Weebit Nano Limited (WBT) Earnings Call Transcript & Summary

Okay. Well, good to be here. As always, good to see you guys and whoever is online. I came here for the Semiconductor Australia Show, which I think was a very good one, So some of you I saw there. And it was -- the goal was really to raise awareness of the semiconductor industry in Australia and especially get some government people to start thinking about it because that's really what's needed here. So I know, I'm Israeli, I'm not here, and I'm theoretically not supposed to have a lot of interest, but I love this country, and I'm kind of involved in pushing that forward and having the discussions and also having the discussions with the ASX and so on. So we're trying to push this forward because there really is a lot to do here in Australia, but I'm not here to talk about the Australia semiconductor industry. The title here says accelerating progress with commercial deals. And I think a lot of you guys are looking at and saying, yes, right. This guy has been telling us for so long that he's working so hard and he's making progress and where is the deal? Where is the beef and where is the -- where are the deals? Well, it's -- I'll say it again, and I'll talk about it when we go through these slides, but it is a very difficult industry. We are really busy. You have no idea how busy we are now with so many potential customers that we're engaged with, so many people that are interested in ReRAM and more than just they're actually evaluating. They're actually looking at how to make it fit in. I know we haven't announced anything, but we have been making progress, and I am taking this big risk of saying that I really believe we should probably be able to announce something by the end of the year. So it's not only up to us. It's always the other -- the other side and these big companies, their lawyers take forever to respond to anything. But it is -- we are progressing with commercial deals. So the quick update, and you guys know the intro, I'll skip that slide, but really talk a little bit -- these are a few slides that I did show also at Semiconductor Australia. And I think it's important to remember. Semiconductors this is a very big and important industry. This is actually -- I believe, I'm prejudice, but I believe this is the #1 industry in the world. This is the most important, most strategic industry, and it's growing very fast. It's growing very fast. You can see here today, it's roughly estimated at [ 600 or not ] today, a couple of years ago, a few years ago. It's expected to reach USD 1 trillion in -- by the end of this decade. It's really growing fast. You can see here the different domains that we're talking about. And I think it's always important to remember that the vast amount of investment, these huge investments that are being made. And actually, you look at -- at this map, and there are several things here that I want to point out. First of all, the cost of building a new fab. It is constantly going up. The new fabs, Intel announced a new fab in Israel that's building, which is $25 billion, Samsung announced a fab, which is $25 billion. I've seen even more expensive ones. This is a very capital-intensive industry. You can see here just in the last few years, there has been $800 billion. There have been $800 billion invested in fabs. And by the way, some of them are what's called 200-millimeter fabs, the older generation fabs, we still need those as well. It's not just the most advanced, the 3-nanometer, et cetera. Even 130-nanometer, even 90 or 65 are being used. And I mean, the point that I was making, of course, at Semiconductor Australia is that this place, you look at all of these -- these are the advanced countries, the U.S. and Europe, and we have Israel and Singapore and China and Japan, Korea, all those -- Singapore over here. And this place is just a void, which kind of looks very strange, especially because Australia is an ideal place for a fab. It has the ideal conditions for a fab. So there's a lot of stuff going on here. There's a lot of investment happening. It is a very exciting domain. For me, it's always a thrill of living it. You guys know our Board, and we have Ash online here. I'm going to skip these slides. A quick reminder of what's happening. Flash has been around for decades, more than 10 years ago, more than 15 years ago, people started looking for a replacement for flash. Actually, I should say even 20, 25 years, people have already been working on it. Just like you realize, again, the time scales that we're talking about. You look at companies like Adesto, which was developing ReRAM or Crossbar or 4DS for that matter. These are companies that were started in the 2005 to 2010 time frame and already working on these technologies then. And other technologies I have here, MRAM, PCM, et cetera. People have been working on this for a long time. These are the time scales of semiconductor. And we need to understand these are extreme, extremely complex technologies to develop. And by the way, Weebit, while we started the company in 2015, we didn't start from scratch. And actually, what we're doing today is based on technology that Leti has been researching ReRAM for more than 25 years now, and we are leveraging their knowledge and their know-how and so on. So there's really a lot of work going on. You look at these technologies, I mean, you have here flash as the first column, and you can see how many red dots there are there. That is really what's triggering this industry to move forward. I haven't talked a lot about PCM in the past, but PCM for a long time, people were also believing that this would work. Actually, Intel invested [ haven't knows ] how many billions or tens of billions of dollars into PCM, and they had it manufacturing. I mean, it was in mass production. They were selling it. Obviously, what happened was every chip, they had to give the customer wrapped in a few hundred dollar bills, so that they would take it, and eventually, they realized they're going to go completely bankrupt if they continue and they stopped it. So it was Intel, it was Micron and others. There's only 1 company today in the world that is still pushing PCM that's STMicro. It's for a limited scope of applications. It's not good for everything, but they are still trying to push it. MRAM, this is really the technology that everyone was counting on. Everyone was sure that this was going to take off. This is it. It's been more than 5 years now