Judy Warner (00:01) Hi, Matt, happy new year. Great to see you and excited that you're gonna be my first guest for our DesignCon 2026 pregame series. So thanks again for joining. Matthew Burns (00:12) Judy, it's always a privilege to be with you. Welcome to 2025. And as always, I look forward to spending some time with you. DesignCon is always a great way to kick off the high tech, high speed year. So look forward to talking about what Samtech is focusing on for the event. Judy Warner (00:27) Yeah. Well, you know, I think I've said it before, but you're the one that gave me the whole idea about doing a pre-game DesignCon series because you said it was the Super Bowl of electronics. Matthew Burns (00:38) I didn't come up with that quote, but one of our Samtech teammates did. So I'd like to take credit for that, but I have to give credit where credit's due. Judy Warner (00:44) Well, wherever it came from, I latched onto it and that's how this became, you know, that we're pre-gaming. So I have a logo with people younger than us with solo cups pre-gaming. So that's what we're up to today. We're drinking from the fire hose, buddy, not a solo cup. Matthew Burns (01:04) No problem, I like it. Judy Warner (01:06) So let's talk about as we dive into the new year and dive into a new design con, know, you're on the front lines of technology speeds. You're talking to lots of people in the industry. So what are the technology trends and things as we go into this year's design con and what are you seeing as the technology drivers this year? Matthew Burns (01:31) No, that's a great question, great way to start out. And I would say at a high level, macro level, Judy, I've never seen innovation ramp up so fast. I've been in this industry 25 plus years. I'm starting to date myself, but that's just the reality. with everything that's happened in AI over the last couple of years, whether it's gen AI or inference or training, Um, just the entire ecosystem has just hit the acceleration pedal and coming up with new solutions looking, and it's not just product. It's, it's, it's innovation. It's R and D. It's how do I, how do I, how do I, uh, design my chips to go faster? How do I design channels to go faster? Uh, how do I process data faster? How do I cluster, you know, dozens, hundreds, thousands of GPUs to handle these, you know, ever growing, uh, AI foundation models. So that's really been the big thing in the industry. it's everyone's saying about it, right? It's the IC vendors, it's the semiconductor providers, it's the IP, it's EDA, it's test and measurement, it's interconnect companies like Samtech. That's really the biggest trend. And I think everyone's struggling to keep up. We're doing it because we have to. It's a... You talked to lot of people, this whole AI ecosystem is a once in a lifetime or once in a generation inflection point. But that's it. And behind that, it's obviously, you and I have talked earlier about how designs get smaller, faster, denser. And that's kind of an underpinning for what's happening with AI. that's been, I think, the biggest driving trend in our industry the last couple of years. And I don't see it changing. Judy Warner (02:59) Yeah. Matthew Burns (03:23) in the foreseeable future. obviously it can continue to evolve, but having more AI processing capability and access to the latest and greatest models is going to define the high-tech industry for the foreseeable future. Judy Warner (03:41) Well, you are saying I was talking to our mutual friend, Steve Sandler, who's been in it 40 years. So he makes us feel good about ourselves. and he's saying that across his whole career, he's never seen this kind of velocity. And I thought we were fast and pushing it all the time before. So I'm trying to get my head around and it, like you said, it's not just. Matthew Burns (03:49) Yeah. Judy Warner (04:13) chip level, right? It's chips attached to boards attached to cables attached to the, so this whole ecosystem has to be brought up in this unprecedented speed. So anyways, I'm getting ready to drink from the fire hose at DesignCon because I think I'm going to see stuff that I didn't even see coming, right? Because I only go to two, three trade shows a year and you're in it. Matthew Burns (04:34) Yeah. Judy Warner (04:41) all the time year round. So it's fascinating and I'm excited to talk to you more about the way Samtech's leaning into that and the kind of things you're coming up with. you know, for our audience, they're design engineers. So what are the challenges because of this philosophy and the R &D and