Weebit Nano (ASX:WBT) at the Stocks Down Under Semiconductor Conference on 30 November 2021
December 1, 2021
video, WBT, Weebit Nano
Weebit Nano CEO Coby Hanoch presented at Stocks Down Under’s Semiconductor Conference on 30 November 2021.
Due to technical issues during the Q&A part of WBT’s presentation, we have transcribed the Q&A below.
See the full transcript of the entire presentation further down.
Is the SkyWater license exclusive?
The license to SkyWater is non-exclusive. We are talking to other fabs as well.
At this early stage are you aware of any impacts of Omicron on any of your many activities?
It is too early to tell how Omicron will impact us, if at all. Over the past 2 years we learned to work alongside COVID, leveraging remote communications and remote work and minimizing travel. I believe we will manage to continue our work without impact from Omicron, and we will update the market if we see any severe implications.
How long it will take for the qualification process?
As we notified the market, our current plan is to finalize the qualification of the first module by the end of 2022.
What does Qualified regarding Weebit Nano go to market mean? What steps need to be taken?
Qualification is the process of demonstrating that the module is ready for mass production. For that we will need to run many wafers and show that the percentage of dies on these wafers which operate according to the specification (i.e. the yield) is very high, making the technology commercially viable. We will need to do several runs in the fab, followed by testing and potential adjustments as needed.
How do write and read speeds compare to flash once you have a commercial product?
At the cell level we are close to 1,000X faster than flash. When you add the module and “overheads” required for a commercial product this will go down to about 100X faster than flash.
A slightly technical question: for advanced System on Chips (e.g. processor <28nm) how is Weebit Nano RERAM (@28nm) integrated on that chip, given they are different fab lines?
Our ReRAM is embedded in the chip, so it will use the same technology as the chip. The advantage of our ReRAM is that it uses standard materials, so it can be adopted by any fab rather easily. We also have the advantage that we can go down to smaller geometries and, as we do so, the designs using those geometries will be able to embed our ReRAM in their chips.
Are you aware of Akida (BrainChip)? Do you have a view on its capability?
We are aware of BrainChip and their progress. I don’t know enough about their technology to comment at this point.
Could you position Weebit Nano against your direct RERAM competition?
We believe we are currently the leaders among IP companies developing ReRAM. Some were either acquired (and it is not clear what the acquirer will do with their technology), or they hit financial and/or technical issues. We demonstrated a working Mb array 3.5 years ago and released retention and endurance data even before then.
There are also larger companies, such as TSMC, with own ReRAM technology at different stages and we are trying to get more technical information on their technology before we can comment. From what we have seen up to now we believe we have a very competitive technology.
Also, we continue to center our development strategy around fab-friendly materials, tools and process flows, with the aim of delivering the most cost-effective NVM solution compared to any other vendor.
Mark: If I can invite Coby to the stage, we’ll kick it off with Weebit Nano. And Coby’s been with the company for a number of years now. And I have to apologize to Coby straight away because in Israel right now it’s 12:14 a.m. Sorry for that, Coby. We’re very happy that you could join us. You’ve hit major milestones just recently, and I’m sure we’ll hear you talk about that. So, I’ll leave the floor to you, Coby. I’ll be back here once you’re done and we’ll go through some questions that we get from the audience. And, yeah, we’ll pick it up from there. The floor is all yours.
Coby: Okay. Good. Thank you. And thanks everyone for joining us. So Weebit…I guess…I think you’re supposed to share my presentation. So, yeah, there we go. Okay. Good. So, yeah, Weebit is, as you said, we’re in the ReRAM space. We’re an Israeli company that’s developing the ReRAM technology and ReRAM is a new emerging non-volatile memory. So those are the memories that retain the data even after you unplug them from the power. The existing technology that’s prevalent everywhere is called Flash, and ReRAM, in general, and specifically our technology has a lot of advantages over Flash, and I’ll talk about that in a future slide.
