Weebit Nano Limited (WBT) Earnings Call Transcript & Summary

My name is Ronn Bechler. I'm with Automic Markets. A bit of housekeeping before we start. If you have a question, we'll be having a Q&A session at the end of the presentation that Coby will be giving. So please put your question in the Q&A box down the bottom, and I'll moderate the questions at the end of the presentation. I might hand the microphone across to Coby to take us through the Q1 FY '23 update on Weebit. It's been a very busy quarter and a lot been going on. So I look forward to hearing the presentation, and then we'll move on to Q&A after back of that. Thanks, Coby.

Okay. Thanks, Ronn. Yes, it has been a very busy quarter and, really, a lot of activity and ramping up to even more activity in the upcoming quarter. Last quarter, we had several key milestones. Of course, the most important one was announced recently with the qualification of our ReRAM, showing that we meet the industry standard requirements for production quality, and that's gotten very good response throughout. The selector, again, another announcement that was done recently. This is something that's happening in the background. But nevertheless, we're continuing to develop the selector and, this time, we even managed to show that we can use it in embedded applications, which is going to be another very important game changer even, I can say. We're expecting the wafers back from SkyWater now, it's taped out and it should be back soon. We'll be doing the qualification there immediately after that. There's really good cooperation now with SkyWater. And in general, there's really more and more activity going on now with partners, Tier-1 fabs, and really even more advanced than I expected in the past. So if you look back a year, and I actually look back 5 years since I joined 5 years ago, it was -- it's really amazing to see all the stuff that we managed to do. But in the last year, we announced that we reached 28 nanometers and then we already are working on the 22. As you know, we're planning to take out also on 22 nanometers. And then the module, we showed that it was fully functional, and we've had to tape out the SkyWater's. And now with the 2 recent releases, I'm extremely proud of the progress made over this past year. Talking about qualification, and it's really important to explain what qualification is and why it's so important. Qualification is a process that's defined by industry standards. We took the definition of JEDEC, which is the common one for nonvolatile memories. It defines a really detailed testing procedures that you need to go to. Now it's -- you need to do 3 separate manufacturing lots. So 3x, you go through manufacturing. From each one of those lots, you'd take different wafers. From each one of the wafers, you'd take random dies. Overall, it's been many, many hundreds of dies that we tested. And it's very rigorous testing for each one of those, going through all kinds of stress measurements. The highlight is, of course, that when you put them in ovens that go up to very high temperatures and checks that the chips are still working. So it's really a very rigorous procedure that takes many months to complete because of that. But once you do complete it, you have a certification. You can show the results to customers, to other fabs, to partners. And you can show them that, yes, we passed it. All of these hundreds of chips, all of them passed it. All of them are acting the same. When we do mass production, all of them will act the same. You won't have different chips doing different things or all kinds of things like that. So that's really a very important milestone. And Leti, though some people tell me, "Oh, Leti is an R&D fab." But Leti's facilities are state-of-the-art. They are top-notch, like the most advanced fabs in the world. Not many fabs in the world, not many foundries in the world have the ability to work at 28 or 22 nanometers. I mean, I think the number is maybe a dozen, at best, and Leti is one of them. So it's really an advanced facility and these tests, too, impress many people. And they actually enabled us to engage with much more advanced fabs. We're talking now to Tier-1 fabs, making good progress with some of them and showing the results to customers that are also impressed by the results. So really, a very exciting milestone that we achieved. Now this is a milestone that we're immediately continuing to do qualification on, even more rigorous and more extreme results. And the way that you do the qualification is you start at a certain level and then you keep pushing the limits, and you keep going to higher temperatures and more endurance. And you just keep showing that the technology can withstand mass production at even more extreme results. So we're continuing to do that work and additional qualification even at more rigorous requirements. I mentioned the selector. The people that have been following Weebit know that we are working on this technology in the background. This is the technology, which originally we focused on 4 discrete or stand-alone chips. And in the stand-alone world, small memory dots, we call them, small memory cells, are really a key requirement. You need to shrink the size of the array as much as we can. It's real estate, just like any other real estate, you want to have as much as -- as much memory as you can on a small -- a piece of silicon as you can, and the selector is an enabler to that. I'm very happy that the team found a way to make this work even in embedded. So we are looking at, in the future, not now because it still isn't ready, but we will be looking at how we can use it also in embedded applications and get another advantage over competition, over other kinds of memory by shrinking the memory arrays even in embedded applications. And this is looking very good, especially for applications that will need very large memories even in embedded like AI and automotive. And it's