Vik Sekar talks with Sean Park, CEO of Point2 Technology, about a third path for data center interconnects. Instead of copper or optics, Point2 sends radio frequency signals down a flexible plastic waveguide. Sean explains how their “E-tube” technology enables 5-10 meter scale-up links, offering lower power and higher density than existing solutions.
This episode is presented by Crusoe.
Crusoe’s Serverless Fine-Tuning trains an open model like Llama, DeepSeek, or GLM on your own data, and you pay per million tokens, not per GPU hour.
About $2 buys roughly 5M training tokens on smaller models.
Get $5 in credit to start: https://semidoped.com/crusoe
Things we cover:
E-tube: RF over plastic waveguide
Active RF Cables (ARCs) for scale-up
Power efficiency and latency advantages
The 1.6T OSFP module design
Near-packaged and co-packaged roadmap
Beyond data centers: automotive and display applications
Mentioned in this episode:
Plastic Cables for AI Clusters: Can Point2’s E-Tube Disrupt Copper and Optics? (Vik’s Newsletter, paid)
This podcast is lightly edited for clarity.
A Third Path for Interconnect
Austin: All right, hey Vik. Today’s conversation is super interesting. It’s scale-up interconnects, but it’s not copper, it’s not optics, it’s radio frequency over a plastic waveguide. This is fascinating. So it’s your conversation with Sean Park, CEO of Point2. How did you come across Sean and why did you want to have him on Semi Doped?
Vik: The discussion is always between copper and optics, but so much of the conversation around this has never centered around a third option. So I first came across Point2 technologies in my Substack where I’ve written about them in quite some depth. And then I ran into Sean at OCP APAC and I stopped by their booth. And so we’re like, okay, let’s get into this and talk on the podcast. It’s so much more fun when we can delve into all the details.
Austin: So, let’s listen to the podcast in the interview with Sean, and then I want to jump on with you afterward and discuss maybe if this can actually ship and if there’s any sort of practical realities to deploying this. So, without any further ado, let’s get into the conversation.
Vik: Welcome to the Semi Doped podcast where we have conversations with the best minds in the semiconductor industry. I’m Vik Sekar and I’ve worked in the semiconductor industry for about 20 years, mostly on the engineering side. I am the founder of Semi Exponent where I do independent semiconductor research, focused mainly on AI and hardware infrastructure. And I also write Vik’s newsletter on Substack, where I explain how technology works and why it matters for the industry.
With me is Sean Park, who’s the president and CEO of Point2 Technology. He’s a seasoned executive with decades of experience in semiconductors, wireless, and networking. Throughout his career, he has held leadership positions in many companies you’ll recognize immediately, including IDT, which later became Renesas, Terasquare, Marvell semiconductor as well. As a CEO and CTO, and the founder at Terasquare, Sean was responsible for leading the company’s strategic direction and overseeing its day-to-day operations. He was also a director at Marvell where he provided invaluable guidance and expertise to help the company achieve its goals. He holds a PhD in electrical engineering from the University of Washington and also attended Seoul National University. Welcome Sean, welcome to Semi Doped.
Sean: Thank you, Vik, for the introduction.
Vik: Yeah, we’ve spoken in the past at OCP APAC where I stopped by your demo and we played around with all the tech you had there. And it’s exciting to have you on the podcast so we can discuss all the cool stuff we talked about for the general audience because not everybody can be at OCP APAC. So that’s why we have you here.
Sean: Thank you. Thank you so much.
Vik: So Sean, Point2 has this very unique idea. And I like it because my own background is RF actually. And so I can kind of relate to it closely. Everybody’s always trying to connect data centers together with either copper or optics. It’s always like team red versus team green or there’s always two sides fighting a battle. But you have this third approach. And you are like, let’s send RF over a plastic cable. Tell us basically how did you come across this idea and decided you want to go down this path of everything else.
Sean: So I have been involved in this data center market for the last 20 years and previous to that, 20 years ago, I also worked on the wireless, wireless systems, wireless chips. And in fact, the wireless signals, we call radio frequency signal, and that is really good for wireless systems, like a signal over air and then connect two different or many different systems all together. And what if we use that RF signal over some sort of wireline communication? And then we can in fact, we can extend the reach of cables in terms of distance and also we can make a very simple silicon solution and distance will be longer and power will be much more efficient than the existing optical system. So that’s the concept that we started.