So Weebit has made very strong progress recently. We’ve taken our technology to the point where it’s showing very positive results. And based on that, we managed to sign our first commercial agreement with a U.S.-based fab called SkyWater. SkyWater is now licensing the technology. So, they’ll be paying us a license fee. And we’re now very busy transferring the technology to them.
Before I dive into the technology, I think it’s important to note, you know, since in Australia many people don’t understand the semiconductor industry, I think the next best thing is to see which people put their names on the company. And in this case, you know, when we look at our board of directors…I mean, my background is 40 years, over 40 years already in the semiconductor business.
I’ve been involved in quite a few startups and a couple of exits. David Perlmutter, our chairman was number two at Intel. He’s the guy who drove all of the Pentium processors until recently, until about five, six, seven years ago. Yoav Nissan-Cohen, he actually did his Ph.D. under Dolf Roman [SP], the guy who invented non-volatile memory. So, he’s been in this industry since day one and he’s also the founder of Tower Semiconductor, one of the largest fabs in the world. So, you know, he really supports us in many ways with his knowledge and contacts.
Atiq Raza who took AMD from a lesser-known company to what it is today. And we also have, of course, our two Australian board members, Fred Bart and Ashley Krongold. And I think the management team as well, you look at Ishai, Amir, Ilan, Eran, each one of them has I think on average about 30 years of experience in their domain. These guys are real experts in semiconductor, in memories, and so on. So, it’s really a very, very strong team, and it’s pushing the company really nicely forward.
So, to just introduce you a bit and, Mark, you gave a good intro to the semiconductor space, but I think that for investors the best way to look at it is let’s see who the top market cap companies in the world are. And just 10 years ago, they were mostly energy and finance, and other kinds of companies. There were only two that were related somehow to semiconductors.
Today 9 out of 10 are basically semiconductor companies or heavily dependent on semiconductor. And I think that’s really the transformation that’s happened in this world. Everything you look at today is based on semiconductors, be it cellphones, or computers, or cars, and all the way to washing machines and, you know, whatever. And really, microwaves and so on.
The investment now…and you talked about the cyclical nature, and it’s absolutely true. Right now, we’re at the point where there’s just incredible amount of money being poured into this industry. Intel spending $100 billion U.S. dollars on fabs. TI with $30 billion. You know, the U.S. announced 52 billion going to be invested in semiconductors. Europe, it’s actually hundreds of billions. It’s not clear yet, the amount. It’s really just amazing amounts of money that are being poured into this industry.
And in this industry, memory is really the key part. As I said, everything has semiconductors and any semiconductor has memory in it. So, the demand for memory has been growing like crazy. You see surveillance cameras being put everywhere. That data needs to be recorded somewhere. Everyone’s backing things up to the cloud, and the cloud, its huge data center’s full of incredible amounts of memories. So, it’s really very, very exciting to be in this industry at this point.
And really, right now memory has become the largest segment in semiconductors. So, when we talk about that, we’re actually talking about the discreet memory or the standalone memory chips. But even in all of the other sectors, the memory has to be embedded into their chips. So again, memory is just all over the place.
And, you know, there’s really a lot of talk now about the non-volatile memory, and it’s really not managing to keep pace. I’ll just mention, you know, the top point here. Some of you might’ve heard that Tesla had a recall of 160,000 cars a few months ago. It wasn’t because of the engine or because of the braking system, it was because of the Flash, the non-volatile memory that was worn out. So that’s just one example of how much this…all of the different industries are being dependent now on semiconductors and on memory specifically.
In terms of Weebit and our ReRAM, the ReRAM technology and specifically our ReRAM is very cost-effective. It’s very easy to manufacture. Weebit, unlike most ReRAM developers or many ReRAM developers, has been very, very focused on using standard materials in the fab. And you need to understand the fabs where you manufacture semiconductors are very, very complex factories. You really…I mean, the simpler fabs cost a billion dollars to set up. The more advanced ones are just billions and billions.
