems to be changing around smart sensors, correct?

Prakash Sangam: 7:43

I think there was discussion of smart sensors as well, right? But the smartness is again, now replaced by ai. And I also saw many of these. You know, for the sake of ai, but they were doing specific things which needs that smartness.

Leonard Lee: 8:01

So application specific, but then like what I was mentioning earlier about the scale down

Prakash Sangam: 8:06

Yeah.

Leonard Lee: 8:07

Scaling down, I heard some. Guys on the floor, a couple of companies talk about how they were able to run a model that was only a few kilobytes. very application specific. we used to call these algorithms, right? But the models themselves have a very, very small footprint, but are extremely capable. Yeah.

Prakash Sangam: 8:30

And that's kind of cool thing. And just to make sure that our audience understand it correctly. This is the place where, you know, census converge, where AI is not LLM, right. These are. Yeah. Yeah, yeah, yeah. Right. So ai got equal to LLM there, right? These are all, yeah. There were some, some I saw, which. We are using, small models SLMs, but none of them are using LLMs. It was really the legacy AI or with their optimized their own proprietary models.

Leonard Lee: 9:04

Yeah. They're still doing a lot of the more conventional sensor fusion Right? So if you look at, let's say, a perception. System. It's more of the old school stuff where you're aggregating data at some point from multiple sensors and then synthesizing, a little bit further away, and maybe even really far away from the sensor. the automobile is a great, example of that. Right. We see small language, maybe SLMs in the band. That's still even really early, right? yeah. But we are seeing more of that application specific machine learning that's moving us closer to the. To the, sensor to reduce the amount of data. Mm-hmm. You know, and then introduce efficiencies that way through the rest of the system. You know, it's like how do we tackle the fact that most, serialized data or streaming data, coming off of sensor is just the same stuff, right? Yeah. how do we, instead of having all this data. Streaming having to be processed. How do we just nip it, that volume very close to the sensor and then, make the system more efficient and streamlined in terms of, the throughput requirements and latencies. Right. So, yeah, it's really interesting this, it's a different universe. It's different, well, actually it's different, but similar to ai, super computing.

Prakash Sangam: 10:32

Even when AO is not talked about, right? Yeah. Like edge intelligence is basically same, right? Yeah. You don't send all the tons of data just blindly. you only send what is relevant and then have intelligence edge, I think was, it was about before a GI came with the picture, so,

Leonard Lee: 10:51

yeah.

Prakash Sangam: 10:52

Yeah. Good.

Leonard Lee: 10:53

You had a panel, was it a panel you interview? You had an interview, right?

Prakash Sangam: 10:58

I had a fight side chat and a, round table as they call them. Mm-hmm. So the fireside chat was, with Tony of E five networks. It was on the, test and, measurement stream. It was well attended. I mean, I was surprised that there, there were enough people there, that they were interested in test and measurement and the log. Tony, from E five Engineering. he a fascinating guy. He basically had, his ex-military, so he had tested, tanks as well as he worked in Tesla for testing. remember the famous, steel ball hitting,

Leonard Lee: 11:35

On the cyber truck. Yeah.

Prakash Sangam: 11:36

cyber truck. And he says he still has that ball with him. He was involved. That testing. Fascinating guy. So we had interesting, discussion how, testing has evolved, before AI and now, with AI now. And there is basically, there is discussion on whether simulation can, yeah. There's so much of it, simulation going. Can it ever fully replace physical testing, obviously, the gist of the discussion was, Of course now there's tons of data which was not there before, which means you can do much better simulations. And, earlier everything had to be physical, testing before you can even understand what is wrong or how it is performing now. You can do quite a lot. With simulation and also we have so much data, so you can do much more analysis before going to the next step. Testing. Yeah. Derisk stuff. There's still testing of course is needed. It'll never replace physical testing, but it is more of a conformance testing kind of a thing, and confirmation testing, if you will, that. You very well, and I have, you have postulates and hypothesis. You very well understand and you are testing it for a specific thing rather than let's collect the data, analyze it, and look at it. That kind of stuff. So it was, Interesting discussion.

Leonard Lee: 12:54

Sensors are interesting in and of themselves for, in terms of the maintenance of. And the calibration. Yeah. Sensor devices. So, you know, it's interesting that you need sensors to calibrate sensors.

13:09

Yeah, yeah.