The E-tube Platform
Vik: That’s nice. So very nice points you hit on here. One of them is, I think when you say radio frequency, everybody imagines it’s some kind of a wireless link. So I think one thing we want to clarify up front is this is not a wireless link between two connections in a data center or anything. This is actually a wired connection, except that the wire itself is a plastic tube or a plastic cable. And we’ll get into why that is important and what the benefits are. And we’ll also get into the points that you mentioned about how it saves energy in the data center and why that’s really important. But this isn’t a problem you’ve been working on like last year or two years, right? You’ve worked on this problem for a long time.
Sean: That’s right.
Vik: Point2 has been around for a while now. So tell us about that journey.
Sean: So we spent many years on researching on the waveguide concept. And like I mentioned, RF over plastic concept is something that we really created first time, I believe. And it’s not really easy because the wireless signal, RF signal, travels along some sort of plastic waveguide. In fact, it will see some dispersion and also some distortion while it travels through the cable. So that’s a fundamental problem we need to solve first. So we spent many years on how we can make the cable waveguide itself a linear medium so that we can transmit data over plastic, but when we receive it, it still maintains the same shape, same quality of the signal at the receiver part.
Vik: Yeah, so it’s not as easy as people imagined that you can just send radio frequency over a plastic waveguide. We call it a cable, but it’s actually what it is is a waveguide. And waveguides have been around since World War II and they have been typically the guiding systems of radio communications for a very long time.
Sean: Exactly. Like the radar system. In a radar system, they put the waveguide pipe on the dish and then they transmit very high power through the dish and put it in the air so that they can detect some object in the air. So that is the first concept of a waveguide is a metal pipe and delivering high power wireless signal to the dish. And then we use the same concept, but it’s a flexible cable and it’s not really a metal pipe. So a metal pipe is a defined boundary condition material. And instead, the cable itself is very dynamic. It moves around, it can be bended. So it’s not easy to put some signal in the flexible dynamic cable and then receive that on the other side and still have a very clean signal at the end.
Vik: Right now you’re still a private company, but recently I was looking at your financing and all that. It seems like there’s a lot of interest. Do you want to fill us in on who your investors are? Who’s really excited about your technology at the moment?
Sean: So, anyone who sees the problem in scale-up network interconnect, they like our solution. The reason is that our solution is silicon plus a simple connector and a waveguide cable. It’s very simple, cost-effective, low power, and easy to build. The complexity is a lot less than optics. So they like our solution. For example, GPU companies, Nvidia, Arm, they invest in our company. And also a foundry company in Taiwan, UMC, they invest in our company. And cable companies like Molex, they also invest in our company. So that proves that our solution is very attractive in this market and in our ecosystem partners.
Vik: That’s exciting. So the moment you say scale-up, I’m sure half the listeners are going to prick up their ears here because it’s actually a problem in our current data center world. And I can imagine that this problem has emerged into your universe in the last couple of years because our GPUs are exploding in the data center and we need to hook up all these GPUs in a rack. Now, clearly when you mention scale-up, I noted that you don’t mention scale-out. So by saying scale-up, immediately you’re saying we have a limited reach in which we want this technology to work, right? So let’s set that expectation clear up front. What is the kind of reach?
Sean: Right. So our first generation E-tube will target 5 to 10 meters. That is a good distance for a scale-up network. So within the rack, connect many GPUs together and rack-to-rack applications.
Vik: Okay, so that’s perfect. Because even when you say rack-to-rack, it’s not really scale-out still nowadays because the scale-up itself has so many GPUs, you can’t fit them in a rack. You have to put them in multiple racks. And now you have to connect them up as if they were one rack for which you need cables. And so 5 to 10 meters covers even a multiple rack scenario.
Sean: That’s right. Yeah, 5 meters we can connect four and 10 meters we can connect eight racks. So that is a good distance that copper cannot reach.
This episode of Semi Doped is presented by Crusoe. They build the power, the data centers and the cloud behind AI. Did you know Crusoe can help you tune an open source model on your own data for just a few bucks? Frontier models are very good, but they don’t know your company’s terminology, your part numbers, or how you like your reports formatted. So, you keep stuffing instructions and context into the prompt of an expensive frontier model to get the job done. But, for simple repeatable work like reports, an open source model trained on your data can actually match a much bigger model. Think of a new grad with three months of on-the-job training. You don’t always need the PhD. And, the open model is cheaper and faster to run. Fine tuning used to mean renting GPUs, working on orchestration and Kubernetes, writing training code, and babysitting the entire setup process. But, Crusoe’s serverless fine tuning removes all of that. Upload your data sets, pick an open model like GLM, Lama, or Deepseek, and they handle the training. You pay per million token used, not per GPU hour. And on the smaller models like Quen 38B, $2 can take you a long way with millions of training tokens available. Once the training is done, you can then deploy them on Crusoe Cloud in just one click, or download the weights and run them anywhere. Your data only fine tunes your model. And, it doesn’t have to be work data either. You can try it on how you write emails. Pick a task, compare the tuned model to your current prompt, and see how it does. Crusoe’s giving all new accounts $5 in credit to get started today. That’s enough to start tinkering with fine tuning right now. Links in the description.