everything that's going on, what are the design challenges they're facing? And if you have any thoughts about what's coming, I think our audience. I would appreciate a little heads up. Matthew Burns (05:13) Yeah, I would say it's been interesting over the last couple of years kind of watching the design challenges evolve with where technology is going. Right? If I say, let's say pre-G, you know, I'm just going to just look at recent history. You you look three years ago, three, four years ago, we're still in middle of the pandemic. It's sort of a pre-Chat GPT world. You know, the focus was speed. How can I get faster, quicker? And that was evolving every Judy Warner (05:35) Mm-hmm. Matthew Burns (05:42) to three years plus or minus. I don't think that trend stopped but because of AI, because of the exponential growth of AI foundation of models, you you're talking about models that have billions, hundreds of billions, know billions sorry too many zeros, you know multiple billions of parameters per system and you each of those single parameters requires a transistor to go back and forth. So you're talking billions and billions of transistors, parallel processing. So it's not one GPU, or XPU, or AI accelerator. It's multiple. You're doing the parallel processing that x86 CPUs just can't handle. So not only is it faster, but it's how do I scale these platforms to be able to handle these models quickly? And so far, the GPU solutions, the XPU solutions on the market are really what's leading it. Judy Warner (06:17) Right. Matthew Burns (06:40) In addition to speed, if you look at the TDPs of these chipsets, they're approaching 2 kilowatts. 2 kilowatts for a single chip. So, OK, I have to route 200 gigabit data in a PCB the size of a laptop. I need to get 2 kilowatts in while also getting rid of all that heat. It's not convection. It's not conduction. It's not a fan. Judy Warner (07:04) Yeah. Matthew Burns (07:10) It's liquid cooling. Yeah, so whether that's a cool plate or a, know, a CDU that's attached to the AI chassis, you know, it's really changed the game in terms of how system engineers, lecture optical software, it doesn't matter, right? Anyone that's in this ecosystem really have to change their thought process. The most important design consideration is how do get all that power in and how do I get all that heat out while at the same time Judy Warner (07:11) is liquid cooled. Matthew Burns (07:40) pushing the envelope in terms of speed performance. it's exactly. it's not just a signal integrity challenge anymore, Judy. It's signal integrity, thermal, power. Some of that's PDM, but it's also power delivery. It's material science. It's working with, is it immersion cooling? Is it liquid cooling? Judy Warner (07:43) and keeping it light and small. Matthew Burns (08:05) The ecosystem of engineers has evolved and grown exponentially just because of the design challenges you have trying to support AI. So again, high speed, but more importantly, power in and thermal out. So from what I've seen in the trade shows, design cons, supercomputing, whatever, the ones I've been at over the last 12 months, that's becoming the theme. Now it's one caveat I would throw in here is, I was at a trade show, I think it was supercomputing, and someone kind of joked that every time you put a new data center up to design an AI-centric data center, you need to put next to a nuclear reactor. And that's happening. I don't know if you saw this in the news, but I think it was September-ish last year. Microsoft is working with Constellation Energy to bring Three Mile Island back online. Judy Warner (08:56) No. Matthew Burns (08:57) Yeah, now, granted, it's not going to come online until about 27 or 28 because of the legalities and everything they have to go through to bring the power plant back online. But that's the reach that the hyperscalers are going through to support their power needs. Oracle is working with nuclear reactor startups to create microcells to support their data. There's all sorts of data points along this standpoint. Judy Warner (09:14) their power. Matthew Burns (09:26) That's where the industry's going. Judy Warner (09:26) In other words, we're not going to count on wind farms. Sorry. Matthew Burns (09:31) There's a piece of it. That's okay. That's okay. Judy Warner (09:34) I just, I'm sorry, I can't help but be sarcastic, all, you know, there's no way with these kind of power demands that we can rely solely on green energy. We're just not there yet. Maybe someday, but it's just, I think nuclear. mean, there's just so many people