really important because we also are managing to make the selector use standard materials, which is really not trivial. And I'm looking forward to see additional progress on this front. SkyWater, of course, is a key activity for us. We went through a long process of transferring the technology to them, doing the tape-out and then having to wait for the wafers. The wafers are now moving very fast. I believe we'll be getting them in the near future. We should be getting them clearly before the end of the year. And as soon as we get them, we'll go immediately into qualification. Qualification is moving -- will move forward, well -- so by the beginning of next year. And the first half of next year, we'll be expecting to finalize it. And we'll be -- we're already talking to SkyWater customers. We're working with them. If some of you guys look at their website, you'll see that there's a page for Weebit there and really good cooperation going on now. In terms of the market engagement, so I mentioned we have really nice results with the qualification. We've already shown intermediate results to other fabs. I think that what's really happening right now is the market is realizing that this is the time for ReRAM. A year ago, 2 years ago, when I would to talk to people, they would always say, "Oh, yes, that's future technology, that emerging technology that will be available some time." Today, I think the market realizes ReRAM is here. Everyone needs it. They're asking for it. We've had more than 1 case where an advanced -- Tier-1 fab told us they lost a major deal with a major customer because they didn't have advanced enough nonvolatile memory. They showed us the specifications, and we put them next to our ReRAM, and you can see the match. We've seen, especially since TSMC, which is the largest fab, announced that they have a ReRAM technology. It really sent a shock wave, I think, through the industry. Everyone now needs to have a ReRAM technology to compete with TSMC. And when they looked around, Weebit is basically the key one that they see. So we have these engagements now, and it's really exciting to talk to all of these guys and move forward. And some of them are already in different stages of evaluating the technology. So really looking very -- I try to be cautious in the words -- in my words, but it really is looking nice. So that's the activity there. We're also talking to potential customers, of course, and pushing them forward in parallel. There's this situation, as you guys know, this triangle between us, the customer that will want to embed our technology in their chip and then the fab. And different customers use different fabs, and they want us to work with specific fabs and things like that. We need to get these triangles formed, and there's a lot of work to be done here to do that, to push that forward, but there's constant progress. And of course, we've been attending a lot of conferences, both technical conferences, and I've been involved in some investor conferences. And our name is really out there now. And even when you look, I said the world is realizing that now is the time for ReRAM. I think nothing exemplifies it more than the report from Yole. I consider Yole the #1 analyst for our market. And these guys in the past were more cautious, but now they basically have the embedded ReRAM market. You can see here going from practically 0, which is, today, to about USD 1 billion in just a few years. I think that -- I can sense that this is what the market wants. This is the direction of the market. I clearly see the demand here. And over the next few years, as we roll the technology out to more fabs and engage with more customers. There's the famous book about the tipping point, and you get to that point where you have several new customers, they're having good results, and then suddenly that whole thing tips over and you have the flood coming. So I'm looking forward to that. So basically, looking forward, and I'm looking at what are our goals. We need to look forward further than just another month or 2, and so we're looking at the targets until the middle of next year. Of course, I mentioned we need to conclude the qualification of the embedded ReRAM at SkyWater. That's going to be a key activity. We'll get the wafers and we'll start the qualification. We're having -- we're making good progress with Tier-1 fabs. And I really set the goal for the team that we need to have a big advanced fab already licensing our technology by the middle of next year. Of course, with customers, we're continuing the work. We said we'll close customer agreements. And we are working on it very strongly to push that forward. And I want to see more of this activity and seeing actual customers signing licensing agreements. We will be -- I was talking about continuing the qualification, and we do plan to continue qualification, take it to more and more extreme levels to show real superiority. Today -- by the way, today, our qualification already shows that we are better than Flash. And moving forward, we want to show real superiority in this domain and very strong results. And then scaling to 22, and we're actually already working on even below 22. I mentioned people don't understand these numbers, 22 nanometers. There's a dozen, at most, I think, fabs that are capable of working at 22. To go down below that, and it's even -- it even becomes more extreme. So we're pushing it. We're pushing the limits there. We want to get to the places where the most advanced designs are, and including AI and automotive and all the really advanced things. So there's really a lot of work going on there. And continued R&D, that's the heart of it all, pushing the technology further, making the ReRAM cell smaller, making the ReRAM cell more advanced. A lot of work, a lot of work. We have an amazing team, and I'm really excited about all this. So that's -- I think that's -- we can stop here and open the floor to questions.