Leonard Lee: 13:10

Test them and then yeah, they're used, to provide that. Grounding. I also had a panel discussion that I was moderating and it was, AI plus, sensors, right? And so, it was a pretty cool discussion around, how AI is influencing or, impacting what we can do with sensors, right? And then vice versa, how sensors are, improving or impacting what we can do with ai. one of the big themes that I. Picked up on at sensors converge is this whole notion of digital twinning, right? Aligns really well with what you mentioned. Yeah. Simulation, because yeah, simulation is one thing that you can do. But what is required is to have, the sensors that can actually capture the physics of a particular, scenario or environment or an asset. And then be able to provide that fine tuned, grounded data set or augmentation Of what you have, whether it's synthetic or, sourced from. Other, ground truths that you somehow acquired. but the role of the sensor becoming increasingly important. And so I think what we're gonna probably see next year is all this interest in simulation, especially as you hear all these folks talk about all the wonderful things that AI supercomputer can do. That's going to provide some gravity to, this growing interest in sensors and perception systems. because I think that is becoming more of a thing, especially as the topic of physical ai. starts to, enter the hype cycle.

Prakash Sangam: 15:08

Correct. I mean, the digital twinning. I mean we talked about in the, obviously test and measurement is a key input to building a digital twin. I heard about in uh, panel and couple of others as well. I was surprised that's not so much of demos of digital twins on the shop floor. Right. Maybe because lot of the, demos were from smaller companies who provide, since it's not the full system where you need Twins and such. But it was some from

Leonard Lee: 15:40

Q and x, there were a few, but no, I agree with you. It's not the holistic digital twinning. Application or scenario, it's like little bits and pieces of it that, if you looked at all of these examples in its totality across all the different stands, then you, you're really talking about digital

Prakash Sangam: 16:01

Correct? Right. So for example, there was NXP Bo, they had. You know, reasonably sized booth. I mean, they are the ones for burning the systems.

Leonard Lee: 16:11

Actually NXP, that was like the smallest NXP booth that I've ever seen.

16:15

Because they like this

Leonard Lee: 16:16

massive booth and I think NXP. Come on guys. Get a bigger booth.

Prakash Sangam: 16:20

Yeah. It was their show in a way. Right. I mean it, and they had some presence in the sessions, but not that gray that much, I guess. So. Yeah. And I would expect them, like Bosch could have really put the digital twins that they're working on, which I'm sure they are, and how these sensors feed in and make that, digital twin possible. Right. That could have Cool. demo and illustration of. The importance of sensors

Leonard Lee: 16:49

that we might see that next year, because that's what I'm saying. I think, yeah. That narrative being stitched together this year, especially as the,

Prakash Sangam: 16:59

If they're listening, then maybe that's the expectations to put on them. Right. And exp bo. And other larger systems players. Yeah, I think the market is expecting, no good, digital twins, not just stocks showing the digital twins of the systems that they're working on.

Leonard Lee: 17:16

Yeah. And so, just a real quick shout out. So my session was called ai. And sensors exploring a special symbiosis. And, the folks that I had on my panel were, Manuel Koni, who is the pro product marketing manager at St. Azita Arvani, our friend, you know? Yeah, right. Yeah, yeah. Yeah, she's a good buddy of ours. Former CEO of Rakuten Symphony, who's doing her own thing with ai. Simon Ford of Belcon. he's a really cool guy, really smart dude. formerly of ARM and, Yuhan Long, who is the founder and CTO of Infra Move. And then Rob Watts, who is the AI architect, one of probably many AI architects at Intel. Yeah, he's a very funny guy. And, I really enjoyed, moderating this panel.

Prakash Sangam: 18:09

Yeah. Sorry. You saw it. Yeah, mine was overlapping, so I only saw the first part. I allowed how, I know I forgot his name. Intel Architect introduces a, yeah, Rob Wa

Leonard Lee: 18:21

Yeah. we invented a new term, it's called Sensor Smushing. So. Yeah, write that down everyone. Sensor swooshing. It's going to be a big theme for sensors. Converge 2026. So, any other takes?

Prakash Sangam: 18:39

Yeah, so I think, I mentioned in the, first part the, wifi has not new, not many people know about it. A long range version. Basically for not just indoor, for outdoor use, right? Yeah. For, medium, speeds. It's called Halo. It's 8 8 80 0 2 point 11. Ah, I worked on it long, long back ago when I was working for Qualcomm. I was, excited to see that, there are many players, all of them talk about it. They contributed to the standard. But they are not announced any product as such, you know, Callcom, Broadcom, and others. But I saw a few companies which have products in the market now, their initial pioneers where, NeuroCom is one, I don't work for them, but, I, brought some of them on my shows and so on. They were the only one. It's a Korean company, but there is, mega chips is building, modules for them, which I went and talked to them. And the many use cases there is a company which manages, the, waste management. They're using, halo, modules and halo systems.