in mass production, but we talked about it many times, that magnetic element. So it's very expensive. It is much more expensive than ReRAM to manufacture. It has rare earth materials, which many people care about. But more than anything, products that were manufactured with MRAM started failing out there in the field because we're surrounded with so many magnetic fields, the memory got erased. So -- and I'll be talking about what's going on now and let me kind of quickly -- I'll get to some of the impacts of that. But really, you look back at the last 12 months, and I think there's been quite a bit of progress despite the fact that it seems like there weren't any announcements. So we went from -- a year ago, we still had only SkyWater. Today, we have also DB HiTek. I think the key is really we had engagements a year ago. We had several -- quite a few things that they were starting, and we were pushing and so on. Today, we're in a much more advanced state with quite a few foundries and IDMs and product companies. There's really a lot of activity going on. The partnering with Efabless, I really hope will enable us to get some companies to manufacture at SkyWater. We -- a comment about SkyWater because I know everyone always asks about it. We believe that SkyWater was going to be a very good start. I mean, I still don't regret having worked with SkyWater. You don't want to work with one of the very big guys when you're doing your very first transfer, having forbid something fails, and you're burned in, in one of the big guys. So I prefer to work with someone that we could look more in the eye. And we are -- as everyone knows, we are using SkyWater, as a part of our R&D. You could say that the R part is at Leti, the D part is at SkyWater. They do things faster and cheaper when you get to more production. The thing about SkyWater is they're getting so much money from the U.S. government that they're just happy. And they kind of don't have that motivation of bringing on board customers because their fab is full with U.S. government stuff. And that's why even when we bring them big name companies, big names that I think everyone in the room would recognize, they are reluctant to really make that effort to really make that push, and that's frustrating as hell for me, but that's the way it is. I hope that with Efabless, we will manage to -- Efabless is more for building prototypes as an initial phase, but still after that prototype, if things go well, they do go into mass production, at least some of them will. So we do hope that, that relationship will help us. We -- for a long time, well, since COVID, we basically kind of put China on hold and for a while, we didn't talk to China. Today, we are engaged again with companies in China very carefully, very selectively. It's not just anyone. And beyond the obvious, we cannot talk to SMIC, SMIC. We cannot talk to Huawei. But we are cautious in the way we're doing it, but there is activity going on. It's a huge market. We can't really ignore it, and now we're at the point, where there is some activity there. A year ago, of course, we were pre-revenue. We now show very initial, very -- not anything significant, but still $1 million revenue is the first step moving forward, reflecting the license fees that we are getting already from these first 2 deals. And in terms of the technology itself, there's a lot of work going on. We have 13 PhDs in chemistry and physics constantly working and constantly improving the technology. And you can see it the fact that we're moving forward, we're qualifying. We did the qualifications at 125. We demonstrated now the 150 degrees, which not only does 150 degrees open the door wide to the automotive market, which is a huge market. Again, every new car today has at least 1,000, if not 1,500 electronic components in it. It's crazy, okay? This is a very big fast-growing market and especially all the autonomous vehicle and stuff, that's even way more. But being able to work at 160 degrees, even for customers, who are just looking for something which is of room temperature or not very hot, it just shows them robustness. It shows them that this technology really is it, it's serious. It can withstand extreme conditions. And so, this has been very important for us. And we have been engaged with even very small geometry foundries. We did get PDKs. We have several PDKs, which are at the 1x nanometer and have done simulations and have demonstrated that we know how to work there. So again, that's part of this engagement with so many of these big foundries and big IDMs is they are sending us their PDKs. And we are showing -- we have, I think, 3 PDKs that are in the FinFET level to 1x nanometer. And we have been doing simulations and showing them that it works. So really, while it seems very quiet, and I know there is a lot happening, and I think it's important to point that out and show it. So with DB HiTek, everyone knows, we did the tape-out. We're waiting for the wafers, hoping that everything goes well, and we do the testing and the testing passes well and then we can go into the qualification. So there's a lot of work going on with them. They are very cooperative. There -- it's nothing like SkyWater. I mean, a completely different world, and it's really good. By the way, they -- we are pushing customers to DB HiTek. They -- I don't know how many of you realize they were burned. They did try to use the ReRAM from Adesto in the past, and they were burned there. And so, they're really cautious moving forward and things. But nevertheless, we are pushing customers there, and we are engaging, and I think that when we get the wafers back and the testing starts showing positive results, it will be much easier to engage with their customers and push things forward. You all know we have the 22-nanometer demos. Those are very successful. They really make a big impact in the market. People see them, these demos, they're impressed by them. It shows, again, the power. GF and many people are asking me about it. GF is -- we talked about it in the past. They're trying to work in parallel with us. They're trying to develop their own ReRAM, and the relationship is definitely not simple with them. But -- these wafers, we're using them. We're doing a lot of testing. We're doing a lot of analysis. So we're learning a lot from these wafers all the time, and the R&D team is still constantly manufacturing more and more wafers to do more and more tests with this. So there's really a