Weebit Nano Limited (WBT) Earnings Call Transcript & Summary

Hello, everyone. Welcome to another sharewise webinar. Today, I'm joined by the CEO of Weebit Nano, Coby Hanoch. Thank you so much, Coby, for joining me today.

Thanks, Caitlin.

So today, we'll be doing an interview style. I'll ask a few questions of my own, and for everyone that's watching, feel free to "put in all your questions in the Q&A chat box down below. So I'll get us started. For those of us who have joined us today, Coby, that aren't too familiar with Weebit, could you just give us a brief overview of your technology and what you guys do?

Yes. So semiconductors today are everywhere. Anything you touch almost has some sort of semiconductor in it. And in this domain, memories are the largest segments. So if semiconductors are expected to reach over USD 1 trillion in -- by the end of the decade, memory is a very big chunk of it, everything needs memory. And in that domain, we are focused on what's called a nonvolatile memory. So those are the memories that don't lose the data when you unplug them from power. If you imagine any electronic device when you turn it on the first time, there needs to be something on it, some software on it that runs some basic data, et cetera. So that has to be on nonvolatile memory. And there are many other things that need to be on it as well. So that's the technology we're dealing with. This market segment on its own, just nonvolatile memories, is expected to reach over USD 120 billion in the very near future. So it's a big segment, and we're very excited to be in it.

Thank you for that introduction. And so how large is this market really? And what is the competitive environment?

So as I said, this is a very large market. The existing technology is called Flash. This technology has been around for many -- for decades actually. And it's been hitting the wall in many ways. So it's slow, it's power hungry, et cetera. We're developing a technology that's called resistive RAM or ReRAM for short. It is much more advanced than Flash on so many fronts. It's 100x faster. It's 100x lower power consumption. It's much simpler and easier to manufacture. That's a key point. We don't require so many manufacturing steps like Flash. And the bottom line is it's just very environmentally friendly, and it is much cheaper and simpler to manufacturer. So if you look at Flash, it adds about 10%, 20% to the cost of a silicon wafer that will be kind of measured against. So other new memory technologies can go up to 30%, 40% and even more of a wafer cost. We are very simple and easy. We go into a manufacturing facility without requiring any changes really going straight in. So the added cost is maybe 5% or 7%. And that's really a key advantage, of course.

And speaking of the ReRAM technology that you were discussing, I know that we've recently qualified your embedded ReRAM at SkyWater's American production facility. Why is this such an important milestone in the commercialization of the company?

So first of all, let me explain a little bit about how this market works. You manufacture semiconductors in facilities that are called fabs. Now fabs are very, very expensive to build and set up, even the simplest ones cost billions of dollars to set up. The more advanced ones can cause tens of billions of dollars to set up. It's really crazy. So obviously, most companies in the world can't afford to have their own fabs. And I mean, I'm talking even about Apple, Google, Facebook, Qualcomm and many others that don't have their own fabs. So there are companies called foundries which have fabs and offer the manufacturing services to the global semiconductor market. SkyWater is one of those foundries. And today, we announced and we'll probably talk about it another agreement with DB HiTek, which is another foundry. So we basically transfer our technology to these foundries so that they can manufacture it. We actually have a licensing -- a manufacturing licensing agreement with SkyWater and we transferred the technology to them. And we -- after developing prototypes and so on, we went through what's called qualification. Qualification is a process defined by a big standards body that define very, very rigorous testing that you need to go through in order to be able to say that you're ready for mass production. The fact that we went through that whole process, and now we have actual chips in silicon that have passed that, that were manufactured at SkyWater, that's a very important milestone. And what basically enables SkyWater customers to engage with us, and now we're working with such customers in different stages of evaluation and discussion in order to kick off manufacturing at SkyWater, obviously, we license the technology through them and then manufacture at SkyWater.

And you spoke of your latest agreement with DB HiTek, which I know a lot of viewers have submitted questions in regards to that for today's webinar. So I think we can get on to that. So first of all, can you tell us a little bit more about this agreement with DB HiTek? It's a South Korean company and one of the world's largest foundries. What is the scope of this agreement?