Leonard Lee: 19:45

okay.

Prakash Sangam: 19:45

And there are. Products from few companies. Murat. Murat also has a mur Yeah. Product of, module. And there's another company called, I forgot its name. So Morris Micro has a chip set for Hello?

20:00

So.

Prakash Sangam: 20:01

And, I think, it's still, smaller companies doing it, but there is more traction. a lot of people are coming in asking about it as well. I was at, WCA Booth. Yeah. And, booth, for own as well. I'm on the board of WCA and a lot of people coming in, especially from medical field, asking for long range, secure. Wifi or any technology that is not Bluetooth or, ZigBee because there are tons of them and they don't, want something other than Bluetooth and ZigBee. So yeah, that was, that was a revelation for me, going, wow. Yeah, yeah, it was that big before because you only get to hear from the big guys and then never talk about, 11 or Halo. So,

Leonard Lee: 20:45

yeah.

Prakash Sangam: 20:46

Yeah. and energy harvesting. There's a lot of work happening on the ON 3G PP as well with ambient iot where you know, you basically harvest the RF energy so that you have extremely low power devices without batteries. they can be there anywhere, like environment sensors and so on. Yes. I saw a few companies who have them and then there is wireless power is, coming back. We worked with a couple of companies, so there are some good demonstrations apart from just wireless charging for phones, which is mainstream now. But there is wireless power within, rather than just this induction charging. There's something called, magnetic, resonance charging, which is, remember long back Apple had this, announced this product and they could not do it. Like I have a tray on which I can throw all my devices

Leonard Lee: 21:34

yeah,

21:34

yeah, yeah.

Prakash Sangam: 21:35

So there, the technology exists. Actually, Qualcomm also worked on it. when I was with Alcom. We worked on it long back, and then ultimately they sold that technology to a company called Vitro City or something like that. I saw a couple of demonstrations of, that, magnetic resonance, wireless power, including charging and so on, which will be interesting for iot, right? So, which means you don't need a battery at the device, so based on the RF that is in or your. Giving RF beams to these devices. They capture rf, and make energy from it, and then power themselves. So,

Leonard Lee: 22:11

yeah. that's a whole interesting feel for sensors, right? yeah. Yeah. It's gonna be interesting how all that transpires because you have the cellular flavor, you have the. wifi flavor, there's all these different, wireless technologies that are Yeah, Laura. Yeah. Yeah,

Prakash Sangam: 22:29

yeah. So iot and B, iot and obviously you don't see much of that at this show, the seller folks. But I did meet a person from T-Mobile. Mm-hmm. Basically, I had a quick chat with him. He was. visiting all the booths trying to sell T-Mobile cell iot for these sensor folks. So I hope they do it more. Uh, I did not. Yeah.

Leonard Lee: 22:53

I, yeah. Well, I think this year the focus was more on, AI because it's such a hype thing at the moment. And the folks at Edge AI Foundation were there, so, correct. Yeah. They, Peter's crew so I've had some chats with them about, where's the state of generative ai, because everyone's wondering, right. Obviously you mentioned SLM before. I think maybe next year we're gonna be talking more about, new. Perception architectures. And so I coined this thing called long thinking sensor data fusion, and it was really coined as a joke. Yeah. But I think it might, yeah, it might actually become a thing. I might actually speak on this next year.

Prakash Sangam: 23:40

you mean

Leonard Lee: 23:40

by long thinking sensor data fusion?

Prakash Sangam: 23:43

What do you mean by that?

Leonard Lee: 23:44

well, you have, reasoning models, right? They do. long thinking, so the Syngen framework and bringing it closer to, sensor data so that you can fuse the data. I see. And then get more, contextual information out of it. More of the. situational, awareness capabilities that you might want to get out of a perception system. So, yeah, we'll see. I kind of battered around the idea jokingly with a bunch of folks, and then I just, this popped into my head and I shared it with them and they laughed and, there you go.

Prakash Sangam: 24:17

Yeah. Attended a session by Edge. AI Foundation, the use cases were very. Interesting and real life, right? Oh yeah. so the sense I got is we talk about ai, and abstraction, but this is where gi basically is happening and that there tons of shall ready use cases that, companies get work and work on and make money on his definite needs.