The 1.6T OSFP Module
Vik: All right, awesome. Now we have to go to really understanding how this works. So, guide us through the whole process of the technology and how it works. We have a simple problem. We have an electrical signal at one GPU and another electrical signal that has to be received by the GPU eight racks away. Let’s say, 10 meters, eight racks away. What is the process of signal conversion between the electrical here to the electrical signal eight racks away? What happens here?
Sean: So, we have four different technologies inside our E-tube platform. First, we need an RF transmitter. For example, on one side of the rack, let’s say we have an RF transmitter. And that converts the baseband electrical signal into an RF domain signal, for example, like 100 GHz to 200 GHz of range of RF signals. And then since we’re using an RF signal, we need to launch it into the cable. And for that part, we need an antenna and we need an RF connector to couple the antenna and the cable. And then once it transmits through the cable part, that is the waveguide, and the other side is exactly the same component: connector, antenna, but it is a receiver. So, all those components, especially for the silicon part, is a fully integrated silicon with an antenna inside a package, and the connector is a simple waveguide RF connector. And the cable is similar to copper. It looks like copper, but inside we don’t have a copper core. So that’s the main difference. And that’s why we can still manufacture that cable using existing copper equipment.
Vik: Awesome. Can you show us the cable now? If you’re watching on YouTube, you’re going to benefit because Sean’s going to show us the cable.
Sean: So I’m going to show you first the plastic core that is a pure plastic material, very flexible. This is polyethylene. And then since we are going to bundle several cores of this into one cable bundle, we need to shield it. So we cover this white dielectric waveguide core and then we put metal outside. This is for shielding purposes. So this cable looks exactly the same as a copper cable. But since we’re using a wireless signal and RF signal, we don’t have a differential pair. We only have a single core. And then we also put multi-channel inside this cable core. So we can reduce the number of cables compared to copper.
Vik: Okay. So when you say that, you will have a differential pair, do you mean you will have two of these blue cables you showed us for every connection?
Sean: Oh no, what I said is that we don’t need a differential pair. Copper, they needed twinax differential pair, but we don’t. So effectively, we can reduce the number of cables by 4X compared to copper. Because we put two channels into one cable core and then we don’t have a differential pair. So roughly it’s 4X better in terms of cable density.
Vik: Okay, that’s very interesting. So when you say two channels, what exactly is that? Are they different frequencies that go inside?
Sean: Exactly. If you think about FM radio or AM radio, you can use many different channels to deliver data over the air. So we use the same concept. We use two channels, different frequency bands, and then we can put two data rates into one cable core.
Vik: Okay, so does that double the data rate across a single cable because of that?
Sean: That’s right. So for example, this is a 1.6T cable. The 1.6T IEEE standard is 200 gigabits per second per channel, but in our cable, we are putting 400 gigabits per second per cable core.
Vik: Okay, so that cable already has two times 200. Okay, so each frequency that’s within the cable, let’s say 100 GHz, like you mentioned, is carrying 200 gigabits per second. The 200 GHz is carrying 200 gigabits per second. So totally you have 400 gig.
Sean: That’s right. In terms of data, this data rate is 400 gigabits per second, but in terms of frequency band, we are using 90 GHz to 225 GHz. So it’s less than, of course, it’s less than 400 GHz of bandwidth because we are using PAM4 and then we are using up-conversion in the RF domain. So in terms of bandwidth requirement, it’s a lot less than the data rate that you talk about.
Vik: Okay, so this is very interesting. So now, clearly there is some part of the circuit, some transceiver circuit that is converting electrical signals into these 90 and 225 GHz signals, right? Do you have a sample of that and can you walk us through what that looks like and what it does?
Sean: Yeah. So I already showed the cable parts. So this is our 1.6T OSFP module sample. So if I open up the shell...
Vik: Oh, look at that, yeah.
Sean: Then I’m showing you two pieces. One is silicon.
Vik: Okay.
Sean: And the other one is a connector on top of that silicon. So I will open up this connector. And then you will see the transmitter on one side, the other side you will see the receiver. In the middle is the antenna.