And it’s really…coming into a fab with something that is not standard that requires new tools, new materials, new processes is just a nightmare. And we’ve been very, very focused, you know. The experience that we have in our management team in our board, you know, the directive is very, very clear. You have to absolutely do things in standard way and make it as simple as possible to go to the fab and ramp up manufacturing as fast as we can. So, we’ve been using standard materials. We basically have two additional masks in the manufacturing process.
So that’s basically about 5% added cost to the wafer. Just to give you a feeling, the Flash adds about 10% to even 20% to the cost, depending on the specific Flash. And other memories like MRAM and whatever are even more than that. So, we’re very cost-effective. Our memory has low power consumption, significantly lower than Flash. We have very good endurance, much higher than the Flash, and so the Tesla example is a good one. We wouldn’t have had that problem. We’re very fast. We’re tolerant to radiation, which is important in some industries. And, you know, the high temperature, again, is a very important thing, especially in cars.
So, when you look at this…at the memory market, it’s really…I would say you can split it into two, and I mentioned it earlier. There’s the discrete or the standalone memory chips, and there’s the embedded memories where you actually embed a memory into a system on a chip. So, since everything has been shrinking so much, you can put a full system on a single chip, and you embed the memory into it. And I didn’t mention it, but it’s a good opportunity to mention the size of the market. This market is growing rapidly and is expected to reach over $100 billion U.S. in just a few years. So, it’s really a big market and a very exciting one.
When you look at the embedded market, this is really where Weebit is focused right now. The requirements from this market are lower than the requirements in the standalone market. So, this is our first market to approach. We have a lot of advantages in this market, and we can specifically address applications, like mixed-signal power management, which is really a very fast-growing domain. Automotive, IoT, AI, etc. So, you know, it’s very exciting now talking to customers.
One of the big advantages that we have is in the embedded space we can go down to very small geometries, to the smaller nanometers, what’s called. And Flash basically is stuck at about 40 nanometers. So, all of the advanced applications today have to jump through hoops to really connect to the non-volatile memory, and we’ll just enable them to embed the memory in a very simple way.
We already announced that we have 28 nanometers, and we’re going down below so we already are entering this domain, which is basically…Flash is not there. You can see here on the right-hand side, by the way, the embedded, emerging non-volatile memory market today is still almost nonexistent, but you can see just how fast it’s expected to grow over the next few years. You know, obviously, Weebit is going to be part of that growth.
So, some of the news about Weebit. The biggest announcement that we had is the licensing with SkyWater, SkyWater licensing our technology and now we’re transferring technology to them. They’re a great partner. It’s really a very good fab for us to work with, and it’s a great cooperation. You know, they are U.S. based, and the U.S. is now making a heavy investment into semiconductors. I think it was not a coincidence when President Biden was announcing a $52-billion investment in semiconductors that he was holding a SkyWater wafer in his hand. So, we’re really looking forward to this, and we’re now focused on transferring the technology and then getting it qualified so that by the end of next year we should be qualified, and then be able to start mass production.
We also announced earlier this year, in the middle of the year, that we have…we’ve taped out a memory module and even a full subsystem, or I would say a full system on a chip that demonstrates how our memory is used. The memory array is important, but one of the key things is to have the memory module, which controls it and manages it. And there’s a lot of intelligence in the memory module that makes it more efficient, lower power, etc.
We went all the way to build a system on a chip that has a processor and all on it so that we can use it as a demonstration vehicle for customers, potential customers. So, we’re going to have the wafer come out from fab in the near future. And then we’ll be…as soon as we test it and verify everything is okay, we’ll be packaging it and be able to start demonstrating it to the customers.
And I mentioned 28 nanometers. This was a very, very important step for us. Twenty-eight nanometers is really one of the sweet spots of our industry. There’s a lot of activity going on at that level. It’s already entering the domain where Flash really has very big difficulty or just can’t go into. We’re focused now on going to smaller geometries. All of the advanced AI, your automotive, etc. is being done at smaller geometries, and we plan to go into this space and be a key player there.
