Semi Doped
The business and technology of semiconductors. Alpha for engineers and investors alike.
Semi Doped
PicoJool's Al Yuen: The Case for GaAs VCSELs in Scale-Up Interconnects
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Al Yuen, CEO of PicoJool, talks with Austin about using VCSELs for scale-up optical interconnects in AI data centers. Al explains why gallium arsenide (GaAs) supply is unconstrained while indium phosphide (InP) is limited, and how PicoJool can leverage existing supply chains to ship in the millions per month. They cover the roadmap from 1.6T to 12.8T, the trade-offs among 8×200G, 16×100G, and 32×50G, and the push to train the next generation of photonics engineers.
- The spec that changed is error rate. AI needs error-free links (10⁻¹⁰–10⁻¹² vs Ethernet's old 10⁻⁶) because GPUs act as one system — one error stalls the whole job.
- Single-mode optics can't hit the volume. Built for ~100K city-to-city links; datacenters need millions/month — a 10–50× gap
- GaAs vs InP is the whole supply story. VCSELs run on unconstrained GaAs; single-mode on constrained InP. 1M units: 8–10 weeks vs a sold-out 8–18 months.
- One platform, three flavors to 1.6T: 8×200G "fast and narrow," 16×100G LPO low-power, 32×50G NRZ "slow and wide" — pick on power, cost, and BER.
- 12.8T needs no new tech — just more of the same: 64 channels (4×16 array in a finger-sized connector) × 200G, scaling channels, bi-di wavelengths, and lane speed.
- Capacity isn't the time consuming part — qualification is. WIN runs ~1,000 wafers/week at ~240K VCSELs each, so 1M units = ~10 wafers. The wait is tier-1 (>6 mo) and tier-2 (~3 mo) quals.
Chapters:
0:00 Meet Al Yuen and PicoJool
2:29 Inventing the active optical cable
5:03 Engineering mindset, copper limits
8:43 Why VCSELs
13:45 Scale-up and bit error rate
20:09 Unconstrained vs constrained supply
21:53 Indium phosphide bottleneck
25:49 VCSEL design and foundry handoff
31:59 Product road map, 200G launch
34:16 Path to 3.2T and 12.8T
40:10 Ordering a million VCSELs
45:10 Ramp timing and training new engineers
Follow Chipstrat:
Newsletter: https://www.chipstrat.com
X: https://x.com/chipstrat
Follow Vik:
Newsletter: https://www.viksnewsletter.com/
X: https://x.com/vikramskr
Follow Semi Doped:
Get more of Austin and Vik daily, free!
Sign up: https://daily.semidoped.com/
Hello everyone. Today we have a special guest, Al Yoon, CEO of PicoJoule. PicoJoule is an optical connectivity company, and we'll get into all the interesting details. But first I wanted to introduce you guys to Al. So, Al, tell us about you and your background. I know you've been in the industry for a long time.
SPEAKER_01Yes, so after grad school at UC Santa Barbara, where a lot of the photonics folks have originated, I went into HP, HP Labs, where we focused on uh photonics research, etc. And since then, uh around 99, I left HP and started my first company called Alvesta. And we created the world's first uh 10 gigabit Ethernet. At the time, 10 gigabit, which is 10 billion bits per second. Um, we were actually trying to figure out applications that how people would use this in 99. Uh, we would make up things like people wanted to stream video someday and all the rooms in their house. Uh and so we had to actually, but today, of course, we're doing 1600 gigabit or 1.6 terabits. So since then, I've gone to various companies. Uh, I ran a division for Coherent, and then uh started some other solar companies and Cling Tech, and then finally ended up at Lumentum in my last gig. And then about two years ago, uh through Playground Global, which is our funder, uh, we started Pical Jewel, and uh so far uh so good on uh what we're doing right now is basically in the um interconnect space, and we'll talk more about that today.
SPEAKER_00Awesome! Wow, uh what a great background. I love yeah, that you guys invented uh early 10 gigabit Ethernet and then had to create ideas to sell people to convince people that like yes, this is useful, uh you know, people will want to use it. That's awesome. Now remind me, you also um helped invent the active optical cable.
SPEAKER_01Is that sure, yeah. So very interesting story. So back in Alvesta, our first uh my first startup, uh again, another small startup at the time called Melanox, which of course now is inside NVIDIA and created this whole uh kind of hyperscale and uh you know the whole InfiniBand, etc. So they approached us, so they had these very bulky copper cables, even back then, you know, so 20 25 years ago, and they said, you know what, the copper cables are very bulky, they could only reach, you know, at that time tens of meters. Now it's even shorter. But they said we would like an optical option, but we don't really want to commit to a full kind of optical solution. So could you put the optics inside the connector? And we said, why not? Right? So basically, we took that same uh transceiver that's typically on a board, and we literally embedded it right into the connector, as you see here. And then we thought, well, that's kind of you know not very uh efficient and clean, and so we embedded that whole connector inside. So the optics uh went inside the connector, and then this is the world's first demo of an active optical cable, meaning electrical to electrical. So electrical comes in, electrical goes out, but inside the electrical to optical transition. So electrons come in, photons carry the information, and then electrons uh go back to the uh to the point B. So that's how the whole active optical cable concept came through Mellanox. And then since then, for the last 25 years, the AOC uh has been the standard workhorse in many, many data centers today.