Leonard Lee: 24:41

Yeah. Pete Bernard is a member of IOT Coffee Talk, so he. Does not have any latitude to exaggerate and be hype all hyped up. So, yeah. that organization is gonna keep it very, very real.

Prakash Sangam: 24:57

Yeah.

Leonard Lee: 24:57

So under his leadership, I'm really excited about Edge AI Foundation. I'm really glad to see that there's that collaboration going on. And, I expect that they're gonna do some crazy stuff next year. So I'm looking forward to sensors converge next year. I think it's gonna be a really good one, and it'll probably be a lot bigger because this is a hot area, right? Whether people know it or not right now.

Prakash Sangam: 25:20

Yeah, yeah, yeah, yeah, yeah. Yeah. I think the biggest issue is they're huge players in sensors, which they don't have to come to us in a event like this. Rest of us are pretty small players, right? So. I think, that's wherein maybe big Blak should come and help Questex make it even bigger.

Leonard Lee: 25:42

Yeah, the thing is the big players are in the audience, so if you bump into some of the people, yeah. The customers of all this stuff are really big players.

Prakash Sangam: 25:51

They're for information and new kinds of sensors and such. I think they should, Come and spend a little bit and make it a bigger platform. Then they can bring in even more people that they can

Leonard Lee: 26:01

Well, I'm on the board of directors and so one of the things that I shared was this idea of doing something similar to TM forum's. Catalyst. and then that might bring some of these larger players that you're talking about

Prakash Sangam: 26:12

Yeah.

Leonard Lee: 26:13

To the conference, to work with some of these. sensor companies as well as AI companies and systems integrators to prove out. And maybe even develop some joint, IP around, advanced perception capabilities. And, these things will obviously, push the frontier of what you can do with, autonomous vehicles. Yeah. Smart glasses, blah, blah, blah.

Prakash Sangam: 26:39

So, sim simple. If you take, uh, you know, agricultural. systems as a market, most of it is driven by sensors, right? Yeah. But there, there was no Honeywell, for example. There were no John Deere there. None of the, you know, rain Bird. I mean, irrigation systems are so much of sensors. Yeah. I was surprised that none of the big guys are there. I mean. Yeah. I mean, a lot of times I talk to them, they have their own events and then they say, why should we come here? But it is that you grow the industry by collaborating on a forum like this. Right.

Leonard Lee: 27:15

Well, one of the interesting things about sensor verge is their location.

Prakash Sangam: 27:20

Yeah.

Leonard Lee: 27:20

A lot of the big customers are located there, so it's kind of cool. It's unique in that way and that's why it draws an international crowd. So you look at the exhibition floor, there's companies from all over the world. And even though, it's not like embedded world. it still draws a lot of interest across the sensor

Prakash Sangam: 27:39

I think majority of the people attending our engineering engineers as well.

Leonard Lee: 27:43

Yeah, it's totally technical and nerdy

Prakash Sangam: 27:46

Talk to people on the. Usually when you go to larger events, right? More Congress, CES and others. When you go talk to people at these booth, their knowledge is very shallow because they're mostly marketing folks. Or higher talent, right? But here you go, talk to somebody, talking to somebody. engineers working on that. Product. Right. So that's I think one of the key values. A lot of, don't, a lot of people don't realize going to the show that Yeah, working on the product that you are, touching and feeling.

Leonard Lee: 28:16

Yeah. But I like to provide that opportunity to highlight the great work that whole community does, the sensor industry folks do, because it is amazing and I think it's far underappreciated. But, Hey, thanks. Sure. jumping on, it's really great to get your thoughts on sensors converge and hey everyone, thanks for joining and making it this far. I would highly recommend that you reach out to prakash@www.tantraanalyst.com. He has his own podcast called, tantra's Mantra, which is great. I show up on it occasionally and we usually argue, which is really fun. And so if it is not for information and education, you might want to just listen in for like entertainment. It's like we'll come back for wrestling of tech. Yeah. And yeah, con contact him on LinkedIn and, get ahold of his reports and, definitely subscribe, to his newsletter. And, ask him about his services and also, make sure to subscribe to his podcast. And, please subscribe to the next Curve YouTube channel. And the easiest thing to do is subscribe to the next curve, a research portal@www.next-curve.com for the constant diet that you need of the tech and industry insights that matter and, PR precaution. Once again, thank you so much and congratulations. Sensors converge on a 40th

Prakash Sangam: 29:46

Yeah. Anniversary. Yeah. Congrat request tax. Yeah. It's a great effort indeed. And really love the show. Take care. Bye