Thanks, Coby. A lot going on over the last quarter and, obviously, a busy run into the end of FY '23. We've got a bunch of questions that have come in, and we'll try and cover as many as we can in the time allotted. [Operator Instructions] Maybe if we turn to, first of all, that slide that you like referencing, Slide 8, which is the Yole market share slide. Can you just talk through -- you said ReRAM's expected market share of 33% of that $957 million annual market size, we got a question asking what's the corresponding 67%?

So by the way, it wasn't Weebit's market share. Yole was looking at the total market for emerging nonvolatile memories. The 2 leading technologies are MRAM and ReRAM, MRAM being magnetic RAM. MRAM has been in the market already for some time. It's actually out in the market first. So there are companies now that are commercially selling MRAM. And so Yole has looked at it and estimates that MRAM will have 2/3 of the market and ReRAM will have 1/3. So what you saw in that slide is the ReRAM part of it. I personally can tell you that, as we talk more and more to fabs and to customers that are already using MRAM, we're getting a very clear message, I can say, that people are not happy with MRAM. MRAM is way too complex to manufacture. It is nonstandard in its extremity. You cannot manufacture MRAM inside the fab. You need to have a separate facility for it. The magnetic material is a terrible contaminant in the fab. And so we're getting the indications. And then by the way, you have so many layers there and everything. The MRAM adds about 30% to 40% to the wafer cost. We add about 5%, 7% the wafer cost. So you can understand just commercially, when people will get to it, we don't require new tools and new materials like MRAM requires. We're much simpler and easier, cost effective. So I believe that, that 2/3 MRAM, 1/3 ReRAM will actually change to the ReRAM side, and we'll see a much stronger ReRAM side once we begin the Israeli mass production.

I think the point to make also is the market opportunity is so large that -- and there won't be 1 specific window that it doesn't really matter. There's opportunity [ we saw that ] even a 10% or 20% share [ of the ] overall market is still massive. You don't need to have 50% share to be happy with the outcome for that.

As you all wrote there, 33% is $1 billion, which is a pretty nice growth rate already. It's very hard to grow faster than that anyway, and it doesn't stop there. It keeps growing. The market is huge. We're talking about both MRAM and ReRAM growing rapidly.

We've got a few questions around nodes, sizes and qualifications and so on. But maybe just a question for those that are slightly newer on this webinar, can you just outline what exactly is qualification? And why is it so important to a company like Weebit?