SPEAKER_00Amazing. So having invented that and then watched it sort of become uh just widely adopted and probably produced, you know, I don't know, in the millions of cables or something, um, how how has that impacted the way you think about like what's possible as an entrepreneur and in this space?
SPEAKER_01I like to think of ourselves more um as engineers, right? So there's a difference between I would say scientists or researchers or what people call RD, right? Research and development, and engineering or product development. And so I always kind of tell people I'm more of an engineer, and engineers solve problems, and they want to basically create products that are that kind of basically meet specifications, right? So if you need a Prius, you certainly don't design a Ferrari, for example, right? So that's a little bit overkill, and and so therefore the product fits the need, the spec costs, reliability, uh in our case, reach or the number amount of power, etc. So, in answering your question, I feel like what we do is really look at trying to solve the specific problem. And today, right now, copper has shrunk to about three meters, four meters reach at 200 gigabits per second per lane, and therefore they can't get the information off each rack, right? So racks and racks of these GPUs, CPUs, lots of compute power. So these racks are getting very hot because they're packing more and more GPUs per rack because they can't exit the rack. And so to exit the rack, you need an optical solution. And so there's many technologies, and of course, vixels-based active optical cable is one of them, and that's what we're basically focused on solving this problem of low-cost, highly reliable. Again, going way back to our roots, which is replacing a copper cable with an active optical cable. So the problem really hasn't changed, it's just that the speed and the aggregate bandwidth is now 1600 times more than the old uh one gigabit Ethernet. So it's exciting. It's been a long journey, but uh, we're still at it, and we see a you know long future for Vixel-based technology uh as we go forward.
SPEAKER_00Nice, nice. Okay, so you have this history as thinking like an engineer of just like what is the problem that's right ahead of us in the industry, and not saying like we need to invent new physics, but just like how can we be thoughtful about um, you know, oh, same cable form factor, for example, in the AOC case, electrons in, electrons out, but we could use optics here, and could we put the transceiver you know on the end of the cable? So, you know, thinking very pragmatically. And now, you know, fast forward 25 uh years or so, and I guess in 2024, you started PicoJoule, and it and it sounds like you're um going to tackle the problem of again of communicating lots of information. And this time it's you know, GPUs that are talking to each other like rack to rack, where uh it's such high bandwidth copper is shrinking to three meters or so. And so you guys again are thinking about how do we use essentially active optical cables, but this time with Vixels. So tell us like why Vixels and what about the problem made you want to start PicoJoule to actually start a company and like get get in the game here?
SPEAKER_01Great question. So um Vixels, for one thing, has been around since '96, right? So it's a technology that's been in product in the field, in data centers since 1996, starting with the first gigabit Ethernet. So you know, I'm more of a historian today. Uh so here's a one gigabit Ethernet that you know HP created, and at the time Honeywell, etc. Very, very early companies that did one gigabit. And so from the one gigabit all the way to today, again, 1600 gigabit or what they call 1.6 terabits, has all been pixel-based. And going back to what I was saying, the practicality of these solutions has to meet capacity, uh, demand, uh, cost, all of those, right? You need to check all those lists. If you're trying to get into a hyperscaler today, you need to basically meet all that checklist. You can't say, oh, I meet everything, but you know, it's three times the cost of your target, or everything is great, but you know, I can't get the reach. So they want all of that met. And so right now the only thing is copper, right? And that's mainly driven because of the cost. Cost for copper, of course, is very, very minimal. Even active copper, where you have some signal integrity, uh, signal processing, et cetera, built in, like an AEC, active electrical copper, uh, the cost is still quite minimal compared to other technologies that are more elegant, longer reach, etc. So getting back to your question, what's very important in a data center, if you have different fiber optic communication, I think everybody hears it and goes, oh, isn't that the over the ocean? They have these transatlantic sub-sea type of cables, and said, absolutely. And so there's many different flavors of optical communication. So when you say that, it doesn't cover all of them, right? So it's a very big umbrella. And what we're talking about today for uh data centers is a very short reach. Today they call it scale up. So it's basically a row, typically 25, 30 meters, right? Or in English terms, uh 75 feet. Okay, so so it's a very short, you know, within your house, basically from one end to the other and uh potentially even shorter. So in that very short reach, what happens in communication is the shorter the reach, typically the higher the volume, right? So if you think about it as you have a lot in the data center where you're trying to connect all these different GPUs, CPUs, or ASICs together. And so there's millions of interconnections, miles of fiber inside a data center, which may be you know a football field length today. But when you leave that data center and you go off to between cities, between buildings, between countries, you have fewer fiber, but you need longer distance. So getting back to the data center now, and long kind of answer to your question is you have to have millions per month capacity to meet the volume demand. So that's one of the boxes, right? You gotta check off all the boxes. So if you're able to demonstrate one or two racks or go to a show and demonstrate just technology, again, technology, science, RD shows the capability that's called a demonstration. But in order to ship in volume, millions per month, you need this whole ecosystem. You need people who make the connector, you need people who make the transceiver, you need people who make the sockets, right? So everything has to be millions per month. And any one of those uh bill of material parts that go into these transceivers, if it's missing, then you can't really ship in millions per month. And that's what's happening with a lot of these new technologies that are fantastic, by the way, right? From a demonstration from a capability going into the future. But Vixel's domination is that the last 25 years we've been shipping in the millions per month. So therefore, there is no invention of a technology or capacity or foundries in order to build these up. We already have that. So now we designed the latest Vixel. So today it's 200 gigabit, which we just announced, and then we put it into the whole ecosystem and they package together, and then voila, we can build millions per month without waiting for the machines to be, or actually even buildings to be uh built up and then new machines and then new processes to develop. All of that has existed for the last uh 25 years to to again service this very short-reach data center application.