So, you know, there’s really a lot of work going on on the embedded side. That’s really where the focus is. By the way, I should’ve mentioned the agreement with SkyWater is non-exclusive, so right now, we’re very focused, on the one hand, on transferring the technology to SkyWater and getting it qualified etc., but in parallel, we are talking to other fabs, and we are talking to potential customers and really looking for those first customers who will sign up. Even, you know, most customers will wait until the technology is qualified but there are always those who will move forward faster and we’re looking for those.
We’re talking to some of those. They’ve already asked SkyWater for their PVK. What’s called…it’s the…what’s needed in order to start designing their fab. So, it’s really very exciting. But at the same time, we’re not forgetting what I call our mid-term technology, which is the discrete or the standalone memory chips. We’re working on that. We believe that we can have a good technology there. And it can replace the Flash chips. You know, there are different domains and we’ll start with one and then move on to the others. Again, you can see this is a very fast-growing market. It’s still going to take us time to actually get a product in this domain so it’s not in the near term, but we’re clearly putting a lot of effort to get there because it’s a very big market.
So, the big news that we had on this front was that we integrated with Leti’s selector. The selector enables us to have a much smaller memory bit and that enables us to have a lot more memory on the same piece of silicone. So that was a very important step in this process. Again, it’s a long process that will…we still need to grow this now and start having small arrays and larger arrays. So, there’s work happening here as well.
And maybe last but not least, if I talked about embedded as our immediate term and discrete as our mid-term, in the long-term we are looking at what’s called neuromorphic technology. It turns out that the ReRAM is actually…and ReRAM still operates in a very similar way to your synapse in your brain. And you can use ReRAM to emulate a brain, not simulate like people do today, but actually emulate it so it’s much more efficient, orders of magnitude more efficient.
Today this is done mostly in research institutes. It’s still in very early stages. We’re giving these research institutes our ReRAM so that they can use it as the basis for their research. We have very good cooperation with them. And we really plan to be there once this is ready to be transferred to the industry. We wanna be there and be the leaders in this domain.
So that’s kind of really Weebit overview. I think, you know, to summarize, we’re a leading ReRAM provider. We’re moving very, very fast now. The technology’s in a good shape. It’s being transferred to a production facility. We’ll be qualifying it by the end of the year, next year, so that it’ll be ready for mass production. So, we have already that first licensing deal with SkyWater, which was a very important step forward.
We’re talking to additional fabs. We’re talking to customers. We’re gonna have more licensing agreements happen next year. And we’re really looking forward to that. I think the key here is our very experienced board and management. And, you know, I’m very, very proud to have this team support me in driving the company forward. So that’s, I think, basically the presentation, and I guess open to questions.
Mark: Excellent. Thank you very much, Coby. Well, we’ve got a number of questions coming in, so let’s kick it off with Katherine [SP]. She’s asking about the license with the SkyWater. Is the license exclusive?
Coby: No, it is not. It is not exclusive, and we are talking today to other fabs. We plan to license the technology to additional fabs. So, I mean, on the one hand, I am very focused on making sure that the SkyWater transfer and all of the work with SkyWater is successful. So, we are putting an emphasis there. But at the same time, we’re already talking to other fabs.
Mark: Right. Okay. A question from Warren. At this early stage, are you aware of any impacts of Omicron, or Omicron…I always say micron for some reason…Omicron on any of your many activities?
Coby: It’s a little early to say. I mean, at this point, it hasn’t had an impact yet. I don’t know if it will or not. We’ve learned a lot over the last year and a half. And we’re doing today a lot of work through Zoom and in other ways. So, I mean, we do need to travel and meet our partners, if it’s at Leti, if it’s at SkyWater and if it’s potential customers etc.
Learn more about ASX-listed Semiconductor stocks with
Stocks Down Under!
Subscribe to Stocks Down Under today!
No credit card needed and the trial expires automatically.