talking about it and sorry to offend you green energy people, but the... Matthew Burns (09:53) Yeah. Yeah. Judy Warner (10:04) These power demands are so vast. That's wild. Matthew Burns (10:08) Yeah. And I think that just goes to, while, you know, we're maybe referring to this maybe slightly sarcastically or slightly tongue in cheek. It does illustrate how disruptive everything with AI has been. Because it's not just designing the chipset to get the highest performance out of it. Power in, power out. It's how do I create the data center where I can put 10,000 of these chipsets or 10,000 of these shells within a within Judy Warner (10:22) Yeah. Matthew Burns (10:37) with a specified footprint power, whatever. So it's, it's, it's becoming more complex and it's going to continue to continue to do so. one last data point. don't know if you saw this or not. Microsoft mentioned, I think over the weekend, first weekend of the year, they're, going to spend $80 billion just in 2025, ramping up data center production, $80 billion capex for investing in data center. that and they're one of eight to 10 hyperscalers that need to invest that much. So it's hundreds of billions of dollars each year, quarter of a trillion dollars each year being invested in this space around the globe. So that really underscores the challenges that designers, system developers, et cetera, et cetera, have to consider. It's not just about putting in the schematic, making a PCB and making sure the thing works. It's this entire ecosystem. Judy Warner (11:08) Right. Right. Mm hmm. Matthew Burns (11:34) that really has to drive things. I think that's, you know, I'm getting long-winded on the topic, but that's going to be continue to be the trend going through throughout 25. Judy Warner (11:42) Yeah. Well, I think it's a fascinating topic and I think our listeners want to know, you know, and get a grasp, you know, what's going on. It's happening so fast. It's actually hard to keep up with. And I think EDA tools are going to play, have a big piece in this, you know, and it'll be interesting to see how that's going to morph as well. I know not too long ago, Matthew Burns (11:57) without question. Yeah. Judy Warner (12:10) specifically to what Samtech is doing and thank God for flyover cables, right? It's like, but I know you introduced a product called SciFly, which I think is a really fun name. Why don't you tell us a little bit about that and for your little piece of the ecosystem, you know, how that comes into play. Matthew Burns (12:15) Yeah, yes. No, that's a nice transition because you can talk about AI for hours and days and never cover everything. So thank you for keeping us centric for where we're supposed to be, Judy. Going back to speed, that's the one part of this whole ecosystem where Samtech contributes to. It's not only the cables, but it's the signal integrity expertise as well. So thankfully, as we're seeing this transition from 100 gigabit to 200 gigabit solutions in a single lane, that's where Judy Warner (12:38) Ha ha ha. Right. Matthew Burns (13:00) fiber technology is a differentiator. The biggest reason is, you look at ultra-low skew twin-axe solutions that we have right now, differential pair skew is 3.5 picoseconds. We're getting down to 1.75 picoseconds on next generation solutions. The loss of the cable continues to get smaller, even as Nyquist gets bigger or higher. We're roughly one third to one fourth the loss of PCBs, so we can extend the reach of the signals. Judy Warner (13:16) Wow. Matthew Burns (13:30) over longer distances. And if you're trying to route signals, just think of a traditional networking chassis or blade. You're trying to take it from the networking IC to the front panel, that may be up to 18 inches or a half meter of cabled solution. trying to find connectors that fit on the front panel, which are typically an MSA, maybe it's an OSFP, an OSFPXD, a QSFP, whatever. On the back end, the question that comes up at higher speeds is, is it possible to integrate the TwinX cable directly on the package of the IC, or do I put it right next to the switching ASIC, or GPU, whatever, depending upon the application? So that's a question that we run into. So what we've done with Sci-Fi is the next generation solution target 224. Judy Warner (14:16) Mm-hmm. Matthew Burns (14:27) really supports both approaches. So we're working with leading chipset designers in Silicon Valley that will actually take SciFly and actually put, you know, the cable comes in, it's connectorized, connector six directly on package right next to