Vik: Okay, so in the middle is the antenna. Okay, I’m going to describe what I’m looking at. So if people are listening to this, they’ll know what I am seeing. So basically, this is like an OSFP connector that has a standard form factor.
Cable Density and System Design
Vik: Right, right. So that’s why I think people were like, okay, I can’t put any more copper in a rack, so let’s go to multiple racks. So what you’re saying is if we can put 144 cables in a single rack, and we can connect them all with plastic E-tube...
Sean: You don’t have to go to multiple racks if you can avoid it, which is the best solution.
Vik: Yeah, if it works, then of course it’s the best solution. It will scale like silicon. Since it’s completely shielded, I’m guessing that just like copper, there’s no crosstalk issues when you put all these cables in a rack, right? They’re all shielded anyway.
Sean: Right. In our case, we really care about the shielding because we need to minimize the crosstalk and interference outside from inside, inside to outside. So we use the same concept as copper. So the outer layer is a copper shield. And that prevents all the interference and we measured the performance is better than 50 dBc. It’s a really, really good number.
Vik: Nice, nice. Yeah, this is nice because not only do you have something that is lighter than copper, but you have something that is probably more flexible than an optical cable. Is that true? Do you think the bend radius is nice on this thing? Plastic cables?
Sean: Yeah, let me show you. So we can, of course, mechanically, we can bend it any shape. But in terms of electrical performance, our bending radius is 4.7 mm. So 4.7 mm is I can still round it across my finger.
Vik: That is roughly 5 mm. So this is a really good bending radius. You just wrap it around your little finger, right? And you’re like, okay, that’s about 5 mm.
Sean: My little finger is still bigger, so I think I need to use a smaller finger, but that’s the smallest I have.
Vik: Yeah, yeah. No, that’s good. Bend radius is important too, because you’ve got to be able to route these cables in a rack and you can’t keep them all straight all the time. So that’s also a consideration.
Active RF Cables vs. Optics
Vik: So, this cable industry is splitting into active electrical cable. Then on the other side you have active optical cable. In the optical cable world, nowadays everybody’s talking about VCSELs or talking about microLEDs. What is your version of the cable called?
Sean: We call it Active RF Cable. In short, it is ARC.
Vik: ARC. Okay. The ARC cables then will go in the scale-up domain, which is very important because I don’t think anybody should compare this technology to scale-out. This is not a replacement for scale-out optics. 50-meter reach, 100-meter reach, that’s something you cannot compete with. Optics does a great job in that.
Sean: Right. For scale-out, it was always optics. In the past, it was optics and in the future it will be optics. That’s for sure. So that market is growing, that is a good market, but that’s an optical solution market, not for us. We specifically and intentionally target scale-up. It’s a very high volume, very important market.
Vik: Yeah, that’s good. So that’s very clear to distinguish because many people I feel like are confused by technologies and there is always a right technology for the application. You have to match it to the right technology because it’s not always about reach. It’s about cost. It is about scalability. It’s about how many errors you can correct for. It’s about temperature stability, reliability. There are so many variables that matter.
Sean: That’s right. That’s right. So reliability is really important for scale-up. And especially for NVLink type of interconnect. So they try to put the interconnect inside the box, inside the system box. And then they want to put the interconnect right next to their XPUs, GPUs, and MPUs. And if they put an active interconnect solution inside the box, reliability really, really matters. If something happens, you need to replace it. But optics, sorry to always compare with optics, but optics is a relatively non-reliable solution compared to silicon. So it’s kind of dangerous for them to use an optical solution inside their system.
Power, Latency, and Co-Packaging
Vik: Yeah, that’s true. In the optics domain as well today you see a lot of talk about putting optics closer to the GPU and XPU. And you have solutions like CPO that come up. And it’s always the question of reliability there because like you say, when you have interconnects that go right up to the GPU inside the rack case, it’s not external to the rack, it’s inside. It’s very difficult to change anything if something goes wrong. Think about it, each rack has 5,000 cables. How do you know what went wrong? How do you know what to change? Now multiply that by all the racks in a data center. It’s a massive problem. But I suppose since you only have a silicon chip there as a transceiver that converts electrical to RF signals, you could very well use the co-packaged approach with RF as well, right?
Sean: Exactly. So that was what I was trying to say. Our first application will be a pluggable use case, like ARC, Active RF Cable. And the second use case is inside the system, inside the box. So that is like a backplane solution to connect many GPUs all together, for scale-up. And in those cases, we only need silicon, which is a lot more reliable than any other optical solutions. So we plan to do that. We call it near-packaged E-tube or co-packaged E-tube.