SPEAKER_00Okay. So yeah, let me reflect it back to you. So the the problem that you're trying to solve is scale up with optical interconnects. Um, and but the key insight that that you have, kind of with your pragmatic hat on, is how can we use uh technology in a supply chain that already exists and can already ship millions of cables, components, parts per month? And so that's one reason why Dixels is so attractive is because they're not a new technology. The supply chain is not new. So even though as we hear about all these other interesting things, because I think anyone paying attention, you know, listens and they say, oh, you know, the broadcoms are coherents or momentums, I hear about um silicon photonics, I hear about EMLs, you know, I hear people talk about micro LEDs, but you're saying, like, hey, don't rule out vixels, even though they're not new, they're still there's there's advantages to vixels. But tell us, like, if if vixels have been around for so long, like why aren't other folks trying to take vixels to the 200 gig per lane times eight lanes, 1.6 T and beyond?
SPEAKER_01Yeah, so like you said, there's multiple technologies, silicon photonics, EML, uh, even this micro LEDs that are coming, all vying for this 200 gigabit type of electrical signal coming in. So from the ASIC GPU CPU, you have 200 G per lane of electrical signal, right? That's coming in. Now, when you go to the electrical to photonic or E to O type of transition, then it can go to different lanes through, again, a different IC that may be a gearbox, they call it. And so you don't necessarily have to run at 200G for uh straight through, which we can, right? So that's the most elegant is if you can go straight through without having to go through an intermediary gearbox, then it saves power, cost, and then latency, which is a delay going through there. So 200 gigabit would be ideal, but what's changed in hyperscale is different from Ethernet. So Ethernet information is sent in what we call packets. So whatever information, oh, where should I go in Italy? I'm going on vacation, da da da. So that information is divvied up by your search engine and sent it in packets, and then that comes back to you. And we're not very cognizant if there's an error drop or there is some delay, because those are in the you know, hundreds of nanoseconds or milliseconds, and to us it's like you know, one thousandth of a second, we're just not gonna have to delay. But for hyperscale systems, which are literally made up of thousands of GPUs acting as one brain, if you will, right? One supercomputer, one high performance kind of cluster of computers, then that latency is super critical because the GPU notices anything in the tens of nanoseconds or hundreds of nanoseconds. So, therefore, the typical Ethernet bit error rate, which is the measurement of how much errors you're getting per second, and so if that needs to drop from 10 to the minus six to below 10 to the minus 10, or what we call error-free today, because any errors then slows down the whole training inference, all the AI infrastructure that's needed. So that has changed and allowed these very, very, you know, typically higher cost single mode. I'll mention that, single mode being longer distance, longer reach, very high performance for long distance, now has come into the data center into the hyperscale systems because of this requirement for very low bid error rate, very high performance computing and connectivity. And now Vixels has to raise its bar from 10 to the minus six typical Ethernet to the last 25 years to 10 to the minus 10, 10 to the minus 12. And we've done that, right? So now we've pushed our Vixels to 200 gigabit. Even though you can use at 100 gigabit and 50 gigabit uh NRZ, you can still leverage that to very low bit air rate. So the answer is that things have changed for hyperscalers to require air-free, and that's allowed all these very high-end single mode solutions now to compete directly with Vixels because of that additional specification that's new to hyperscale uh AI systems.
SPEAKER_00Okay, okay. So because the capacity for sustaining errors is is much, much lower. We don't want all these GPUs to just be waiting. Um, that sort of changes the game from like the cloud SaaS day to now, when everything's acting as one big computer. And uh so this bit error rate has to be much, much lower. And so vixels which are short reach have to either like essentially come down to meet that, or what you're saying is these other technologies that are longer reach, higher power, those already were closer in the demanded, or like the necessary bit error rate. And so people are saying, oh, why don't we take those technologies and bring them into short reach? But but surely that has like power and cost trade-offs of taking something that could talk at a low error rate, like over a long distance and trying to bring it in. And so you guys must be taking a different tact and saying, like, no, no, no, like let's just make pixels error-free, essentially. And and presumably that's uh like a cost or a power trade-off that that you'd rather go with, or is it back to the the manufacturing supply chain capacity?