the IC. It's a much more complex design because it's not just an IC being placed on the PCB, but it's the die going on the package. Our connectors are right next to the die. We have to work with their... packaging vendors to come up with a solution that secures all of that to the same substrate, put the solder balls on the substrate and then drop it down on the PCB. It is. So you'll see some information from us on next generation prototypes of this solution working with our tier one vendor at DesignCon. So we're really excited about the opportunity. We're really excited to see the application. And it's pervasive. It's not just Judy Warner (15:05) That's amazing. Matthew Burns (15:24) networking, but it goes into storage and servers and AI and things like that. Additionally, we also have the near chip version of SciFly so that in instances where the design is too complex for a connector on substrate application, we have the near chip or ASIC adjacent solution that keeps the trace length from the ball to the connector, maybe a half inch or an inch, depending on how the trace goes. So in those types of applications, it's all about how do I optimize the breakout region from the signal on the chip to the breakout region on the interconnect? How do I make that interconnect not only support 200 gigabit plus performance, make it small, make it dense, make it reliable so that it can have multiple mating cycles? That's really what SciFly is designed to do. So we have SciFly for interconnect on substrate, and then we have the near chip. near chip component version that allows for both approaches. So we've demonstrated SciFly reaching 224 gigabit. We actually worked with, this is public, I can mention this, we worked with Synopsys at Supercomputing demonstrating their latest 224 gigabit IP with a one meter long version of SciFly. The bump to bump channel was 40 dB at 56 gigahertz Nyquist. and we had a prefect BER of about E to the minus nine on it. we're confident that this is going to be a game changer for a number of ASIC to front panel or ASIC to back plane applications. And we're really excited to talk about it at DesignCon, demonstrate the solution with our chipset partners and help the industry to adopt it. Judy Warner (17:17) Well, one of the cool things about you and I doing this every year is I think about what were we talking about last year? And it's like, now you're just like, oh yeah, 200 gigabits per second, whatever. Last year we're like 200, 224, you know, and we were just, it was just becoming viable, right? So a year later you're talking like, oh yeah, that's... Like, where are we with 224? And then in a bit, I'll ask you about looking ahead since you're on the front lines of all this stuff. Matthew Burns (17:53) Fair enough. 224, 200 gigabit is still ramping up. I mean, there's solutions on the market that have been publicized by the solution providers. Some of that's in the AI space. We're starting to see it come out of the network space a little bit as well. So 224 is reality. Is it widespread? Do we have widespread market adoption from all the equipment vendors and every single data center application? No. We'll probably see that over the next two to three years. And even though we have production ready solutions, ready to go. There's production ready solutions available on the market. And there has been probably the last, called the last 12 months plus or minus. I'd have look at product release schedules from providers, but 224 is going to continue to ramp up. You and I talking about it for five minutes, Judy, make it sound like it's trivial, but it's not. You know, the solutions that were working. Yeah. Yeah. Judy Warner (18:45) No, it's not last year was like, we're going to have something in the booth and it actually works. You know, it was, it was, you know, an anecdote, but now you're saying we're way down the road. It's not widespread adoption. It's still a lot of stuff, you know, doesn't run at that speed. It's novel and it's interesting though. And it is no small thing that we can do that. Matthew Burns (19:12) No, it is. you it's, it's, look at the, the, the contributions from, like I said, we talked earlier, I know this is a whole, an entire theme of your entire podcast series is that it takes a village. So you mentioned the EDA vendors, you know, interconnect obviously from companies like Samtech, Silicon vendors, you know, the test and measurement guys, you know, the key sites, the Rodeo and Schwartz's, the, the Enritsu's. Sorry, I'm not trying to, I'm not trying to, not mentioning anybody, but