Vik: So that’s NPE or CPE, if you want to.
Sean: That’s right. NPE and CPE. Exactly.
Vik: Right. And one other nice thing maybe I should ask you to explain about, at least what I think is nice, is the fact that you don’t need any kind of gearboxing chip. For those listening, if you don’t know what a gearboxing chip is, a lot of the interconnect solutions today have a wide but slow approach, which means that you have to convert a 200-gig lane into four 50-gig lanes. You have to make that conversion because you can’t launch into an interconnect that only runs at 50 gig per second when your incoming signal is 200 gig per second. So there’s a chip that sits in the middle called a gearboxing chip that converts 200 gigs to maybe 50 gigs. But for your case, the nice thing that you have in the scale-up domain that things like micro-VCSELs and microLEDs don’t have is that you don’t need a gearboxing chip. Is that right?
Sean: We don’t need a gearbox chip at all. So our electrical interface is 100% compatible to existing or future IEEE data rates. So we take a 200G data rate as it is. And we just convert that to the RF domain and we deliver that signal over a plastic cable.
Vik: Okay, that’s very helpful. Because these gearbox chips are actually usually leading-edge silicon chips. They’re not 55-nanometer chips, right?
Sean: It’s very expensive and very power-hungry.
Vik: Okay, perfect. Power. Let’s talk about power. It’s a very important question, especially in the world of data centers and scale-up because the one thing that has stopped optics from becoming prominent in the scale-up domain—and Nvidia has said this at many events that even I have been to—is that it’s too energy expensive. It’s not about cost. They’re like, it sucks too much power. I cannot spend per rack anywhere between 20 and 40 kilowatts just for networking. To put this in context, in the cloud era of data centers, the entire rack with all the compute and everything in it used to only be 15 or 20 kilowatts a rack. Now our data center racks have become like 200 kilowatts. And now we are talking about spending what used to be a whole rack worth of compute only for networking. And Nvidia has always said, no, we will stay with copper as long as we can because we can’t spend this much energy on just networking. So what does the E-tube offer in that aspect?
Sean: So, our power efficiency is a lot better than optics. If you think about why optics consume a lot of power, it’s because they need electrical-optical conversion. And for that purpose, they need a laser, a photodetector, they need some interface silicon like a TIA and driver, and sometimes they need a DSP. In our case, we only need fully integrated silicon chips. So power efficiency is a lot better than optics. We always claim that it’s 3X better. And for example, this 1.6T first generation, we only consume less than 10 watts per module. So that is in terms of efficiency is like five picojoules per bit. And for NPE and CPE use cases, our power consumption will be less than two picojoules per bit.
Vik: That’s comparable to the CPO co-packaged optics level of promised energy efficiency per bit, right? That’s the same range.
Sean: Same range, but I think our solution in terms of power consumption, it will scale as the silicon scales.
Vik: Which means you can go to a smaller node and you can save the transceiver power?
Sean: Exactly. That’s a fundamental benefit of all the silicon technology we’ve had. In the past, for example, I have been working on Wi-Fi, Bluetooth, those kinds of wireless chip designs at Marvell. Every year, we reduced the power consumption by 30-40% by migrating from one process node to the other process node. It will definitely happen to us as well. So, tomorrow we’ll hit two picojoules per bit. And the day after tomorrow, we’ll hit one picojoule per bit.
Vik: The fact that you don’t need digital circuitry or a DSP or forward error correction at that module level means you have better latency.
Sean: That’s right. In terms of latency, DSP versus our solution, the difference is 1,000 times. So our latency is in the picosecond, nanosecond, below nanosecond range.
Vik: Okay. Yeah, that’s very important because in the scale-up domain, especially during training runs, when you have these all-reduce operations, the training run becomes as slow as the slowest link in the network. So you need all of them to have the lowest latency possible. And a DSP usually doesn’t allow this because to do the DSP, you need all the bits to arrive, and once the bits arrive, you do some error calculation and recovery and all of that stuff. All of that takes too long and too much power. DSPs suck power too. You avoid all of that.
Sean: We don’t have that. We don’t have that and then latency is a lot better.
The Road to Adoption and Qualification
Vik: So, within the pluggable module itself that you showed us today, what is your biggest challenge within that module? Clearly some part of it has to be the hardest for you. What do you think is your biggest challenge within it?