lot of good stuff happening with these wafers. I think we talked a lot about the business model, so I don't really think that we need to go into it. We are -- today, the big focus is really on foundries and IDMs. It should say slash IDM up there. Well, that will make this slide much more difficult. But we are trying to engage with foundries and IDMs because at the end of the day, the product companies won't really engage until they don't know that their foundry has our technology. Now it's true that the foundries need to know that there is a customer that there is customer demand, but I think today, they know it. So I'm going to skip these -- some of the slides here are more for new people and new investors. Looking forward, I mean, semiconductor market is huge. We've been talking about that a lot. There are many different segments, where our ReRAM fits. Weebit today is engaged with medical, with automotive, with AI, with power management. Actually, if you ask me what is the #1 market that today sees real value in ReRAM and is willing to make that extra step and move forward, the power management guys. They today really need ReRAM. And so, everyone needs it, but the level of urgency with power management is stronger. I mean, what's happening, first of all, so many devices today are battery operated. I mean, everything needs to be charged today, right? And when you're charging, it's no longer just a simple cord going into the wall. You always have these boxes. I don't have it here, but I mean, you know, those boxes on the cables or the thing that you plug into the wall is actually a pretty big thing. It's not just a simple connector that goes into the wall. They -- the reason is that power management today doesn't just power. You wanted to preserve the battery. You wanted to lengthen the lifetime of the battery. So -- and I think many of you know, you want to start charging slowly, then ramp up and charge strongly and then slow down when the battery is almost full because otherwise, you will ruin it and things like that. So you need beyond the power management logic, which is the traditional one that everyone always had, now you need a microcontroller, and that microcontroller cannot work without a non-volatile memory. Now you can have them, and that's what people are doing today. You can have 2 separate chips. You can have the power management and then the microcontroller on a separate chip, but that's not optimal. If you want to combine them, what happens is the power management is an analog circuitry and the microcontroller with the non-volatile it's a flash. The flash is a front end of line technology. It [ sits ] next to the design, and it actually interferes. It causes a lot of challenges. The fact that ReRAM is back end of line and comes on top, that's a huge benefit for these guys. And that's why we have -- the engagements with the power management guys are even more intense and more advanced than with others. And we talk a lot about BCD. That's definitely for power management, and we have a very nice blog that came out, which I think can really help you guys understand some of what is this thing. So that's these -- the different markets that we're talking about, the growth opportunity for ReRAM, you guys know the slide and the numbers are less important. I think it's really the fact that the market wants to move to ReRAM. And maybe this slide is potentially the most important one in the presentation. What you can see here is the really important transition that has happened in this market. And we're not talking about a long time. Just 2 years ago, basically, the market was just starting -- we started seeing ReRAM reach the manufacturing stage, and there was that big announcement. The announcement of Infineon, which is one of the biggest automotive companies. The fact that they announced that all their future MCUs will have ReRAM in it, that was a shock wave in the industry. It really was a shock wave. Many people suddenly woke up and said, "Oh my God, this thing is real. It's here. It's no longer the future. Weebit changed its tagline at the same time more or less because the next NVM is here suddenly, it's not that future technology. Weebit at the same time, was working with SkyWater to qualify. A little bit later, TSMC started pushing in power management. Again, I mentioned power management earlier. Those were the earlier things that they engaged in. We also, last year, engaged with DB HiTek and there was a lot of activity there. And on the IoT side as well, TSMC mainly. UMC, we're really not seeing them. It seems like their ReRAM has a lot of issues. So we're not really seeing any customers picking up. They did qualify it. But even I think you guys know the UMC technology originated from Panasonic, and Panasonic sold its semiconductor division to Nuvoton. And half a year ago, Nuvoton announced that they have this new MCU with ReRAM in it. They did a big [indiscernible] and everything. And for a while, I was totally confused until we finally realized they manufactured at TSMC. So even Nuvoton didn't use their own ReRAM at UMC. They went to TSMC. And I don't know, we still need to understand what's going on there, but TSMC is pushing it forward. Moving to the end of '23 and TSMC, actually, this is really end of '23 and moving into 2024. TSMC once a year kind of in the middle of the year, they announced their road map. And their road map is really -- I mean, a year ago, in June of '23, it was all MRAM. You looked at the slide, it was all MRAM, ReRAM was mentioned in the corner. They still had a total focus on MRAM. Guess what, June this year, and I have the slide -- a copy of the slide here, I'll be able to show you guys offline, their road map slide totally the opposite. Everything is ReRAM with just a small mention of MRAM. And the reason is the fact -- and we don't -- I guess, I don't see the comment here, but the fact is that consumer companies, the big consumer companies have been vocal about not wanting to use MRAM, okay? Normally, by the way, the big consumer companies, they like to throw responsibility on the foundry. They basically say, we will use what the foundry recommends. And if it fails, the foundry is to blame, and we will go strongly against them and after them and all of that. So the big consumer companies normally will not voice their preference because they want to be able to go after the foundry if something fails. In this case, they actually have become very vocal. They said, we know MRAM