my brain just went south on test and measurement vendors. Yeah. Judy Warner (19:34) Mm-hmm. Yeah, no, it is the ecosystem, right? It's the representative of the ecosystem. Matthew Burns (19:42) Yeah, the laminate providers, et cetera, et They're all innovating. They're all moving forward. And everyone's working together to come up with these solutions that make sense where everybody has their differentiation. So I think you're going to continue to see 224 to ramp and become more and more of a reality in 25. So we see it with our SciFly solutions. We see it on the front panel with OSFP and OSFP XD. Judy Warner (19:49) Bye. Matthew Burns (20:11) you know, there's increasingly, there's increasing amount of chipsets available with 200 gigabit transceivers, FPGAs, GPUs, CPUs, 800 gigabit ethernet, 1.6 terabit per second ethernet is reality, know, 8 by 200 obviously is giving you the port speeds that ethernet's working on. So we're there. It's... Over the next six to 12 months though, you're going to see it ramp to full scale adoption and go from there. What's interesting too is that I always tend to think of AI slash data center as the tip of the spear in terms of market performance. behind that, and maybe this is just a side point, we're starting to see faster speeds in embedded. We tend to think of embedded as gigabit, single gigabit per second. Judy Warner (20:53) Mm-hmm. yeah, embedded is huge right now. Hmm. Matthew Burns (21:08) But when you look at the FPGAs from Altera, when you look at the FPGAs from AMD, they're 56 gigabit, they're 112 gigabit. So, you know, what was bleeding edge in the data center call it four or five years ago? It's bleeding edge for the embedded ecosystem. And if you're working with an embedded engineer that's used to working on 10 gigabit design, moving up to 50 gigabit is not trivial. You know, that's trivial for a data center person, but if I'm someone that's in factory automation and I got a Judy Warner (21:22) Mm-hmm. Mm-hmm. Matthew Burns (21:37) enable or mill arrow or commercial satellites or something along those lines and you need to support those type of data rates all this ecosystem SI concern becomes a new reality for new applications. So it's interesting. It's a fun time. Oh yeah without question we didn't even touch that. Judy Warner (21:50) Mm-hmm. And so does RF, right? Yeah. Which is a whole other mess for design engineers. It's like, sorry, you guys, we just keep throwing more and more on these guys and gals plates. I'm like, you know, anyways, I don't want to wax too heavily on that because I will. just think, wow, how do we? Matthew Burns (22:03) That's. Judy Warner (22:23) How do we support these folks, right? The best I can do is make a podcast with people like you. So for a moment, I'm going to give you a shameless plug moment. I know you're unable to be at DesignCon this year and I'll miss seeing you, but why don't you tell us specifically about the things Samtech is doing and showing at DesignCon so those that are attending will know to look for those things. I know you often do a lot of presentations. and demos in the booths all of that. So talk to us a little bit about that. Matthew Burns (22:58) Yeah, so first and foremost, we talked about the demonstration of Sci-Fi HD. So I think that's going to be one of the biggest demos that we have in the booth. But we have a full 20 by 40 space. It's booth 939, so we encourage everyone to come visit us at the show. We've got six or eight different types of presentations ranging on topics from PCI Express and PAM4. PCI 6 is obviously 64 gigabit per second PAM4. PCI 7 is coming out later this year at 128. What about 256? PCI 8 is going to come out. PCI 6 will start working on that eventually. We've got the who's who of SI luminaries, the Scott McMurrows of the world, the Rich Melitzes, the Brandon Gores, the Steve Krueswicks, the Jignesh Shahs. We have a number of technical presentations where these thought leaders, these luminaries will presenting their viewpoints, their study, their research, their work at the show. We'll also be demonstrating other solutions that we have. You mentioned RF, something else that we're really excited about is our next generation Bullseye platform. Bullseye is a solution that we've had on the market for, I don't know, 15, 20 years. It's really designed as an SMA replacement. So if you think about Judy Warner (24:14) Mm-hmm. Matthew Burns (24:23) an IC characterization board with dozens or hundreds of transceivers. If you put SMA down for all those lanes, you're talking, you know, dozens or