Sean: In fact, we have spent many years on E-tube technology. Nothing was really easy. First, think about the frequency range that we are using. We are using a 100 GHz, 200 GHz range of signals, right? So nobody has done 200G circuit design before. So that was the first challenge that we faced before, but not anymore because we already have silicon. And on top of that, we still need to build our own ecosystem. It’s relatively easier because we can still make use of existing copper manufacturing equipment and the ecosystem. But still, we need to convince cable manufacturers and we need to convince other ecosystem partners so that they can move on. So that was a difficult part in the past.
Vik: That’s also because you have these near-packaged and co-packaged E-tube ideas. Would you consider also being part of the Open CPX MSA? This is a consortium of companies who are coming up with a solution to put—it’s an open ecosystem of providers for those listening—so that multiple people can supply the equivalent of co-packaged whatever, co-packaged copper, co-packaged E-tube, co-packaged optics. You fit into that ecosystem, right?
Sean: In fact, we are part of the consortium.
Vik: Okay, that’s great. Yeah, that’s good because now we have so many different competing solutions in the co-package space that I think yours would be another one of them that is interesting and provides more options to people looking to do scale-up.
Sean: Exactly. We’ll give freedom for system engineers to decide.
Vik: So in terms of actually getting this stuff into the hands of a customer, let’s say a hyperscaler or somebody building. What would need to happen to this technology so that it is currently now going to be used in data centers? What do they want from you?
Sean: We have spent many years to convince our partners and also our customers, including hyperscalers, GPU companies, including Nvidia. And right now, we are at the point to deliver our product in their hands so that they can test in their labs, their environment, and they really need to use it in their system. So that’s the next step. After that, of course, we need to go through the qualification process. That will take many months.
Vik: And then the qualification process itself, do they have to do reliability testing and what are the steps in the qualification that usually are required for interconnects in scale-up like this?
Sean: Right. All those qualification processes are very similar for E-tube and also similar to copper and optics. But usually, they call it a gorilla test. So they’ll plug in and plug out and twist and then they will run scripts to all different use cases for each equipment. It’s not an easy process, but no matter what, we need to go through it.
Vik: Yeah, they’ll put your E-tube cable to the test, right? They’ll put it in temperature, they will give it mechanical abuse.
Sean: Exactly. Mechanical abuse and then some script to try to find out the break points.
Future Scaling and Unexpected Applications
Vik: In the broader scheme of things, we are in an era where we really don’t have a one-solution-fits-all in the scale-up domain. I think we are at a very exciting time in history in data centers and interconnect technology because all these years copper has been the way.
Sean: I believe three technologies—our solution, optics, and copper—will coexist for a long period of time. Because copper is still useful for very short distances. Some people want to use copper, so they will design their system to accommodate the short distance of copper. And that way they can keep using copper. And some people have already heavily invested in the optical solution, and they must use it. So they will use it. And some people who have a more sophisticated idea, they will think of our solution as a more reliable, cost-effective, and less power solution. So they will use it. So I think for a long period of time, they will use three solutions all together in different places.
Vik: Amazing. And do you see RF E-tube technology also scaling to faster lane rates as we go ahead in the future?
Sean: Yeah, so we are already working on the next generation. So per cable core, it will be close to 1 terabit per second data rate. We are already working on it. So that will be our next generation chips and E-tube solution.
Vik: Fantastic. That’s always the question to ask because at some point we’ll be like, oh, if E-tube runs out of steam, just as in copper, then what do we do? We have to then resort to optics. But the nice thing about waveguides—and this is more of an electromagnetics thing—is that there is no real physical upper limit to frequency in a wave guiding system.
Sean: Exactly, exactly. If you think about copper, why does copper have limited bandwidth? Simply because copper has too many electrons, and those electrons when they flow, create a skin effect. And because of the skin effect, bandwidth is limited. So, what I’m saying is that the bandwidth limitation simply comes from the existence of electrons. And this plastic, it doesn’t have electrons. So, in terms of bandwidth, it is unlimited. Bandwidth is unlimited. So, once our silicon works, we don’t have to really change this cable at all. Just scale the silicon, then it will be used for a higher data rate.
Vik: That’s fantastic. Just before the show, we were talking about how we got into this technology and all that. And you mentioned that in spite of all your technological advancement on this interconnect technology using RF, your first customer was not related to RF at all. Can you tell us that story?