SPEAKER_01Yeah, so it's again, it's a quite complicated matrix of items you have to check off, right? So those who were shipping uh silicon photonics and EML for long reach or DFBs, etc., single mode, very high performance devices have been around for, again, similar time, 25 years, and they've been used for long reach because they have that super high performance, very few fibers, and then what we would call WDM, wavelength division multiplexing. So fibers are very expensive when you're going hundreds of miles, hundreds of kilometers. And so you want to use very few fibers, but then pass more information through more wavelengths, more colors in that same fiber. For again, short reach, when you're talking about tens of meters, you're not as uh kind of uh locked into the cost of the fiber. It comes down because you're only 10 meters versus 10 you know kilometers or extending. So therefore, the capacity for volume supply chain of the very, very high performance, low bidder rate would be a natural, right? They come in, hyperscalers want low bidder rate. Let's go with the Ferrari, right? Let's go with the super high speed, you know, high performance single mode. But they've been used to building in the maybe 100K, 100,000 kind of volumes, because you don't need as many of those between cities, between countries. And all of a sudden they come into the data center, even though their performance is excellent, cost is a little bit higher because of the single mode nature of packaging. However, just the infrastructure to build millions per month now is 10, 20, 50x what exists. So that's brick and mortar. That's basically saying I only have one pizza oven, and I've been used to you know, a small uh clientele that I can make maybe about 20 an hour. Someone goes in, I like to place an order for 5,000 pizzas and I need them in an hour. So you're going, I need I need a basically a hundred pizza ovens. That's the exact same problem that the very high performance, single mode, long-reach traditional having coming into so silicon photonics, EML, excellent, excellent technology, very, very good bid error rate, all of those, but then the infrastructure needs to be built up, and that's what you're seeing, right? You're seeing a lot of announcements with people holding shovels and saying, we're investing in the next, you know, supply chain, the buildings, et cetera. And that's great, right? Certainly bringing more manufacturing, not only in the world, but back to the US, and those are all great for the the industry as a whole. But vixels have been around for 25 years and shipping in the millions. So we're not building, we're not putting shovel to ground, et cetera. We're just changing the actual vixel performance, the chip, and then getting uh leveraging the existing infrastructure. And so what we call for vixel capacity, it's unconstrained, right? So there's constrained, meaning they're sold out. We we're sold out through next year. If you're gonna place an order, it's gonna be have to you know be an eight month, 18 months lead time. A year and a half from now, we can get it to you. For us, unconstrained just means we have a certain lead time that's basically only limited by our cycle time of building through the factory. So if for us, you know, a vixel uh run and then a packaging run may be four, eight, twelve weeks, but that's limited just to the fact that we have to build it out and ship it, but it's not limited by the constraint of the supply and ecosystem of machines or pizza ovens. We have plenty of pizza ovens, place the order, and we'll get you your order in the cycle time that we uh we commit to.
SPEAKER_00Gotcha. Okay, that's that's very interesting and a great sort of competitive advantage there for you. So on the constrained side, is that what we hear about for listeners like when they hear about like indium phosphide being a bottleneck and just like or is it particular, like where in the supply chain is it constrained? And then for for you, like what's different about vixels that does make it unconstrained?
SPEAKER_01Yeah, so for indium phosphide, a lot of the silicon photonics and EMLs, etc., is based on this material, indium phosphide. For um vixels, our technology has always been gallium arsenide. I don't know if for the listeners that may be okay, one three, five compound versus the other. What's the difference? Indium phosphide is material constraints from the very beginning. Like you can't even get a base substrate, right? Even before you process it, just the substrates for indium phosphide are limited before you get it all done and basically made into a product, whether it's EML. Silica photonics or uh vixels. The indium phosphide substrate, bare material is already limited. Gallium arsenite unconstrained. So we start with that. And now you go through the fabs, and then, of course, fabrication, etc. Depends on foundries. They're usually very large companies that do foundries, companies like PicalJoule and others. We don't own large clean room factories that make these. We design the vixel. We design the individual kind of vixel chip device, and then we use foundries to manufacture, right? So those foundries are available, but they can't get enough indium phosphide starting material to do that. Now, after that, again, once you get to the chip level, right, you dice up, you go, okay, great, I've got the laser, I'm ready to go. Now, to get to the from the chip to a pluggable device, right, to an actual optical engine, if you will, there's a ton of stuff that happens, right? Now you've got uh laser drivers, you've got boards, and then for single mode, you have to have all the machines that align that particular Slocan Photonics EML to a very, very small core single mode fiber. So those machines have to be readily available, et cetera. And so Vixels again has that millions per month type of volume infrastructure, doesn't need to be built up, and then we get back to from the very beginning, indium phosphide material constraint, and then you have to build it into lasers, and then finally you have to package it into transceivers. And all along that uh supply chain, it's not used to building millions per month, so all of that has to be uh built up, like hardware machines, alignment machines, testers, etc., to go get there. Again, Vixels has all of that infrastructure existing already.