So DB HiTek is already a much larger and better known foundry than SkyWater, they have hundreds of customers, including many of the big names. It's the Intel, Qualcomm, Sony, Texas Instruments, Toshiba and many big companies are customers of DB HiTek. So this agreement basically defines that we will transfer the technology to them so that they can manufacture it in products for all of these hundreds of customers that they have. And that's very exciting. I have to comment, I know many people are asking now, why did you make this announcement and you didn't say that it's price-sensitive. And I have to just make the comment here. We clearly think it's price-sensitive. We clearly think this is a major, major milestone for Weebit with a potential revenue, which is very, very big from all of these customers. For some reason, the ASX without consulting with us, decided to drop that and just release it as a regular announcement. But I want to be clear, this is a very important milestone for Weebit.

And we have another question here submitted asking that you have had an agreement with SkyWater for a few years, but it hasn't generated any significant revenue. So why will the agreement with DB HiTek be any different?

It's a good question, and I want to clarify a bit about how things in the semiconductor world work. When we sign an agreement such as this one with DB HiTek, we now enter a phase of transferring the technology to them. And again, manufacturing semiconductors is an extremely delicate, very, very complex task. So transferring this technology to a fab will take several months. Once we go and manufacture the test chips, and that takes several months. And then we need to go through the testing and qualification and so on. And again, that takes time. I explained in different interviews, part of the testing is putting the chips in ovens for thousands of hours at high temperature to see how they -- that they resist that and so on, thousands of hours means months of just waiting until you actually get the chips back. So it did take us about 2 years with SkyWater to go through this whole process and have our chips or our technology qualified. And now we are engaged with their customers. It will take us, hopefully, somewhat less than 2 years to get it qualified at Dongbu, but that is the order of magnitude that it takes. Now it's important to understand, the first customer you ever work with seems everyone is always extremely suspicious, everyone is extremely cautious. The fate of their product is on the line when they use a new technology. And so obviously, SkyWater customers have been extremely hesitant to move forward and be the first ones who actually use the technology. Now that they see us qualified, they see things moving forward, I think now that they'll see the announcement from Dongbu from DB HiTek. They -- all of these things make it easier. And as we move forward, each of these agreements will make everything simpler, we'll be able to move faster, we'll be able to do things -- I mean the time scales that we have to go through, I mean you can't shorten it. I told people that you can't have a baby with 9 women in 1 month, and you can't -- you can shorten this, you have to put things in the oven for 1,000 hours or 5,000 hours, you can't shorten it by putting it in multiple ovens for a shorter period of time, you just have to wait. So this is basically what's going on. But we are moving forward now, we told the market that we will even start seeing some initial revenue before the end of this calendar year. Again, it's not going to be huge revenues, but there will start being some initial revenues from these things and it's going to start moving forward, and we will have, obviously, more agreements moving forward.

And we also have a few questions asking about GlobalFoundries. Someone wants to ask, following tape out, when will you get wafers back from GlobalFoundries to start testing and qualification?

So we are expecting these wafers in the near future. We're working with GlobalFoundries. GlobalFoundries as I think many of the listeners know, they also acquired another ReRAM technology whatever remains from the Adesto what's called CBRAM technology which is the version of ReRAM. And so they're working with us in parallel and we are now waiting for these wafers in the near future. We will be doing testing and working with GlobalFoundries. It's a good example to show that these foundries, even when they already have a ReRAM technology, it doesn't preclude them from looking at other ReRAM technologies and working with them. I'm mentioning this because I know that people are asking some of the bigger foundries like TSMC or UMC announced that they have ReRAM and it's important to understand that for these foundries, ReRAM is a means to an end, it enables their customers to manufacture more wafers, more silicon and buy more wafers and the fact that they have a ReRAM doesn't mean that they won't look at other ReRAM. And if customers come and say that they prefer Weebit ReRAM, as long as they buy a lot of wafers, the foundries will be glad to work with Weebit as well.

And moving on to talking more about the discrete memory market. I know that Weebit's current technology is for the embedded memory market. Can you tell us about the solution you're developing for the discrete memory market, which is I know part of your midterm strategy?