Sean: Oh, yeah. So, I have been thinking about the data center, that’s why I started this company using E-tube technology. But our first customer, they were display companies. For example, they wanted to have a very thin wallpaper-type of display. And in order to make it so the thickness is very thin, 3 mm, they cannot put silicon in the panel. It’s an OLED panel and they cannot put a decoder or encoder inside the panel at all. So, they put every TV function outside of the panel, they put that in the set-top box. So, between the set-top box and the panel, they need a high-speed interconnect because they cannot decompress and they cannot decode in the panel at all. So, they need to transmit raw video data, uncompressed video data to the panel. So, the data rate suddenly becomes 100 gigabits per second, 400 gigabits per second. Very high data rate for consumer electronics applications.
So, they looked for some other interconnect other than copper. Copper is too thick, they cannot really fit it in. And they looked for optics, but optics is too expensive for consumer electronics, and also is power hungry, and in terms of thermal, it’s too much thermal dissipation. So, they don’t want to use optics. So, they looked at our solution and they thought, oh, this is the one, very thin, low power, longer distance. So, they put some money for us to develop the TV application, consumer application E-tube solution. So, that was our first customer.
Vik: That’s amazing. Is that actually adopted? Or did they find another solution out of it?
Sean: Unfortunately, they delayed their project and they stayed with the copper solution.
Vik: Yeah, yeah. No, that’s a fantastic thing. It gave rise to a technology that now has potential uses in something much, much larger than set-top boxes and TVs. The data center interconnect is a far bigger opportunity than any of that stuff. So, it’s good in a way, it’s okay if they didn’t adopt it. At the end of the day, you have a far larger opportunity because of that one problem you tried to solve in a different area.
Sean: Yeah, exactly. But I think eventually our generation, people will use more data in the future. Not just in the data center. For example, one of our investors is Bosch. Bosch is the number one automotive component company. And I think they invested in our company because we have a very cool long-distance and low-power interconnect solution. I think they’re thinking of using our E-tube solution in their automotive network. So, in the near future, I’m pretty sure we’ll be in the automotive market as well. Of course, we need to spend a lot more cycles and time with our customers in the data center first.
Vik: Awesome. Thank you so much, Sean. This was very interesting. Thank you for teaching us all of these cool technologies you’re working on. Definitely check out Point2 technologies. Your website has all this information that people can go check out. And you’re often at conferences. So, if any of the people listening here, they should stop by and check out your demos like I did at OCP APAC in Taiwan. It’s really fun to see this stuff in action.
Post-Interview Discussion
Austin: All right, Vik, that was a fascinating conversation. Now, what I’m super interested in is your perspective on will this actually ship?
Vik: Let’s start with the timeline, because this technology is relatively new. It’s old but new. It’s been around for a long time, but it’s not been in deployment. And now you are seeing a lot more alternative technologies coming up. So we would put it at least a few years away. It’s definitely not something that you will see in mass production in hyperscalers or big data centers today.
But that being said, today most of the scale-up in a rack happens with copper. And it’s okay at 200 gigs per lane. You can even run bidirectional, which means you can run transmit and receive on the same copper cable. It all fits within a single rack, no problem. We are now going from single-rack scale-up domains to multi-rack scale-up domains. Now distances are increasing. And now when you put racks next to each other, you need to hook them up, you need much more reach than copper can even provide. And now the question is, what are you going to do? Are you going to use optics? Maybe, but it’s very power expensive if you’re going to use pluggable transceivers. So the option is, okay, let’s go to co-packaged optics or near-packaged optics. Yes, but that’s an entirely new ecosystem that nobody’s ever done before. So it’s a big step up from what we know how to do.
So you’ve got that copper doesn’t work, optics works, but it’s overkill perhaps. And so what are you left with? You’re left with alternative technologies like microLEDs or micro-VCSELs. The problem with these technologies is that they don’t run at the speed of copper, which means 200 gigs per lane runs within a single rack today, even bidirectionally. These technologies don’t do that. Micro-VCSELs could probably run at 50 gig and microLEDs run at maybe three gig, so they are much slower. So what you have to do is convert the 200 gigs per lane that’s coming out of the XPU or the switch. You need to convert that into something much, much slower. And that is done with something called a gearbox chip. The gearbox converts this fast but narrow to slow but wide. And that has its penalty because this chip needs its own energy to make this conversion. It’s another point of failure potentially. And the complexity is just higher.
