 So we are a company called Silicon. We are recently acquired by Applied Materials. We are experts in the low-power smartwatch wearable market. We have our GPU 2D rendering IP inside. This is the Mbeak Apollo 4 product, which is using a Cortex-M4 core with this one here, Mbeak Apollo 4, which is a Cortex-M4 core, along with the display subsystem is made up of NEMA graphics. The Pico XL, which you see here is inside there, which generates and registers the graphics in this embedded system on a smartwatch. And it also has our NEMA display controller, which generates very vibrant colors that you see here on the display. And this is running at less than 100 megahertz, and it has a memory of 1 megabyte of S-RAM and 1 and 2 megabytes of M-RAM. So very, very small memory footprint. It's a sexual device. When it says small is mighty, what is the expertise? What do you do at Silicon? I think Silicon, do you want to talk a little bit, Yanis? Yanis is our software lead. So what we do, we do actually is that we have a really small device, but it is really efficient. So what we do is bring features that normally you would find only on the high end of GPUs, for example. And you can have on smartwatches really elegant fluid animations, but the power consumption will be really low. So you will have a really long battery life. You can have fluid graphics with up to 10 or 12 days of... Are you running on Cortex-M? Cortex-M, M33, M4, yes. RISC-V, yes. So what is a secret? How do you do graphics on the Cortex-M? And it's good. The secret really is that the IP was designed from the ground up for this specific application. OK, so we knew the embedded world really well. We knew graphics really well. So we found a really good balance, internal features, performance, and the power consumption. Nice. Do you want to show us another demo? Is actually the watch you're wearing... Yes. ...things to look at? Yes, this. All these watches. So can you describe what is this watch? This is a Xiaomi Watch S1 Active. It has a resolution, I think, of 460 by 460. So what is... Nice big display. Yes, very colorful, very bright. Exactly. Somehow long battery. Yeah, it's from, I think, 7 to 12 days. So from my experience, I think like 7 or 8 days with the screen always on like this, it's possible, yeah. And the graphics... With my Samsung here, I could barely do one day. Exactly. That's exactly what we do. It's easy to watch power. That's exactly what we do. But it's checking all my vitals. Is it vitals and stuff also? Yes, yes. I always have them, yes. And you can see here... So you are in actual products out there. You can see... Yes. There. What do we see here? So you can see we span the entire range of the segment from the feature watches, the fitness trackers, like the Xiaomi Mi. All the way up to the high-end segments, such as the garment that you see here on the left. We power the displays in all of these devices without 2D graphics rendering and our display controllers. And, yeah, Iris, can you show... Shabak, the demos here that we have shown... Yeah, maybe more of the VG demos. So maybe also the PVG demo. Yes. So what is the Pico VG? Pico VG is our new CPU that will launch from October. And it accelerates vector graphics. You can see here a map, an SVG file that's rendered with vector graphics. All right. Yeah, so the beauty of this processor is it can either do to accelerate rasterized graphics or vector graphics. And vector graphics you can see on, like, maps or fonts. And it takes up a lot of the utilization on the Cortex-M. But now that we are running it on our accelerator in the NIMA PVG, you only get... You only require 20% utilization on the Cortex-M versus AT. So all this is rendered on the NIMA GPU. So you work with the chip makers? Yes, we do work with some chip makers. And you can see here that one of them is listed here, which is the Ampig company. And so they integrate your IP on the PSOC? On the PSOC, yes, we're an IP vendor. There's others in the next floor. Next door that you can see big names. Next door, which are also integrating our IP. These smartwatches have big names. Yeah. Like world leaders, I must say, that these guys, they like selling those smartwatches. And it's crucial to have something that's next door and has one battery. Exactly, exactly. That's the value proposition. The future, though, is more also adding extensions. So we obviously extended from our original GPUs. We extended, we added vector graphics capabilities. And now we also are adding in the latest NIOX architecture that you can see here. We are advertising, we have pamphlets. It's adding AI capabilities on the GPU. So we add compute shaders, which are very common in smartphone markets. Compute shaders are very common in the smartphone market. And I've been water falling down into the MCU market. And I'm running on the next generation NIOX as compute shaders. So you can do post-processing, pre-processing, classic computer vision, image processing, and also machine learning. Here we have a demo where you are doing some person detection on the NIOX GPU in very, very low power consumption. How long has I think Silicon been doing this kind of technology? So maybe you can answer that. Yeah, we started in 2007. But with graphics, we have been developing only graphics products since 2010. And yeah, so we are quite a long time in the segment. And initially I think that we are actually the first GPU that exclusively targeted the embedded market like smartwatches. And are there new hybrid Cortex-AM that could be taking advantage of your technology only when it's on M and doing something else on A? Actually, there are a couple of watches here. I think one of them is that one, the OPPO Watch 2, which have like a big little subsystem. On the big subsystem, it's running whereas on the little. It's, I think, an RTOS-based system, which of course consumes very low power, very low energy. And when it's running on low power mode, then it's using our graphics. Nice. And what does it use when it's not on low power mode? I'm not really sure. I'm not really sure with the specific. You're always going to be on Cortex-AM only? Cortex-AM on the high power. Cortex-AM on the low power. You do stuff on A, yeah. Good, all right. We can, yeah, we are scaling all the way from, we scale from 0, 1 quarts all the way to 64 quarts in the future in real applications, in real chip sets. We have in the MCU space alongside Cortex-AM, but the plan is to move into the space of Cortex-AM in the future as well. So I love the smartwatches. And what's the big idea here in the embedded world? What are people asking you for from you? Yeah, it's all about cost, power, and performance in that very limited amount of footprint that you have available to you in that small silicon. So you're working with sub-3 megabytes of SRAM, and you're working with, you know, we are consuming less than 10 milliwatts, which is so significant in these very small phone factors.