Yes. So that's a very good point, and I would like to clarify the terms. Today, Weebit is focused on what's called the embedded market. Many people might have heard of the term Moore's law and know that everything in semiconductors is shrinking rapidly. Today, things have shrunk so much that you can put a big and complex system on a single chip. So that's what's called the system-on-a-chip. And you have their processor and you have sensors, communication. You have all kinds of things there. And obviously, you have to have memory on that chip. So these product companies like the Apple, Google, Broadcom, Cisco, Sony of the world, they want to embed our memory into the system-on-a-chip, and we will license this technology to them to put in their chips. So that is the market that we're focused on today. One of the things to understand, the big advantage of ReRAM over Flash is the fact that our technology is a bit addressable. You can actually access every bit on its own because it has what's called a selector. And in Flash, we actually have to read the whole block, change the bits that you want to write it back. That's why it's so slow and so power hungry and all of that. The only -- today, we are using as this selector, the simplest selector in the industry which is called the transistor. And that works great. And in the embedded space, it's really a good solution. The challenges with these transistors that we use are somewhat large and so the memory array is a little larger than what people might want sometimes. So in the embedded space, that's not an issue because the memory is just one part of the chip. But when you go to the discrete and discrete means a stand-alone chip that is 100% memory. In that space, already everyone wants to cram as much memory as they can on a small piece of real estate as they can. And so the selector size becomes an issue. Weebit has announced already a while back that it's developing an advanced selector. This is a very advanced technology, very complex. It takes years to develop, and we're working on it and we're making progress there. However, we -- now that we announced that we have this technology qualified for the embedded space, we have people come to us and say, "Hey, guys, the discrete market isn't a uniform one big blob. There are different segments in it. And in some of those segments, the technology that you use with that transistor is already good enough and you are actually competitive there. And so now we started working on a discrete solution that will be based on this, what we call a 1T1R, one transistor for one resistor. And we're looking at that and we're moving forward also in that direction.

Okay. And so a viewer wants to ask in regard to the discrete, I know there's a long way to go, but what are the key milestones that Weebit needs to achieve to get it to market from where you are now?

Well, one thing that we obviously need to achieve is to have a foundry or a fab that will manufacture it. So part of it is we're working and I think the viewers know we are working today with many major foundries and many IDMs. IDMs are companies that have foundries of their own to manufacture their own products, like Intel manufactures its own Pentium. So there are some companies that despite the high cost of a fab, they manufacture their own products. So we are working with leading foundries. We are working with leading IDMs, we're trying to get this technology out to more of them and have one of them as a potential for manufacturing this discrete chip and that's also part of the work that we're doing.

And does your agreement with DB HiTek impact discussions with other foundries or IDMs at all?

I believe it will. I believe that, as I said, there's always that human reaction to new things. We're always cautious with something new and SkyWater is a great partner, but it's a relatively smaller foundry. So having someone one of the top 10 that's licensing our technology, I think that's going to reduce that concern level of other foundries, it's going to make it easier for us to move forward. I think every new foundry that we sign up will lower the barrier to entry to the next one. That's how it normally goes. The same thing, by the way, with product companies, product companies today are very concerned to bet their product on a new technology that has -- they've never seen operating in mass production silicon before, and it's normal. As we will move forward and we're getting close to that point of having the first product companies sign up and work with us. As we make progress, as people see that our technology is more accepted in the market, the barriers to entry will go down. And eventually, we will move forward. I think Weebit today, by the way, and I'll sidestep a little bit. But when we talk to the foundries, they realize that Weebit is right now the leading provider for them to work with. You have a couple of -- 2 or 3 foundries that have their own ReRAM technology. But all the others in order to be competitive, they need ReRAM technology. Their customers are asking for it. The world knows that this is the next thing. And right now, Weebit, as far as I can see, is the only independent provider of ReRAM that can work with them. And we have a very strong company. I think they're very impressed by our R&D. We've very strong teams across the board in all of the disciplines that are required for ReRAM. We have a big management focus and so I think it's really a matter of time now, and this agreement from Dongbu or DB HiTek is really going to help us push forward.

And a viewer wants to ask, does the ReRAM device contain any critical metals?