will fail in the field. We already saw it happen. We're not willing to use MRAM period. And the result is TSMC and others really understood the message and now everything is focused. So I think now we can really see moving forward, there is a clear understanding in the industry that ReRAM is the technology. ReRAM [ is it ]. They want to use it. I think that's one of the things that's also given a push to these foundries because the foundries kept telling us, yes, we know ReRAM is great. We can see the results. We're demonstrating all these great things, but we need to know that if we invest in technology transfer, we spend these 9 months and all of this manpower and all of this work and wafers and everything to bring up ReRAM, we want to know that there's customers behind it that are going to pay. Today, they can already see it. I mean the market is -- it's going through that transformation, and that's what's enabling us to make more progress to move forward with these guys. So it is very difficult. The foundries are very cautious. They've spent billions on their facilities. They don't want to take any risk. We are going through unbelievable due diligence with them. The product companies -- think about it, a product company has a new product that they expect to have hundreds of millions of dollars or billions of dollars of revenue over the lifetime. And they see all of the advantages of ReRAM. And everything is clear, but still the one thing that they haven't seen, and this is repeated over and over and over again, they haven't seen what the industry calls silicon proven. Now silicon proven is not just 1 or 2 demo chips or showing them a few hundreds of chips that are working. Silicon proven means mass production, millions of chips going through and demonstrating that you can actually manufacture them in mass, and it works. Now one more thing that I think is important for you to understand. Semiconductors is not like a lot of industries. If I manufacture chairs, basically, all the chairs are basically identical, right? I mean -- and if there are small variances, who cares. But in semiconductors, the manufacturing, actually, every ReRAM cell is different than others, okay? There are no 2 completely identical ReRAM cells, okay? The manufacturing process is so delicate. We're talking about nanometers. We're talking about practically putting almost every molecule in its place. The most minute things impact here. And so, people want to see that when you manufacture millions of these, the variance is very small. They want -- I mean, they know that they won't be identical, but they want to know that the variance is going to be very small, that it's really ready for mass production. And even though we did go through the JEDEC qualification, and we did all of these things. But it's these kinds of things that sometimes we have a product company, who says, yes, we know it's going to make our product cheaper, lower power, more competitive, all of these great things. But there is that risk. It's not silicon proven. And we're going to try to make our current product still work and be competitive and whatever with flash. We need to see someone else do it first. Yes. And that's really -- it's hard to explain, but that's -- it's ongoing discussions and convincing -- and convincing now. At a certain point, one of these guys realize that, hey, the potential here is so huge, I'm going to take that risk. And I've been through this already with previous companies. And I mean, it's not the first time that I'm going through this with semiconductor and these kinds of products. But we will get there, by the way. Again, I'm expecting to have a deal before the end of this calendar year. I really, really, I am working very hard on it, and we don't have a lot of time. So I am taking a big risk already making that commitment, but that shows you that we really are making very good progress there, and we will do that. And then this one company, others will start being more -- I mean, DB HiTek already helped us get off these other guys, and more engaged another one and another one at a certain point, it will get to the FOMO phase, where suddenly everyone will say, "Oh my God, we don't want to be left behind. And that's when you go into what in the industry we call the tornado, where suddenly it all goes crazy. So this is really what we're seeing and look at it in just 2 years, we've gone from ReRAM is that future technology to having ReRAM become more and more standard. You now have -- I mean, TSMC ReRAM is silicon proven already. And people -- again, it gives people confidence even though it's TSMC, they see that ReRAM can be mass produced. Now it's not Weebit. So that's still that last thing that they're concerned about. We need to overcome it. But at the end of the day, these foundries that are not TSMC and UMC, they need ReRAM, and Weebit is the only supplier. So like it or not, they're working with us. And they know -- I mean, they're not working us, and I really believe they're not working us just because we're the only one. They're working with us because we are demonstrating quality. They can see a quality team, a very strong team with a lot of experience with a ReRAM that's already proven. We've been demonstrated at 150 degrees, and we've been qualified at 125. We'll also qualify at 150. We'll be pushing that forward in parallel to so many other activities that we're doing. But they're seeing this. And again, it's -- you don't see it, you guys look at Weebit and maybe think oh, this company went to sleep or something. But we are really working very hard, I can assure you. And to prove it, well, first of all, a few months ago, I think you saw Dadi and Atiq and myself, I cannot recommend to anyone to buy the share or not buy the share, but I personally think it was a great investment on my part in the summer to buy some shares. Right now, you saw the notice of meeting come out. And I think there's a very important message there. And again, I believe it shows the quality of this Board because the directors basically said, we need to listen to the shareholders. We need to -- we're not living in this vacuum or in this bubble or whatever. The shareholders are not happy with how things have gone up to now that we don't have a deal yet and things like that. So there will not be an allocation of any equity to the directors this year, and it is a very extreme step. I mean, I don't remember this kind of step being taken in other companies in the past. But