SPEAKER_00I see, yeah, that makes a ton of sense. Um, I love your props, by the way. Uh, I liked the little vixel that you held up. So tell us more. So you yeah, that's awesome. Um, maybe tell us what we're looking at and then walk us through exactly what you design and then kind of where it gets handed off and built and packaged, and yeah, where where your responsibilities end.
SPEAKER_01Yeah, so basically, again, background is super important, right? So there we call kind of this tree of knowledge of vixels, right? So you have this line from Honeywell through Finassar, and then Finissar goes into 2.6 and then goes into coherent. So this is a coherent line. And then for us, was obviously HP went into Alvago, went into Bracom, and there's this kind of HP BraCOM line. And then finally, there's this uh, I would say Pico light E2O, and again, throwing some old names from 25, 30 years ago, uh, they go into JDSU, which is another another big name in the 2000.com time, and then JDSU spins off Lumentum and Vyavian. So we're in the Lumentum arm. So all of these three major arms have a lot of vixel knowledge, and so the strength of PicoJoule is we've tapped into, and we have people, designers from all three of these branches. So imagine that all the know-how, again, not patented know-how about recipes, and you know, I I used the example, you can hand three different chefs a recipe for making souffle, and most likely you're gonna get three different types because it's really difficult to get a perfect souffle, if you will. It's not just getting some uh, oh, you you get the eggs, you crack the eggs, you beat the eggs. So it's a lot more to that. The same thing as in the vixel. So answer your question: what do we do? So we take all that know-how, and then we design the epi layers. So you see all these little kind of lines. So these are epi layers that are designing this vertical cavity. So it's not an edge emitting. So edge emitting basically is a flat chip, and then the light comes out of the edge. Surface emitting, right? Vixels, vertical cavity, surface emitting, comes out of the surface. And so we design all of the uh kind of an internal cavity of the laser, uh, all the dopings and not getting into too much detail, but all the process. So after we design, we hand it over to a uh Epi uh foundry that grows this material. They give us back a uh Epi wafer that's unprocessed, and then we've been working with Wynn semiconductor in Taiwan. That's our foundry. So once that epi wafer is ready, we hand it over to Wynn, and then they process it. Uh they process it into their clean room process, and then they make the actual final uh vixel device. And then the beauty of the vixels, also compared to an edge emitter, is at the wafer level, you can start testing and probing each one of these, 100% tested, what they call known good dye, before you have to singulate it and dice it up into arrays, etc. So that advantage is huge because if you have to add the more work you add in before you quote yield the device, whether determine whether it's good or not, the better the cost, right? Because you you always want to yield upstream. The more value you add, and then you yield downstream, you lose all that value added. So wafer level testing for vixels is really, really advantageous versus the the other kind of edge emitting type of technologies.
SPEAKER_00Nice. That's man, I'm learning so much history and so much sort of one Vixels 101 here. That's awesome. So you you work with Win, you design the Vixel, you have an Epi uh partner who helps with that. Win is able to even do the testing uh at the wafer level because it's a vixel, no one could die. Ultimately, it gets built into the device. Um so they're they must be doing the package, like some level of packaging for you too, Win?
SPEAKER_01Right. So uh Win is only on the uh wafer uh processing. So they come to the wafer and then it's diced up into individual vixels, and these individual vixels we work with our partners to build into either active optical cables, right, or transceivers, et cetera. And they're again another um foundry, if you will, but this is a packaging company. However, with that said, companies like TSMC, right, Taiwan Semiconductor, now is going into co-packaging, meaning after they make their silicon wafer, they'll start packaging the optics directly on top of their wafer, which is co-packaged. So there's another whole field that's growing called CPO, etc., where the traditional semiconductor foundries that are kind of wafer processing now are stacking up different technologies together, and they call it 3D kind of wafer level kind of packaging. Uh so that's all emerging. For us, we only do the wafer at WIN, and then that Win uh vixel goes to our module integrators partners, and then they'll build it up into the active optical cables or transceivers.
SPEAKER_00Gotcha. That that's helpful. So then um take us back to your roadmap. I know you mentioned a 50G version, 100G, 200G. Can you tell us more about and and I know you also mentioned a recent launch. Um so so tell us more about your roadmap, what you launched, what you announced.