No. I think that's one of the things that we're very proud of that we're using standard materials, no rare earth materials or things like that. We are very environmentally friendly. That's, by the way, in stark contradiction to the other technology that is out there and to be transparent and to give the full picture. There are, today, 2 main technologies that are trying to replace Flash or starting -- I don't want to say trying, they're actually starting to replace Flash. MRAM or magnetic RAM has actually been already in mass production for a few years now. And it's gaining traction, but not as fast. And part of the issue is that MRAM is, first of all, very complex to manufacture, it requires rare earth materials, it uses magnetic material, which is a big contaminant in foundries or in fabs. So you need to have a separate production line that's kind of protected. The added cost there in the manufacturing is very high. So it's both not environmentally friendly and very expensive. That is why analysts today see ReRAM as the big promise. And even though at this moment, MRAM still has the very large market share, they expect ReRAM to be passing MRAM, overtaking MRAM and in a few years already, the analysts are expecting to see the majority of the wafers manufactured with the newer memories being ReRAM.

Okay. And Sorry, I have a lot of questions coming through. Which agreement do you think will be signed first? The commercial agreement with the Tier-1 fab or a customer agreement at SkyWater?

Well, the answer is very simple. We signed -- yesterday, we signed an agreement with Tier-1 fab. So that's done. We are working on additional agreements with additional Tier-1 fabs. DB HiTek was clearly not the only one that we're working with, and we're making progress with them, with IDMs and so on, that's continuing to work. In parallel, we are making also progress with SkyWater customers. We had -- sometimes, you have things that are out of your control. And I think it's important to understand we've had a situation with customers at SkyWater who are ready to move forward. And then they said, "Oh, we need to embed not just on ReRAM, we need to embed some other elements and SkyWater doesn't have that other IP that we need, so we're stuck, we can't actually manufacture our product at SkyWater because of that. It has nothing to do with Weebit. But Nevertheless, we lost that opportunity that we thought we were already going to close the deal there. So you have all of these things that happen. Dongbu or DB HiTek is a much bigger foundry, a much more advanced one that has a much richer IP library. And I think that's also part of it. So we expect to make much better progress working with DB HiTek.

And will the -- will you start the tape-out process with DB HiTek straight away? Do you anticipate any delays? Or do they have the process availability to begin straight away?

So we have to first transfer the technology to them and in parallel, we need to transfer our test chip to fit their manufacturing process. Every foundry has its own process and what's called its own PDK, process design kit. And so we -- process development kit, I'm sorry, and we need to adapt now the test chip to this specific process and make it work there. So these works on transferring the technology itself, what's called the recipe. And it is actually just like a recipe in the kitchen. You need to transfer what are the steps you do with which materials in order to manufacture ReRAM and in parallel, we need to have the design team work on the test chip and transfer it to the DB HiTek PDK.

And speaking to Weebit Board and management team. I know we have a lot of industry professionals on the Weebit Board. So I was wondering are there any key risks associated to closing the deals, if any of these executives were to start working for the company?

Well, first of all, I'm extremely proud of my Board. I think it's very rare to find such a caliber Board in general and definitely in smaller companies so having Dadi and Atiq and Yoav on the Board is just awesome and everyone is always so impressed by it. Having said that -- and by the way, they are contributing a lot, and they are giving a lot of advice, and we talk on a regular basis. It's really great. Having said that, the company management is also very experienced. And today, we already have a working technology in our hands. We already have a qualified technology in our hands and people -- you look at the VPs in the company, people like Ishai Naveh, our CTO, who has been working in ReRAM for, I think, 17 years now and has been in this foundry and nonvolatile members for more than 30 years, 35 years, I think, actually more than that, even more than 40 years or whatever. It's really a long time. Ilan, our VP R&D, who is responsible for the design side. Again, he's been doing memory design for ages. So the company has the team and the people reporting to these 2 guys are also -- there are so many very, very, very experienced people there. I'd hate to lose any of my Board members. It will obviously get low, but it's not going to be something that kills the company. We have a working technology. We're talking to customers. We're moving forward. And the company is very strong, it can survive anyone who will leave it today.

And for those who are new to Weebit, Weebit is based in Israel. And so our thoughts are with you, Coby, I know you and your family are based in Israel and with the company. And so I was wondering whether you foresee any disruptions of the conflict to the business operations?