it is a message that the directors are listening. And my equity grant, which is proposed for the AGM is totally tied to performance. You can see that there's -- I am committing to get more foundries and fabs, I mean, foundries, IDMs on board and get product companies on board. So I do believe we will achieve those milestones, and I'm committing to it now in a public way that we will do it. So I think you're getting these messages from us, and we are doing everything we can. It's not all in our hands. We need to negotiate this. And some of these companies, I mean, it drives me crazy. They send us comments on the contract. We react within overnight. They get the answer the next morning and then their legal is, oh, we have so much time. And a week or 2 later, they send us some comments and things. So it just takes forever with these big, huge dinosaurs or huge companies, but it's moving forward, and we're engaged with more than 1. So we are pushing it forward. I want to talk about something that is also very important because, again, people ask me about what is ReRAM good for? And I keep saying it's good for all of these different markets. But there is a market today that you cannot ignore, and everyone talks about, and I want to talk about it, and that's, of course, AI. So one of the things that is happening in AI, now a few words about AI because I'm assuming a lot of you don't really know how it works or what happens there. AI has 2 main functions. There is learning and there is inference. So you first need to teach the AI system how to do whatever you want it to do. And then there is using that knowledge in the day-to-day life to actually do things, and that's the inference. So I'll take an example, which I think will be maybe the easiest one. I want the AI system to recognize a terrorist when he's walking down the street, okay? So the beginning is to show the AI system a lot of pictures of this terrorist from different angles and profile and whatever and teach it how -- what this terrorist looks like, okay? Now that needs to run on these big, huge data centers. It requires a lot of compute power. It requires a lot of memory, and it's a big thing. At a certain point, the system knows how to recognize. And now you want to move to the inference stage. And you have all of these cameras along the street or in the airport or wherever, and they're looking at people. Now -- it used to be people thought, okay, we still need a lot of compute power, and we still want to do -- we'll just -- everything we film will stream up to the data center, and we'll have the data center do all the inference calculations and stuff. Well, guess what? There is so much to stream up there. The bandwidth is just unbelievable. And these cameras very often are working on batteries or whatever power is an issue. And that's when -- sorry. And that's where the understanding that inference needs to be done at the edge in that camera in the airport or wherever came in. Now -- how do we do that? Well, we have this chip, which has an AI engine, and it works. It runs whatever algorithm and so on. But we need to load it with all of the, what you call, coefficient that are used in order to do the AI computation, okay? That's what we got from all of the learning, right? This needs to be in a non-volatile memory. It needs to be somewhere where when we turn the power -- the system on, it's there, right? Now what are we going to do? Well, normally, these chips are 28, 22-nanometer and going to smaller. We can't really have embedded flash in them. So we have flash outside, as a separate chip, where we put the coefficient. We turn the system on. We load the flash into a dedicated SRAM here, and we'll use it. Now SRAM is volatile. So the system now you can't really turn it off. You turn it off, you lose everything and so on. SRAM bits, by the way, are very big. SRAM bits are normally 6 or 8 transistors. Flash and ReRAM are basically 1 transistor. So you can -- the number of coefficients that you can load is smaller. And that's where people have realized, hey, we can actually replace this SRAM with ReRAM. ReRAM can be done at 22 nanometers or even below. Let's go and put ReRAM here. We already have a huge saving because we're throwing away 1 out of 2 chips. So that's already a big saving. And then we can now turn the system off. I mean, if nobody is walking down the street, no sense in having the system on. We can just power off and when someone starts walking down the street, we turn it on again and so. And since ReRAM sales are so much smaller, we can actually have more coefficient. We can be more accurate. We can be better. So this is a trend that we are definitely seeing, and there's a lot of interest in it and research institutes are already engaged with us on it. And this leads to what you know already, eventually, and we are already engaged with research institutes on the neuromorphic side on what's called -- I mean, it has so many names now in memory compute, neuromorphic and others. It's basically -- as far as you're concerned, it's the same thing. So -- and we're heading towards that. So Weebit is focused on this. We are working with more and more research institutes all over the world on this thing. And I think that's, again, just to see how ReRAM is really relevant for all of the markets, including some of the more advanced and interesting ones. Okay. So I'm basically towards the end. You know we have a very strong team. We have the device and process and analog and digital teams and the whole cooperations have already presented the slide in a different form, but you know this, and that's really the secret of success of Weebit. And moving forward, well, basically, we threw here on the slide names of big foundries and IDMs, and I guess, end product companies, some of them, mostly -- basically fabs and IDMs. I'm sorry, yes, all of them are fabs and IDMs. So you can see here the names. We're engaged with really the vast majority of these guys. There's so much going on with them. And the goal is to start announcing some agreements with them. So we are working on getting something done before the end of this year and then moving into '25, getting some more of those. You saw it in my targets for between now and the end of '25, get 3 licensing agreements, get 3 product agreements and get the DB HiTek qualified. I think those are really the most important things. That's really where we are. I know everyone wants to hear more news. We're working hard on it.