SPEAKER_01As I mentioned early on, 200 gigabit per lane is kind of the benchmark, right? It's a bar you have to clear. And EMLs, Silicon Photonics have all done that. And now vixels uh have reached that. People just announced our 200 gigabit. We'll start sampling next quarter. And so that's for a very simple transceiver where you have eight channels of 200 gigabit coming in. The aggregate or the combined bandwidth of that, eight channels of eight by 200 is 1600 gigabits or 1.6 terabits. So that's the that's the standard uh ramping today, right? There's 800 um gigabit transceivers as well. That's also shipping, that's 8 by 100. And then the next generation or today's generation is 8 by 200. So that's the 200 gigabit pixel that we announced. However, there's many, many different flavors of that because of the specification. So aggregate bandwidth 1.6 T, right? Check. But there's different ways to get there if you want very low power or very low bid error rate. So running 200g, I liken it to a Ferrari, right? It can give you the 200 miles per hour, but it's a very high-end, relatively expensive because you need certain signal integrity, signal processing, all of that that adds to that. So now I say I want a very low cost, low power, but I still want low bidder rate. And then what people have done is let's slow it down, let's use the 200G performance and the vixel, but then actually run it at 100 gigabit, right? So now you have a excellent, excellent vixel that gives you a lot more performance. And if you use it at half the speed, you really get much better bit error rates or the signal to noise kind of uh ratio or relatively or relative intensity noise drops as well. So that's 100 G. So but you need more lanes, right? So to get to 1600, you need 16 lanes of 100. And then recently something came out called microvixels, and that's going even slower down to 50 G, and then they call it NRZ. So instead of PAN 4, which has four levels, 0, 1, 2, 3, now we go back to the original NRZ, which is 0 and 1. So now you have the use of the entire 0 to 1 signal to noise, which again reduces your bit error rate. But you need more channels, so you need 32 channels at 50 G to get to 1.6 T. But all three we are shipping, and all three are in demand by customers, depending on whether they want what they call fast and narrow, right? 8 by 200, or they want a LPO, linear drive, no DSP, low power, and that would be a 16 by 100 G. And then finally, if they want really, really low bit error rate, very, very low power, then they go to the 32 by 50 G uh NRZ, which they call slow and wide. We we tend to call it fast and wide and then faster and narrow. But uh, you know, when you're in the high-speed interconnection, we try not to use slow in any of our marketing.
SPEAKER_00So that's good, that's good. Okay, interesting. So yeah, this this is really cool. So you're saying, okay, there's many different ways to get to 1.6 T. You could have eight times 200, um, which would use PAM for require a lot of DSP and power, um, but it's it's definitely possible. Or you could do 16 times 100 or 32 times 50, and you need like less DSP for each of those. The the the um fast and wide 32 times 50 has like much less like uh because it's NRZ, so less like DSP and stuff. Um yeah, this is all very fascinating. So will that approach still hold once you move to 3.2 T? Is it gonna be kind of that combination of different possibilities?
SPEAKER_01Great question, because people say, right, and anytime you have a technology, they always say, what's the roadmap ahead? What's the future? Is this the end of the road? Okay, 1.6 T, we get it, Vixels can do it, but is there a 3.2 T? Is there a 6.4 T? Is there a 12.8 T, right? I mean, Andy Bechtelsheim, notorious, right? He he's created an XPO that's literally gonna give you six uh 12.8 T in a big pluggable today, right? So if they're thinking way ahead, they're they're planning way ahead, because no one has ever told us in the last 30 years, oh whoa, whoa, we have way too much bandwidth, right? So we have to have that bandwidth, and exactly like you said, what's the future? So, number one, we can use something called bi-dive, bi-directional. So, meaning we can just add another wavelength, not the complexity of a WDM where you have eight or sixteen wavelengths, like single mode uh that would do, but we basically just add another wavelength to our existing one, so two wavelengths, passing them in both directions, so bi-directional, that doubles the bandwidth without changing anything else except for you, just add another laser uh at a different wavelength, and you can leverage the entire ecosystem. So it's from 1.6 to 3.2, we could add another wavelength. The other ways to do it, of course, is to increase double the speed, right? So can we do a 100 G NRZ? That's in the works, right? So we're developing a 100 G NRZ. Today is 50 G NRZ, but we're developing a 100G NRZ leveraging our 200 gigabit vixel running at a 100 G NRZ, etc. And then in the future, we can go to more channels, right? So the beauty of vixels again, surface emitting means for edge emitters, you can only have a one-dimensional array. So you can have a one by four, one by eight, one by twelve, but it's it just makes you a long bar, if you will. But for surface emitting, we can have a two-dimensional array, meaning I could do two by four, two by twelve, two by sixteen, and then essentially couple the light very elegantly with an optical fiber uh bundle to do that. And then in that case, I'm kind of unlimited, if you will, right? So I can go up to up to 64 channels today in a 4x16 connector that's the size of, let me just show you. So a 4x16 fiber is this size, right? Here's my finger. So uh and that's that has 64 channels in there. If I run them at 200, that gets me to 12.8 T. So, in essence, the technology of today, without having to go to 400 G per lane, which we're also looking at, but at 200G with more channels with more colors, like another color for Baidaye, you double, you triple by size, et cetera. So that roadmap to 12.8 T, we believe is very solid, very clear, without even having to invent any new technology to get there. And then with new technology, it just gets easier if you can do a 400G per lane, et cetera.