Okay. So first of all, we use the terms, and I apologize -- in the industry, there are a few terms which are used interchangeably quite a bit. Nodes, geometries, nanometers, they all refer to the [ same technology ]. So when I say a small geometry, I'm talking about something that is smaller than 22 nanometers. Similarly, when I say small nodes, I try not to use the terminal just to confuse people less. So -- and as I mentioned, the market has been going down steadily. There's Moore's Law, for those who know it, and everything has been shrinking. I saw -- a month ago, I saw an announcement that Samsung is planning by 2027 to be manufacturing at 1.4 nanometers. That's like -- it's not understandable at all. I mean, I just can't understand how they can do that, but the market keeps shrinking. And the number of companies that can really work in small geometries, as I mentioned, is very, very small. It's really a challenge. So we're pushing forward. You asked about qualification. And as I explained, it really is the best indication. The market can get the tour ready for mass production. So by the way, I have people say, "Wait a minute, what do you mean that Weebit is the only one out there now?" And maybe talking a little bit about the competitive landscape, just a few years ago, there were really 4 smaller companies that were focused on ReRAM. CrossBar was maybe the leading one. They were already working with production fabs, et cetera. But the problem was that they were using nonstandard materials. They raised 150 -- close to USD 150 million, but at the end of the day, they didn't manage to get it working. And some Chinese firm took over, and we don't see them in the market anymore. There was another company called Adesto that was acquired by Dialog. Adesto had several products. Dialog acquired them. And then Dialog was acquired by a big Japanese company called Renesas. And Renesas, it doesn't have any focus on memory or anything. So the ReRAM technology there has kind of been put aside. And Weebit now is really the leading player in the market. So we're very happy that the spotlight is on us. The qualification results show that we've passed where the others were. And now we're in a good position, ready to go, and doing this move to commercialization.

Thanks, Coby. And just to confirm, the qualification that was recently announced, what node was that done at?

That was on the 130-nanometer technology from Leti. Those are the wafers that we did. If you remember, a year ago, we manufactured the wafers for 130, the modules. And so that's what we tested.

And the 130 is what SkyWater is starting with as well?

Yes, yes. We will -- the SkyWater manufacturing is also with 130, so we'll be qualifying there. We are taping out soon at 22, and we'll -- of course, we'll have to wait until those wafers will get out of the fab. But once they get out of the fab, we'll be starting to qualify those, and that's already going to be a much more advanced node, yes.

And then the qualification that was done, you mentioned in Leti's fab, and so...

No, the qualification is done in special labs. Actually, most of the qualification was done in our lab. We set up a lab in our office. A lot of it was done in our lab. And then we worked with specialized labs that have the specialized ovens and everything to do that work. So qualification was actually done by Weebit's team in Israel, mostly.

Okay. And then you mentioned about the plans to tape out 22 nanometers. Will that be done within Leti's fab? Or would be some other fab be involved?

No, that's with one of the Tier-1 fabs. That's going to be done at a fab of one of the big fabs.

Okay. And you mentioned Tier-1 and Tier-2 fabs, but the company is talking to other's Tier 1 and the Tier 2, looking at different node sizes in terms of what they're wanting from Weebit, or it's just they're not so worried about 130 or 22. Like how are they thinking about it in terms of their engagement with you?

Normally, when we talk about Tier 1, we're talking about the ones that are going below 28, going to 28 and below. That requires huge investments, tens of billions of dollars, and even more. That requires really mass production to justify all those costs. Some of the Tier-1 fabs also work at large geometries. So some of them work at 180 and even beyond 180, 130, et cetera. So -- but some of them also really have a big focus on just the smaller geometries. It's kind of across the board. But the key element here is, when you go below 28 nanometers, you need to have very, very deep pockets and really mass production.

We've got some questions that came in from people that weren't able to be on this, but they've e-mailed it through. And looking at embedded and discrete, there's a question around what prospects are there for moving discrete, as in stand-alone technology, forward now that the selector has been tested? And what do you need to do in development for it?

The selector technology, I mentioned it when we started working on it. It's a very, very complex technology. We're going through different stages of development, and there still is more work to be done there. So I don't want to confuse people. It's -- we went through a certain important milestone, but there still is -- in order to get the selector technology to the point where it can be qualified, we still have a lot of work to do.

And so then when you think about sales avenues outside of embedded, that's further down the track as well?

Yes. I think I always said that the near-term low-hanging fruit is embedded. That is my big focus. That is where the initial revenue will come from, and we're focused on that. The discrete is a big market, but it requires the selector, and that's going to be the midterm. And I always said the longer term will be more and more fixed. So that's kind of -- that hasn't changed. We're very focused right now on getting revenues from the embedded. And the selector technology and the discrete is continuing in the background.