Okay. Thanks, Coby. My name is Danny Younis. I’m from Automic Markets [ and IR ] for Weebit. So for those in the room, who want to ask a question, just raise your hand and I'll bring a microphone across. For those online, if you want to type in a question, you type in at the bottom of our Zoom box. We already do have several questions online. I might go to online first, while those in the room are actually thinking of a question. Okay. Coby, the first question is, is there a chance of getting a deal with an end customer through SkyWater?

Yes. That's a good question. Thank you. It is -- I haven't given up. I mean, I told you I'm totally frustrated with them, and I am totally frustrated with them, and they know I'm totally frustrated with them. But we are not giving up. And just a few months ago, as I mentioned, we had a good opportunity. We brought it to them. They didn't want to work with it and whatever we're taking that customer now to DB HiTek and hoping to get it done there or with another foundry that might pick it up. So we -- I'm not giving up. At this point, that's the only place, where I am qualified today, a mass production facility. So I don't have a choice. I'll keep pushing. Hopefully, we'll make it happen. And by the way, that's what Efabless is about, and we really hope that maybe even through Efabless, we'll manage to do something. So we haven't given up on SkyWater. And I still hope -- I'm committing to 3 product companies next year. I really want to see at least one of them at SkyWater.

Okay. The next question is around the fabs. So what are the ideal conditions for a fab?

Well, fabs in order to build a fab, I guess, someone picked up on my comments that Australia is ideal for fabs. You want it to be in a place, where there are no earthquakes. Japan, by the way, has a lot of fabs, even though they have earthquakes, and they have a big issue on how to protect them and so on. Even Israel is a place, where you do have earthquakes. But a place, where the ground is solid, where you have clean air, where you can go outside and be further away from the pollution of a city and have that and where you have clean water. I think those are the 3 top requirements that you need. Australia has them. So yes.