So well, first of all, thank you. I'm getting many, many messages of support, and it's really heart-warming to really get all of these very kind messages from shareholders, from partners and everyone. It is a very, very crazy and difficult situation. We went through and we are going through a very brutal attack. We have the sirens going off on a daily basis and depending on where you are in Israel, it can be multiple times per hour or multiple times per day or it's definitely not easy to live in such a condition. And -- but on the other hand, I'm very happy to say that, first of all, all the Weebit employees and families are okay. We have a contingency plan for this, unfortunately, living in Israel. This is not the first or second or third time that Israel is under attack. We have -- most of the employees have shelter rooms at home. For myself every time the siren goes off, we head with the family to the shelter, we wait 10 minutes until we know everything is okay, and then we can come out. But I do want to make it very clear with all of these harsh conditions, the Israeli industry has a long standing policy, and it has proven it of not allowing the terrorists to win and part of it is not allowing them to disrupt our work. So our lab is in a protected environment. It's continuing to work 24/7 without any disruptions. The employees, each one of them, we have the settings for them that they can continue to work. And I would say even our resolve to make sure that we don't miss any milestone, that we don't change any plan is very, very strong. We told the market that we don't expect this to impact our ongoing work. And I mean it, we've done it in the past. I've been through these situations at Weebit and before Weebit, and we've always managed to continue to deliver on the milestones and we didn't change work plans because of these crazy bombs.

Well, I'm really glad to hear that you and your family and all the Weebit family are safe and I know all the Weebit investors are thinking of you. So I'm glad to hear that they are safe. And we have some more questions still relating to DB HiTek. I'm just conscious of time, but we'll try and get through as many of them as possible. And Coby, please tell us more about the licensing fee with the DB HiTek deal. Will the fee be upfront or afterwards or another time? How will it be calculated?

So I can't go into the exact details of the deal, both because it is confidential, and we're committed with DB HiTek not to talk about the details. And also, I think people can understand that if we do let some of the information go out, that will become the ceiling with all of the other foundries. They always use that as their open negotiation point. So we don't go into the details. However, there is obviously a licensing fee here. There is obviously an NRE fee and engineering fee for the work that we're doing to transfer the technology and so on. We -- I can't talk about when we will receive which money, but we will be getting paid for the work with DB HiTek. And I guess you'll start seeing some of that in revenue reports as things move forward.

And who are actually some of DB HiTek's major customers?

If you go to their website and open their annual report, you'll see some of the names there, companies like Intel and Qualcomm. And I think the public names that are known, companies like Sony, Toshiba, Mitsubishi, they have quite a few big name customers that manufacture in their foundry that now that we have this agreement with them, we will have access to. So we will be working with DB HiTek to go over their customer list to identify who are the ones that are the best candidates to start off with. We clearly plan to work with them in parallel. We don't expect to wait 2 years until we're qualified or 1.5 years or whatever it will take and only then start engaging with customers, we will start engaging with DB HiTek customers much earlier than that. We will try to push them forward. I hope that now confidence level in Weebit is growing. So we really hope to get customers who will agree to engage with us before qualifications and try to do some of the work in parallel. And it's going to be very exciting now.

And I know we've already moved past talking about the discrete memory, but we just have another question here regarding, can you advise the status of the discrete NVM work stream, given past customer feedback that a discrete memory module could actually be feasible even without a workable selector?

Yes. So to be exact, the transistor is also a selector, and we are looking at that. There's just a lot of work going on in parallel, obviously, work with DB HiTek, for example, had higher priorities than working on that discrete project. So there's progress being made. We need to have -- I mentioned earlier, we need to have a foundry that will also manufacture this. So obviously, the work with the foundries is very important. We are swamped with so many projects and so much work, we just have to prioritize. Right now, my priority is bringing on board more foundries and bringing on board customers, product companies that we can work with, those are the top priority for us. And in parallel, we are working on this discrete project.

All right. Thank you. Well, I'm sorry, I couldn't get to everyone's questions. I know there's still a lot coming through. I'll try and get them sent over to Coby potentially. And if he has time, he can get back to me and I'll send it to everyone that has attended. But thank you so much, Coby, for joining me today, even despite what's going on in Israel. I'm glad to hear everyone is safe and thank you for all those that are watching. The replay will be posted on to our YouTube, so you can find it there. But thank you so much, Coby, for joining me today.

Thank you.