SPEAKER_00Sure. Fascinating. So it's just the same 200G vixel kind of over and over. It's just what if you want to put it in an array and you get more of those, or or do you are you having to sort of invent a new vixel to try to get it the 100G version to run at NRZ?
SPEAKER_01Um, so we're just starting tests. We believe that the 200G Vixel has the capability to go to 100G NRZ. Uh so it's it's not a new vixel, uh, it's just basically using a different uh coding uh with respect to the signal coming in, uh, running at NRZ instead of PAM4. Um so that's just to take advantage of the zero to one, using up the whole signal to noise ratio as one bit as opposed to four bits, right? So that's the the difference.
SPEAKER_00Gotcha. Yeah, this goes back to your kind of engineering pragmatic mindset of just taking a Lego block and figuring out different ways to place it or different ways to use it and to sort of unlock this whole roadmap in the future roadmap. That's pretty cool. So, okay, so if uh a hyperscaler, you you talked about um unconstrained gallium arsenide and working with win, um, and they're you know used to making this stuff, they've got all the pizza box, uh, pizza ovens they need. So if if uh big hyperscaler comes to PicoJoule and says, you know, we want a million of your pizzas, um what what does that look like? Like how does that actually happen?
SPEAKER_01Right. So if if they want a million vixels, then we would give them typically an eight-week, uh 10-week kind of lead time. That's kind of the basic. Obviously, the we can accelerate that and and you know, kind of push and have engineering carry certain wafers, but typically, you know, eight to ten weeks on the vixel side. So you'll get a you'll get a wafer or you get individually diced vixels. If you want transceivers, then that eight weeks tags on a certain number of weeks to package it all into the transceiver, right? So, but those are again constrained strictly by the process of packaging, not about ordering equipment or having to build capacity, et cetera. That ecosystem works and it just leverages this to the typical cycle time we call it of building out. So typical cycle time, eight weeks for the vixel device, and then potentially you know, four or six weeks for the additional module uh and uh after that. Uh and so you're looking at anywhere from 12 weeks to 16 weeks to get to the full uh module starting from a epi reactor growing epi and going all the way through. Um and so we'll continue to drive that uh you know lower as far as lead time or cycle time, but also yields. I didn't really mention too much. Yields are a way to say if I make a vixel, how many known good dye can I get out of this wafer or this uh particular process? And the higher the yield, up to obviously 100%, the fewer wafers I have to run through, the higher the capacity of this factory, if you will, right? So if every wafer goes through and I get 100%, then I need fewer wafers, and therefore I need less capacity to for the demand. But obviously getting to 100% is very hard. But vixels have been perfecting that process for many years, for many decades, and now we're leveraging all of that. It's not new, it's not something that has to be established, it's not based on new technology. So that's very important is that since we've uh Wynn has been doing it for about 10 years since we transferred that uh for consumer electronics application back in 2016, they have a capacity of up to a thousand of these wafers per week, right? So that capability, and then the other thing is there's about 240,000 vixels on each wafer, right? Because they're really tiny. So they're very, very tiny. And so you that adds up to you know a million yielded, maybe 10 wafers. So it's it's you know, the capacity is huge for datacom. Um so I think we have no worries for us once we get to the 200G and the product specs are met with the customer, reliability, qualifications done, then we just ramp readily with win and they're ready to go.
SPEAKER_00Nice, amazing. It's it's quite compelling. Um, I think normally when people hear like, oh, there's a startup that's trying to compete in a space that has these huge incumbents, um, it's you know, the question is like, well, how is this startup going to compete? How are they gonna get to market? How are they gonna find customers? How are they gonna build up supply chain? All these things. But but what I hear you saying is that like you're you're taking sort of industry veterans that have some process know-how across like the probably um similar ways of thinking as competitors, just you know, you've been in the game for a long time, and then tapping into an existing supply chain. And also at the end of the day, it's not like you have to go win, you know, 50 customers, but there's a probably a handful of big customers that would be really make a difference for PicoJoule at the end of the day if they said, yes, you know, we'll we'll take some of yours. And then, but but the most important point, I think, of all of this is the unconstrained gallium arsenide, like being able to make a million, you know, 10 wafers with 240,000 on it, whatever you yield, you know, we're talking millions of pixels very quickly. Because I think as we see in across all of semiconductors, whether it's um memory or CPUs or AI accelerators or whatever, there's obviously there's just so much demand and such constrained supply that truly, obviously, you want to compete on cost and on engineering performance. But I do feel like there is just a little bit of like, if it's good enough and it's in production and you can install it into my data center, like you know, game on. Um, so it feels like you know, like you guys have a strategy that will allow you to, you know, deliver shipped pixels as soon as you possibly can.