Okay. The next question is around DB HiTek. So regarding the DB HiTek qualification, are you awaiting silicon? Why is there a delay in obtaining silicon?

There isn't a delay. It's just that's how much time it takes to manufacture. I think I presented in the AGM last year, the slide with just how much time things take. Manufacturing normally is about 6 months. I mean, you can try to rush it and do things, which cost a lot of money to have it faster. And we are trying to get DB HiTek to give it higher priority. But in general, 4 months to 6 months is the normal time to get silicon back.

Okay. The next question is around how should we ready ourselves given the fact that TSMC has serviced the bulk of the power management business, and we are left to...

The message is clear. So well, first of all, TSMC has definitely not taken the bulk of the PMIC business. I would say even on the contrary, TSMC has ReRAM at 40-nanometer and at 28-nanometer. They're working on the smaller geometries. PMIC normally wants 65, 90, 130. So to a certain extent, yes, they are -- I mean, they have 40-nanometer BCD. By the way, again, they started off. The first one that they did was 40-nanometer BCD for power management, but many of the power management use larger geometries. That's number one. Number 2, many of the companies like to have a second source. People don't want to be completely reliant on TSMC, and they really want to find the second source, and we can definitely be that second source. Number 3, as you all know, I really don't consider TSMC as my competitor. And I honestly believe that when we will continue to evolve and show that our ReRAM is strong, and we'll have customers, who come and want some tailoring of their modules and things like that. At a certain point, I really hope to have a large volume customer who will go to TSMC and say, "I want Weebit ReRAM, I'm going to buy a lot of wafers, please, let's work on it. And TSMC, today, if you look, by the way, at TSMC's enablement library at what I always call the supermarket, you have their IPs that TSMC developed and competing modules of the same type of IP. It's not like unheard of that TSMC developed an IP and then put in their library IPs for the same thing from other companies. So we can definitely go even after those guys, but yes.

Okay. I might go to the room. Does anyone in the room have a question? Please raise your hand.

Hi, Coby. Could you give us a description of the typical evaluation process in terms of -- is there a standard road map? Do you go from simulations to physical engineering? And in that, do you have the capacity to service the needs of the foundries you're talking to at the moment?

So with the foundries, they basically -- first of all, they just want us to demonstrate a lot of parameters. It's voltage, current power, leakage and so on and so forth. So we actually go and demonstrate to them using what we have, the chips that we have today, we demonstrate, we show them, we show them the results of the qualification. Then they will come back and say, you know what, we want you to give us this die size. We want to be able to put this many -- whatever kilobits, megabits into this square millimeter type thing at a certain voltage. They will challenge us. And so, our team will go and will -- I mean, they send us their PDK. That's the beginning. I mean, the first step is really getting their PDK and analyzing it and understanding how their technology works, okay? So the PDK is process design kit. And this is like the basic element that you need. It describes their process. It describes how their transistors work and how everything works, and then we can take that, and we can put it into our simulations, and we start simulating. And we simulate and we tell them, okay, now they will come with extreme demand. I mean, sometimes they will come and say, we want you to operate at this voltage and we'll be there, oh my God, this is super ultra-low voltage, how are we going to do it? And you have the teams working together. Again, that that benefit that we have that the process guys are working with the analog guys and everyone, and they look for how do we get a joint solution because the process guys alone can't get there, or the analog guys can't get there. But together, we can manage to achieve it, and we go, and we demonstrate it. And then they will try to challenge us in many different ways and want to understand. Now sometimes it's the manufacturing process. Even though we're using a standard machine, but they use the machine in a different way than what we do. And so, they want to understand it. And then they want to see maybe we can change something in the way that it will be easier for them or things like that. Sometimes it's, oh, one of the machines that you want, it's fully booked. We already have it at capacity. We're going to need to buy a new one. So we will have some expense. We need to see how we deal with it. And there's endless combinations to this thing and a lot of simulations. And many times, the way this industry works, suddenly, they will tell us, "Oh, we're sorry, we have whatever crisis on another line, or we need to bring up another line. We need to take these people over there. We're just pausing right now. We'll be back. And they continue communication with us, but kind of the whole evaluation is put on hold for a while. It's very common that these kinds of things happen. You guys might remember, I think even a year ago, I was feeling pretty good about the potential of closing a deal even in '23. I know you guys have good memories and people are saying, wait a minute, he was already here, and he talked about closing a deal. We weren't nearly as advanced as we are now. But I had a good feeling about that customer. And then suddenly, they -- we need to set up a new production line. 10 months to 12 months, that's how much it takes. So they disappear, right? It's those kinds of things that happen. But that's more or less how it goes. It's a lot of simulations and a lot of discussions and a lot of demonstrating our capability here.