hundreds of SMAs. It becomes bulky. It becomes expensive. It becomes, you know, it's just a challenge to be able to attach and reattach the cables for those solutions. So, know, Bullseye basically gives you four to one space savings as well as giving you phase matched RF cables allowing you to terminate the end to the cable, you know, one millimeter, one point eight, one point three five, one point eight five, two point four, et cetera, et cetera. The latest version of Bullseye is targeted at 90 gigahertz. We've actually tested up to one 10th. So, you know, when looking at a two 24 PM four great, great solution for testing that we're starting to see the, the seedlings, I guess of interest in 400 gigabit, serial lane. So again, I, purely conceptual, seems like pure science fiction, but once we get in 26 and 27, we'll be talking about the reality of 400 gigabits per lane. So we've seen some testing capabilities from Keysight at 400 gigabit. We know that some of the IP providers, without naming anybody, are starting to look at that technology. And we are too, in terms of how we can test it with Bullseye and some of the performance we have with our TwinX cables. Judy Warner (25:24) Right. Matthew Burns (25:50) So that's just a little bit of a foregling in terms of what we're going be demonstrating or talking about at DesignCon. We're really excited about the event. As always, Judy, it's always a good way to kick off. We mentioned earlier, it's always a good way to kick off the new year looking at what the ecosystem is working on for the remainder of the year. So we're looking forward to it. Judy Warner (26:14) So this is sort of a fun part of this little annual podcast we do is this is where I ask you to put on your turban and get out your crystal ball. And with all the technology and the ecosystem development and speeds and frequencies and all this, what do think we'll be talking about on this podcast in 2026? Matthew Burns (26:38) Uh, AI. I mean, you can't go wrong there. Yeah. Yeah. Uh, it seems like it, but I think there's a, there's a, there's a large level of truth to that. you know, AI and the data center obviously will continue to be, uh, a big topic. You know, something I saw just to kind of continue on that trend. I read somewhere that something along the lines of less than 5 % of the fortune of the global fortune 1000. Judy Warner (26:40) Right from now until forever. That's all we're going to talk about is AI for the rest of our lives. Matthew Burns (27:08) is actually using AI for their internal processes. So we've been talking about gen AI, we've talked about inference, we've talked about training. We haven't really started talking about enterprise AI, right? How is it that these global Fortune 1000 companies actually use the capabilities, the powers, harness AI to actually improve their processes so that they can drive value to the bottom line of the company? Judy Warner (27:19) Right. Matthew Burns (27:32) Obviously the hyperscalers have figured that out because everybody's, know, they're, they're, everyone's paying them for access to the AI. but how, you know, if you're, if you're a bank or if you're, if you're an interconnect, interconnect manufacturer like Samtech, how do you use AI to improve your processes? You know, we look at from machine vision standpoint from, you know, to, look for manufacturing defects. but there's so many more like applications. So I think you're going to, you're going to, we'll see that becoming more and more of a reality over the next six to 12 months. A lot of the analysts in this space predict that, Judy Warner (27:39) Right. Mm-hmm. Matthew Burns (28:02) There's been so much money that's been invested in AI capabilities, know, companies aside from the hyperscalers are going to start to look for real ROI. So I think that's going be a big thing. You know, we just mentioned 400 gigabits about how we're starting to see, you know, seedlings of interest in the technology. I think it's going to be more of an everyday conversation as we go into 26 a year from now. Will you see products at 400 gigabit? I don't think you'll see products before 27. That's just my opinion. Judy Warner (28:18) Right. Matthew Burns (28:31) But I wouldn't be surprised that a year from now that we start to see people talking about, we're going to have a 400 gigabit solution that does this. Ethernet, the entire Ethernet ecosystem, they're talking about 1.6 terabit per second port speeds when we get to 3.2 terabits. It's just a matter of time. Judy Warner (28:32) Okay. Matthew Burns (28:57) and gigabit technology