We've had a couple of questions around some of the previous conversations on partnerships and letters of intent and so on that we talk about in terms of the Indian Institute of Technology Delhi on the neuromorphic project, the XTX, SiEn. Can you give us an update on how those various collaborations are progressing?

So I think I'll break that question into 2 questions. The IIT Delhi and also Politecnico di Milano and other research institutes and the Technion in Israel, those are institutes, those are research institutes that we're working with them on the neuromorphic technology, neuromorphic and processing in memory and other advanced technologies. And that's going very well. Every once in a while, we announce a paper. Actually, with IIT Delhi, we -- I think there was a paper in a conference just a month or 2 months ago. So that's progressing very nicely. China is a very good question, and I think it's a question but should be answered on its own. I think everyone here is aware of what's happening in China, with the geopolitical tensions between the U.S. and China. I will add that during the corona period, it was very, very difficult to work with China. You couldn't travel to China. Communication was more difficult. So both of those things have pushed us to -- I mean, we managed to get an agreement with SkyWater, and that was a key step forward. And that we're now in discussions with many, I guess, I can call it, Western foundries, Western customers. That market is huge. And right now, we're focused on that. I think it's -- I'm looking for ways to lower the risk for Weebit, and I'm sensing that it is lower risk and higher potential for us to focus on U.S., Europe, Western world. And it's not like we totally cut off our ties in China, but we're clearly not pushing strongly there. And by the way, you still can't travel to China. So it's -- I guess, I'm very happy with where we are today. It gave us a huge push with the big fabs in the Western world.

Cyber threats and cyber risks have become quite topical here in Australia off the back of some [ risk ] inflow. Can you outline to us how Weebit's memory technology helps improve cybersecurity?

There are actually several ways. The ReRAM technology -- first of all, many of the cyber applications are such that require AI and so on, and they want to go down to the small geometries. And they go down to 22, 16, 12 and even smaller than that. As I mentioned many times, Flash does not scale to those geometries. It gets stuck at 40, 28, if you really push it, and it's not economical already. So what they need to do is they have to have a separate chip for nonvolatile memory that's kind of glued on to the main chip and so on. But the mere fact that you have separate chips opens a communication line between those chips to eavesdropping. And that's one area of concern of people. In embedded ReRAM, you don't have that. We are embedded in the chip. We're an integral part of the chip. So that's one of the advantages that we have. Weebit is also -- I mean, our technology is immune to interference. I mentioned MRAM before, and one of the concerns about MRAM is that because it's magnetic, if you go near a strong magnetic field, things can be erased or things can be altered. If people are smart enough, they'll find ways to do it even in other ways. ReRAM is inherently immune to radiation, which is another point. So we have several advantages in the security space, and some security companies are actually looking at that.

Keeping at sort of the high level, and we've talked about cyber risk, what about quantum computing? Do our chips have a role in quantum computing?

It's very exciting...

Memory technology, sorry, not actually...

On memory? Yes, yes, because there's also the embedded. Quantum computing is very exciting. But at this point, at least, it's a different kind of technology, and we don't really have -- we don't really have a role to play there in quantum computing. So we're looking at it out of personal curiosity, but that's not where Weebit is heading now.

If we think about -- and you had that slide at the end about key milestones to deliver over the remainder of FY '23. But a question that's come in that's asked what other technology developments are we expecting, for example, qualifications at higher standards, smaller nodes. Can you give a bit of a sense of how you're thinking about where you want to take the technology over the remainder of this financial year?