Coby, 2-part question. The first part is great to get the initial revenues this year. Can you give us a bit of a feel for the existing agreements, what they may look like as 2025, 2026 roll through? That's number one. And then the second part is for the agreement -- for the current negotiations you're working through, are we looking at single geometry negotiations? Or are we looking at something bigger, older, better?

So the first one to -- for those who don't know how revenue recognition works. So revenue recognition is not tied to the payments. I mean, it's been already a long time since the tax authorities realized that salespeople are really creative and play with the payments and things. And so, what they basically try to do is take the expected revenue for this project and recognize it over the lifetime of that project. So they kind of look at the expected lifetime, they look at what's the expected revenue. And what you saw right now is the part that was recognized from DB HiTek and to a certain extent, from SkyWater that they were willing to recognize. So now whoever wants to start doing extrapolations of what part of the DB HiTek do you think we finished, knowing our time scales, et cetera, and try to extrapolate how much revenue was expected, it's already up to you. So that's kind of on the revenue side. Payments are, of course, tied to reaching milestones and things like that. I won't go into what the expectations are from the future deals. Again, I just can't go into those things, but I'll stop here. The second question, I forgot.

Second question is just around the geometries, are we talking to just around 130 or 65 or are we talking...

We -- several of these guys and especially when you talk to IDMs and so on, once they want to get into a technology, they want to know that they can have it over multiple geometries. And here comes, again, that fear of do we want to sign an agreement, which is on one geometry and then expand it to other geometries? Or do we want to try to have it as a more like what we call an architecture license, where they really take the technology and have more freedom to transfer it to other geometries. Each one of these guys has a different philosophy. Some of them are asking for more than just one geometry. We'll see. I mean, I don't know. These negotiations are going back and forth and sometimes they change in the middle. And so, I don't know. The answer is when we'll close deals, we might disclose the structure.

Are there any other questions in the room?

Thanks for your insights, Coby. Just a question about capital. Is there any likelihood that the company is going to need capital to continue this development? Or how well are you placed?

So we will be issuing our quarterly report, I guess, tomorrow or the day after. But the last report that has been made public is the Annual Report. It stated that -- I mean, in that report in the end of June, we had $63 million in the bank. And fiscal year '24, we basically spent $25 million. So simple math, you can see that there's more than 2 years that we have. Obviously, we are still ramping up. And obviously, you don't want to get to the point, where you don't have money. But at this point, we have 2 years. So money is not the issue at this point, and we're not thinking about it. I am totally focused on closing these deals. I'm literally not thinking at all about any raise or anything like that.

Okay. Then we're coming up to close. So we're still receiving a lot of questions. We're running out of time. So if anyone has any further questions, you can contact us and Coby or us will respond to them. But maybe one final one. I think this one has been asked about 4 or 5 times online. So I think it's probably the best one to finish up on. And the question is, Coby, you've been saying now you would close the deal/agreement either with a fab or a customer in 2024. Has that changed? It seems as if now that goalpost has changed in 2025. Is that correct?

I think I said it several times already that my goal is before the end of this calendar year. So it's a big challenge. It's -- we just need them to move faster. We're -- again, there are negotiations going on with more than one, but at least one is already close enough to the finish line that I really hope their lawyers move a little bit faster and we just closed this already.

Okay. We've probably got time for another one. Has we get scaled into the teens with another company that is currently under NDA? Or will all scaling only take place with Leti?

Well, I think the only comment I can make is, I think I mentioned we have 3 PDKs that are in the teens. I cannot talk more than that.

Okay. Then on that note, we'll wrap it up here. Just to reiterate, if there are any further questions that you require an answer, just please contact us and will respond to those questions, me or Ashley.

Okay. Thanks. Thanks, Ash for joining us.

Thanks, everyone. Thanks, Coby.

And thanks, Eric and Danny for organizing this.

Great to meet you.

And of course, thanks to everyone for coming.