SPEAKER_01Uh, which is Yeah, so the model has been around for decades and the silicon, right? I'm in Palo Alto in Silicon Valley. Most companies, most chip companies designing uh integrated circuits or uh CPUs, GPU, they don't have their own foundry. A lot of people use Intel or even AMD uses TSMC, et cetera. So AMD is this huge chip company, they don't have their own foundries today, right? And so TSMC, Intel, Global Foundry, I mean, many foundries are the quote factory floor, the the clean rooms of all these startups. So just because of our small startup size doesn't mean we can't ship in the millions per month and compete directly with the very large presence of uh other optical suppliers and companies, our competitors, et cetera. So that's the beauty, right? So we can stay very lean. You know, our motto is stay small. And that has to do with many things. Our name is Pical Jewel, right? That's low, very, very low power, and that's one of the things that's driving. And then we're also very well experienced small team, but we can have huge uh benefit by leveraging wind semiconductor by working with our supply chain, and they've Got the factories, they've got the clean rooms, and then we can ramp very quickly by providing our designs and our unique kind of specialty, and then partnering with these uh large companies that are already shipping, and they would just drop ship, right? So we don't need a big large uh company in order to ship in the millions per month.
SPEAKER_00That's amazing. Yeah, definitely punching above your weight. That is awesome. So um when is your high volume ramp targeted for?
SPEAKER_01I I if I recall, the maybe the press release said something, but uh yes, uh the press release said we're starting to sample uh next quarter, uh, and that's in different flavors, right? So we've got customers for the 50G NRZ, 100G uh LPO, and also the 200G. So all of those will begin to sample, and the process of from our device to actual product shipment or revenue for our case is a period what they call qualification, right? Qualification or reliability testing. So everything has to kind of not only meet spec at zero hour, but it has to be predicted to last 10 years or a number of years, et cetera, through what they call accelerated aging. They test it out at higher temperature, higher uh bias conditions, and then they uh kind of estimate back at the normal operating condition, can it last 10 years in the field, et cetera? And that period can for uh very small companies like tier two kind of customers be as short as three months, right, to go through that. But for tier one, because they obviously have uh much more to lose if they have an issue with their connection, uh, could take more than six months to get those uh up and running. So ramping at win is available today. However, we have to go through this qualification cycle with our customers before they get the orders, everything is uh approved, uh, meets their specifications and qualification, and then we start ramping. So we we uh also said that we'll most likely start ramping in early 2027 next year.
SPEAKER_00Okay, okay, got it. Thank you for the education here. So, yes, sampling and then the qualification process and then ramping, and it's probably yeah, as you've been saying throughout the whole thing, you're not concerned about the ramping. Obviously, it's you you you had to build the product and let customers kick the tires, and once they say, let's go, then it's off to the races. Well, uh, you know, we've covered so much. This has been so amazing. I've learned a lot. I know the listeners will have learned a lot. I guess is there anything else, uh, any last things about PicoJoule or anything that we didn't talk about that you were hoping to cover?
SPEAKER_01Um, I guess one thing that's very interesting to me, right? As you can see, uh there's the the image that you're seeing is a very experienced uh elderly startup person. And uh one thing to point out is that very few people went into hardware and photonics in our space because young people over the last literally, when the dot-com came out, it was 25 years ago. So those who are coming out of the workforce were more in the application side, what we call the software side, right? So people wanted to go into computer science, et cetera. And so what we found is the aging kind of experienced hardware folks in photonics needs to transfer all of this knowledge. So that's one of my passions is to kind of bring on the next generation, the generation after that for the photonics, because we don't, you know, just like vixels and other technology, we see many, many decades. And as I said earlier, no one's saying, oh, way too much bandwidth, right? So we see bandwidth increasing and bandwidth demand increasing with robotics and autonomous vehicles and what have you, right? Everything is going to be uh bit and connectivity kind of constrained. And so we want to basically spend time to educate, train, and so we're trying to hire you know hardware engineers and then train young folks, maybe without the experience, to kind of be the vixel designers, be the transceiver designers of the future. And so that's really exciting for us, right? Because we've got all this knowledge, you know, 10, 20, 30, 40 years, and it's really kind of feels wonderful to have the hardware excitement again, not only in the markets, etc., but the investment community. Silicon Valley is booming with photonics and hardware. So, you know, we don't take it for granted. It's a great opportunity, and we definitely want to basically take young kind of entrepreneurs, young engineers, folks that are interested in this space, you know, along for the ride, and then they take it from there.
SPEAKER_00I love it, I love it. Very inspiring, it's very cool. Um, and you know, uh that it's never been a better time, probably for interconnects and photonics and optics type folks. And uh I love that you know, you industry veterans want to bring up the next generation and give them the opportunity to learn from folks like you uh and to kind of uh revitalize, rebuild, um, make sure we have like a reinvigorated workforce so that you know we uh for my generation and the generation uh of my children that they they can keep having more and more data moved around faster and faster.
SPEAKER_01That's right. That's right. Yeah, so really enjoying our conversation, also. Thank you.
SPEAKER_00Awesome, cool, thank you, Al, appreciate it.