So we're constantly pushing the envelope. And also, there are 2 directions: first of all, just pushing the envelope; and second, picking what we have and making it robust so that it can go to mass production. Those are 2 parallel activities that we're working on. So I talked about qualification. At this point, the qualification was done at what's called the industrial grade. We're now starting qualification for automotive grade for more advanced. The requirements, in terms of temperatures, endurance, et cetera, are more extreme. And we'll be pushing that. There are multiple grades of qualification that you can get to. We were confident that we can go several grades forward already. We have very high confidence that we can achieve those. The question is how far. And that's where the other activity comes in of making the technology more robust without going into technical details. But you want to make -- when you -- the memory at the end is chemistry, and you have atoms and all of that in there. And when you write a 1 or you write a 0 into the memory, it's not digital, it's analog. And you want to make sure that the differentiation between a 0 and a 1 is strong enough and wide enough so that it's very obvious, and then you can do a lot more things. So we're constantly pushing to make the 0 and the 1 more distinct, a wider gap between them, which will enable us to move forward to more extreme qualification levels and so on. So there's really a lot of work on all of that, just making the technology more robust. And that's something that's never ending, by the way. That's something that it won't end in 5 years or in -- I don't know what. You have to keep constantly pushing the technology to give better and better and better results because that's what the market wants. In addition, as I said, we're working now. There's a big focus on the 22-nanometer. We're already analyzing geometries, which are lower than that, and working with some of the Tier-1 fabs on looking at how we can go down to smaller geometries. And we'll be pushing in that direction. The whole market is pushing in that direction. So a lot of very interesting work.

You mentioned some of the low-hanging fruit and the focus on embedded. A lot of question that's come in on the embedded selector. Can you explain what types of markets it will excel in, and if the embedded selector can actually create whole new markets and products?

An embedded selector has several really important advantages. The most important one, of course, is just size. The selector enables us to generate -- to build a much larger array on a much smaller piece of silicon. So in other words, on the same piece of silicon, you can put a lot more memory, which is actually what most people are looking for. AI is the type of application that no matter how much memory you give them, they always want more. So if we own the same piece of silicon that they have reserved for memory and can give them more memory, this is huge for them. That is a very big advantage. So we're looking at that. Another thing, which is really further out, that's something that we are looking at, once you start using the selector, you can actually start thinking of 3-dimensional. People here have heard in the past that there's 3D NAND. And the way that Flash is dealing with the fact that it can shrink is -- it's going up. Now it's -- they're building skyscrapers of memory. And on the same piece of silicon, they have more and more memory just because they pack it in piles. We actually was -- as a selector, we will potentially be able to do the same thing and start going through dimensional. So those are further-away milestones, it's not something that will happen tomorrow. I don't want anyone to misunderstand me. But getting the selector to work embedded, once it works and it's ready for mass production, and then getting it into the embedded will open -- will make us much more competitive because we'll be able to have much more memory on the same piece of silicon and potentially even both 3-dimensional at a certain point.

We have a question that's come in, and I'll state the question, but I'll then rephrase it because it's a bit hard to actually answer. And the question is, if everything goes according to plan, when do you estimate mass production will occur? Maybe I can just -- because, ultimately, you don't know what's going to plan because -- if you gave an indication of time frames and what you want to do by the end of FY '23 in terms of signing up a new fab or a new customer, but let's just say we announced -- Weebit announces that a new fab or a new customer has been signed up, sort of roughly what's the sort of time frame between that and actually having mass production start?