will drive that. So I'd say those are two or three things I'd look forward to in the future. And then from Samtech's perspective, it's always going to be faster, smaller, denser. So that's our perpetual opportunity, perpetual challenge, and that's what makes us who we are. So we continue to answer the bell when our customers bring that to us, whether it's data center, AI, compute, or more traditional embedded industrial applications. So that's where I see us at 12 months from now, Judy. Judy Warner (29:02) Mm-hmm. I read something today just on LinkedIn where somebody was opining about photonics, you know, and I think they're saying that they think there's going to be explosion of photonics startups this year. So I thought that was kind of interesting. What do you think about that? Matthew Burns (29:46) It's inevitable that photonics will become more and more of a reality. Samtech is confident that copper has its place at 224, especially inside the box. We think for short to mid-range cabling, box to box, chassis to chassis, copper can play an answer, can provide a solution there. Especially if you look at active copper, at 224 active copper, maybe you get three to five meters. Judy Warner (30:08) Mm-hmm. Mm-hmm. Matthew Burns (30:15) of length. But when you start getting out to seven to ten where you're linking top of rack or you have to GPU clustering, you need optics to do it. The copper just can't make it. And frankly, it's the heat, it's the thermal response. The photonics are just, they weigh less because there's no weight there in the fiber. There are latency issues. There's only a certain amount of length you can go. Judy Warner (30:39) Right. Matthew Burns (30:45) with the fiber optics and still keep the data rates that you need. for most use cases, you only need to go 10 to 15 meters. So whether that's front panel to front panel, which is typically what we see now, or we start to see the transition to co-package optics, Broadcom has done some nice work on that. TSMC has done some nice work on that. This is all public, right? I'm not commenting on anything that hasn't been publicized. Judy Warner (31:09) Right. Matthew Burns (31:13) I agree with the pundits that optics are going to transform this ecosystem. We don't typically look at co-package optics or front panel optics. We tend to look at midboard because they have some unique applications. But we think that there will be, we know there's niche applications in the data center for midboard optics as well. So again, just because of distance. So that's going be, that'll be something that we'll talk about more and more as time goes on. And I would be. I would be shocked if that's not a major messaging theme at DesignCon here come up in a few weeks. Judy Warner (31:45) Yeah, that's what I was kind of wondering about. saw that and I'm like, hmm, you know, I see these little blips on the, on the meter. like, huh, I wonder if that's going to show up at DesignCon. Again, because DesignCon often has novel, bleeding edge stuff that you don't, you know, and even our audience isn't going to see day to day, but it's helpful to track where we're headed and, and to sort of get educated up and. And, you know, do a heads up of where we're going. So it's kind of fascinating. Well, Matt, thank you again for your time. Welcome back to the, to the new year and all the challenges ahead. I appreciate everything you and all your team share with our audience. And I do, I am going to put the link below for our audience, which is the Sam tech. Matthew Burns (32:27) Thank you. Judy Warner (32:39) DesignCon page and that'll tell you where all the SI and all that you have an amazing team of subject matter experts that are going to be presenting and I'll put those links below so you can find them easily. So Matt, thanks again and I look forward to seeing you if not at DesignCon somewhere down the road soon. Matthew Burns (33:00) Thank you very much, Judy. It's always a privilege and I love getting on your platform and talk about tech. It's always a bunch of fun. Thank you so much. Judy Warner (33:06) Thank you. For our audience, thanks so much for joining Matt and I today. I trust you enjoyed this geeky conversation about AI and where technology's heading. Please go check out the show notes and check out Samtech's DesignCon page. I will be walking the floor. I'm not having a booth this year at DesignCon, but I'll be around and about, and I hope to see you there. We will see you next week in this continuing series. And until then, remember to always stay connected to the ecosystem.