 So hello. So we're here at the Iron Tech Con. So who are you? I'm Kalpesh from Open Silicon. So what do you do? So we had Open Silicon do everything from concept to silicon. So here we are trying to demo the 2.5D technology. So what is that? So we named this project as AOTAR. The 2.5D technology here, the two dies sitting on the interposer. The die is a dual core A9 with a CPU subsystem and a DDR interface running at one gig. The interposer is 500 megahertz, 64 bit wide, full duplex and two channels. So what's special about, what does it mean to do this? What is it? What is it for? So this is for, say, you want to reduce your cost, you have done some design, you already have a die with a CPU subsystem. Why do you need to design it again? Just use that for a next chip and reduce the cost faster time to market. So what is this chip? Is this real? It is real but it's a test chip. Test chip to prove the technology. So 2.5D interposer, what does that mean? So that is a passive silicon on which the dies have been sitting. So we take two non-good dies. You take two, what's the ring? Non-good dies. Non-good. Non-good. Non-good dies. So it could be like an ARM Cortex processor, it could be some other memory interface or maybe an entire custom video control. Or something else. So we are taking those two and you have already made them. And now rather than integrating again, making a bigger SOC, we are saying we could just take this passive multi-layer metal and then integrate this on top of it. So it's not three-dimensional, you're not stacking like memory, like Micron does with three dimensions. But we are doing 2.5D, which is 2D for the actual dies. And this half, because now we are putting this interposer, which is just a narrow layer of metals that are used to interconnect these two parts. So this reduces time to market drastically for customers. So is this your technology? It's in 300 plus designs. 85 million units shipped. Is this what you're talking about? Those are everything that OpenSilicon does, not just 2.5D. In addition to that, now we have come up with this new technology for 2.5D and moving towards 3D in future. So we can stack. Stack? Yeah, in future. So OpenSilicon, can you explain what does OpenSilicon do? Is it an open source silicon? No, we are a turnkey company that does designs for anybody that needs silicon solutions. So they may give us either a concept or a spec, and from there we deliver the actual silicon to them. So they themselves do not have the ability to build an entire complex chip that they may want to for any application. Networking, signal processing, cameras, anything they want to do. And then we build the entire thing for them. How old is OpenSilicon? Ten plus years. Ten plus? Yeah. And so what are we showing over there? What is this? Yeah, we're showing different IPs, that is the 2.5D, then we have the HMC IP, the SIRTIS, and we're trying to show that we can do anything from spec to boot. So special specification right to your boot to your chip production samples. So what is this one? This one is a demo proving the communication between the two dies over the interposer. All right. So how big is OpenSilicon? Three hundred plus employees. Where? All over the globe. And Fortus is here in Melpitas. Melpitas. And what's going to happen in the future? Is there a roadmap or is there something announced? There's a bunch of things happening already. Yeah. All right. So we are trying to do bigger and more complex SOCs for customers. And then one of the new technologies for instance we are coming up with is the hybrid memory cube, HMC. HMC? HMC is the hybrid memory cube. Yeah. Which is another new IP that we are building and that will be the next gen memory technology. So today we have DDRs, so in future we will have memory that is stacked. Okay. This is the memory that's very recently been standardized. Like POP? No, there's nothing to do with that. Yeah, not quite. Okay. So memory itself is being stacked. And then that can be used to increase bandwidth to like 160 to 240 gigabytes per second or greater. So that's another IP. Are these technologies proprietary to open silicon? Yeah, these are proprietary to open silicon. These are things that we are working with partners and working by ourselves in this case to help customers reach boot stage very quickly. So what do you do here at ArmTechCon? Do you talk with potential customers? Yes, financial customers who have come and talked. We've also talked to potential IP vendors who may be interested in using their IP in our silicon. Alright, so thanks a lot. And all of this, all of what open silicon does is all ARM? Not necessarily. We do other things too. But a lot of customers ask for ARM. Mostly is ARM? When you say 85 million units shipped, what is it mostly? We have different mips, we have all kind of mips, time area, we have different cores. Different cores are good. ARM is one of the cores. Is it the best? I don't know. What do you think? No? Our customers can answer.