The beauty of RF over a plastic waveguide is that it runs natively narrow but fast. So it’s very much copper-like. It’s not microLED, it’s not micro-VCSEL, it’s not indium phosphide lasers. It’s very much copper. And if it’s very much copper, it uses very simple electronics. It’s just a chip and it just goes over a plastic cable and that’s about it. There’s no complex integration, there’s no putting III-V materials like lasers. You can’t make silicon produce light. And so you have to put all kinds of other materials—indium phosphide, gallium arsenide, gallium nitride—on top of silicon. So much complexity, so many integration problems. This does not have any of that problem. All you have is a silicon chip that’s not even cutting-edge silicon, it’s like 55 nanometers. We could go to 22 nanometers and still get so much benefit, but it’s still not even a FinFET, let alone a gate-all-around three nanometer. No, it’s none of that complexity. It’s a simple chip. It works on simple electronics that are based on radio frequency stuff, which is very, very well established over many years of chips in phones and we know how to do this stuff. And you can run this without any gearboxing chips because it runs natively at 200 gigs per lane. So you can just put it end-to-end and if it has up to 5-meter reach, you can at least connect two racks. And if you can have 3 meters, you can even connect up stuff within a single rack and plastic is much less weight and much more density is possible within a single scale-up domain. So it has all these advantages that the optics world doesn’t offer right now.
Austin: Yeah, interesting. Okay, so I like that you make the point there’s no III-V material, there’s no indium phosphide bottleneck. This is just silicon. It’s older technology that you could get built at a foundational node at many sort of trailing-edge facilities. One question I have, is the antenna also just silicon-based then?
Vik: Yeah, it’s completely integrated on the chip. The beauty of this is that because of the frequency being 90 GHz or 225 GHz, the wavelengths are actually pretty small, which means that you don’t have to have antennas external to the chip. Because antennas are always a function of wavelength, like a quarter wavelength could be an antenna, or a half wavelength. So you can make these antennas on the chip because they are actually small enough at close to 100 GHz frequency. So everything is integrated. You’ve got transmit and receive are just antennas and they’re all on the chip. So all you do is take a chip and put it into the OSFP module. Done. That’s it.
Austin: Fascinating. What’s very intriguing to me is like Credo sort of invented this new market with AECs and it was pretty much just plug-and-play. You know, you just use the copper cable but it has some DSPs to help make sure that that signal can transmit further and be cleaned up and re-timed as necessary. But anyway, to the actual data center operator, it’s sort of just plug-and-play. It’s just like buy this and use it, everything works. It kind of feels like what you’re saying is that not only are these active RF cables simple to make and conceptually, there’s no big concerns about the materials or bottlenecks or anything, but it also feels like it’s kind of just plug-and-play. Is that a fair statement if you’re a data center buyer?
Vik: Yes, it’s completely plug-and-play because the electrical interface is the only thing that the rack will see. As long as the plug goes into the electrical side and it works, what happens beyond that is not anybody’s problem but the cable maker. Your cable maker’s responsibility is to take this signal electrically from one place and deliver it electrically in the other place. Whether you want to do that with microLED or optics or RF, it’s not anybody’s problem. The concern is, is it reliable? I think Sean mentioned in the podcast, they are going to put this to the test. They want to make sure that it works because if you’re going to pitch anything to the scale-up domain, stuff has to work because there are way, way, way too many cables even in a single rack. We are talking about 144 GPUs will have above 20,000 cables. You can’t have any one of them fail and that’s a single rack. Now multiply that amongst racks. The reliability requirement of the cable is extraordinary. And I don’t think this has been something that any technology has proven out at scale. We don’t know if CPO can do that. We don’t know if NPO can do that. No technology has that data right now. Copper is the only thing that has worked and we know that we can deploy a data center full of copper cables and it’s going to work. So ultimately, data centers and hyperscalers want to stay in copper as far as possible, but it’s becoming extremely difficult. So the search for a new technology is on.
Austin: Yes. Yes. Okay. Well, it sounds like reliability is the big thing. It sounds like the technology checks out. My thoughts around the cost being lower probably checks out just with the simple process node and keeping everything silicon. And even the plastic waveguide was quite simple. And then the mention of it actually reducing the amount of connections that you need, cables you need. So it feels like the cost checks out. Really the question is reliability. And I think to your point, we’re early in NPO deployments and CPO deployments, even like rack to switch or whatever. But obviously the big question for Point2 and for this technology broadly is will it be reliable when there’s hundreds of thousands of them?
Vik: Yeah, I fundamentally love the idea that there is another option that is not copper and not optics. And if it’s RF, it’s a bit closer to my heart because that’s my whole expertise that I have built over a couple of decades in the semi industry. So, personally, I’m rooting for this to work just because it’s RF and I love RF. So, let’s see. Exciting stuff.
Austin: Exciting stuff. Let’s see. All right, listeners, I hope you enjoyed this conversation with Point2 and our thoughts afterward. Leave us comments below, send us an email, and we’ll see you in the next episode.