So first of all, the first answer is, with SkyWater, we are moving forward, and we're expecting the wafers, and then we'll do qualification. And we still need to see how much time exactly that qualification will take, but it's order of magnitude of a few months that we'll do it. We already did qualification once. By the way, that's another advantage of the fact that we did qualification on the Leti wafers. We kind of -- now we have that experience, and we know what we need to do with SkyWater. So we're hoping to have that qualification done even a little bit faster. And then once we're qualified, we can go into mass production. But the other side of the story, and I always talk about this triangle, you have the fab and the fab needs to be ready, but then you need the customers and the customers to come on board. The customers today -- and especially the atmosphere in the semiconductor space is -- it's a little reserved, I would say. Whoever has been following the semiconductor market, the big players have had challenges and share prices and everything. So there is even more hesitancy than normal. But we are moving forward with them. We are making progress. We're now working with SkyWater, talking to several of their customers about adopting our technology. Those customers, once they embed the technology into their design now, they need to design their product with our memory units and get it traded for tape-outs. And that takes several months, that can take a year, depending on the design, and then they go to mass production. So that's kind of the order of magnitude to give you the feeling with SkyWater. And you can follow the process we did with SkyWater since we signed the agreement and so on. With the larger fabs, I expect things to be -- to go faster, transferring the technology, getting things done, not orders-of-magnitude faster, but clearly, both, we have more experience and the larger fabs have more resources to put on it and push them faster. So -- it's -- the answer is it's a ramp-up to get to mass production. We need to close an agreement with the fab, transfer the technology, which, as you guys know, can take a few months and then you do the tape-out. You wait for the wafers, and you get it back, you do the qualification. And in parallel, you work with customers. I think with the -- as we move forward, we'll be able to bring more and more customers on board in parallel to the technology transfer and get things moving faster.

In terms of the FY '23 goals that you've set out today, what are the biggest challenges in your -- in the company's ability to achieve those aspirations of those goals for FY '23?

Time, which, if everything takes long, and I'm a very impatient person so this drives me nuts. It's -- I don't know where I'm going in semiconductors, but that's the way it works. So in semiconductors, you can send wafers for manufacturing. You just have to sit there for many months. And if they rush them through the fab and everything, you can get it in maybe a little bit faster. But no matter how you look at it, it takes a lot of time. So I think one of the key things is things just take time. We're also in a period where the customers -- because of what's happening in the marketplace. And we're hearing about companies that their results -- as semiconductor companies are looking at their results and everything, if normally, they're very cautious, now, you can -- there's a little more of that obstacle to overcome. We're pushing it. I know we'll make it. I don't have doubts that we will -- we won't get those customers on board. And we will get those fabs. It's maybe a little bit more of a challenge, but we're making good progress with them. We have very, very exciting talks with some of them. And it's going to be -- for me, it's really a very exciting period right now.

Maybe on that note, because we've hit 45 minutes and we're running out of time, we might wrap up this meet-the-CEO interview. There's obviously going to be a lot of opportunities for people to meet with you, ask you questions face-to-face in a couple of weeks' time when you come out to Australia. So just for those online...

Looking forward to it.

We're looking forward to it, too. It's been almost 3 years, Coby. But we've got you in Sydney on Monday, Tuesday, Wednesday, which is the -- I'm just checking the days now, 14th through the 16th of November; Melbourne on the 17th and 18th of November; and then Perth on the 21st of November. So if anyone -- and there's a series of events, both in Melbourne, Sydney and Perth, for a group briefing for investors. And I think, Coby, you're also -- you're obviously presenting at TechOpps this week. And potentially another conference, while you're out here, we're just waiting on confirmation. So there's a lot of activities going on in the AGM. But we certainly encourage shareholders or potential shareholders to take advantage of Coby and Dadi, David Perlmutter, the Chairman, will be out at the same time in a couple of weeks, to meet with them face-to-face, come to the briefings and hear more about Weebit, the technology and the growth opportunities that lie ahead. Coby, maybe just on that note, if you'd like to just wrap up and leave us with a thought before we finish up today, that would be great.

Yes. So it's -- as I said, it's very exciting times for Weebit. You guys have seen, we have several key milestones ahead of us. We're waiting for the SkyWater wafers. We're taping out the 22 with -- there's more activity, of course, the customer activity and everything, really exciting times. For me, it's very exciting to come back to Australia after -- it's almost 3 years now. It's getting close to 3 years. I really missed it. I really look forward to meeting people face-to-face. And so I would love to have all of you guys come to these briefings and meet you in person. And Dadi will be with me, which is also, I think, an important point. So together, we'll be able to answer your questions. Looking forward to it.

Coby, thank you very much. And to everyone online, thank you for joining us this afternoon. Have a great rest of day. And hopefully, we'll see you all when Coby and Dadi are out here in a couple of weeks' time. Thank you.

Great. Thank you.