Vik talks with Ilan Sever, Chief Memory Architect at Weebit Nano, about the ubiquitous but often-overlooked non-volatile memory (NVM) inside most chips. Ilan explains why the long-standing solution, embedded flash, stops scaling at advanced nodes like 22nm. He details how Weebit’s Resistive RAM (ReRAM) provides a low-cost, high-endurance alternative that is particularly well-suited for demanding markets like automotive.
Things we cover:
The role of non-volatile memory in MCUs and analog chips
Why embedded flash stops scaling below 22nm
How Resistive RAM (ReRAM) works by forming conductive filaments
Integrating ReRAM in the back-end-of-line with just two masks
Achieving automotive-grade endurance and qualification
The future of ReRAM in AI and in-memory compute
This podcast is lightly edited for clarity.
Introduction
There is a memory in almost every chip that nobody talks about and it holds a few kilobytes to a few megabytes. It has no like spot price or a super cycle. But without it, most chips, analog, especially, can’t work or even meet the specs on its data sheet. So for 30 years that memory has been embedded flash, but as we get to smaller technology nodes, it really stops working. So today’s guest designed embedded flash and then spent the rest of his career on what comes after it.
Hey, welcome to another episode of Semi Doped podcast. I’m Vik Sekar, the founder of Semi Exponent where I do boutique research on semiconductors and AI infrastructure. I also write Vik’s newsletter on Substack, one of the top 20 technology newsletters on the platform covering AI and semis. I come from two decades of working in the semiconductor industry and my work now is taking that engineering background and explaining what matters in the technology and why it matters to the markets.
Check out Vik’s newsletter on Substack and for technology research, reach out to me at Semi Exponent. And our guest today is Ilan Sever, Chief Memory Architect at Weebit Nano. He has spent over 30 years in non-volatile memory, semiconductor IP and SOC design. He was previously group CTO at Dolphin Design, director of IP and libraries at Tower Semi, and design manager of Flash memory at ST Microelectronics. Ilan, welcome to the podcast.
Ilan: Happy to be here, Vik.
Awesome. So, one of the things that we don’t often talk about is what exactly are non-volatile memories because when I mentioned flash memory, that’s there pretty much everywhere, but it is so ubiquitous that people have even stopped thinking about it in the mainstream. It is assumed to be there, but it’s still a very important part of all of our electronics, and it’s not just like DRAM and SRAM, that’s the whole problem, right?
So you’ve worked on this stuff before. So let me start by going through your journey. What made you get started on working on flash memory?
Ilan: Yes, thanks, Vik. So yes, non-volatile memory is basically a fundamental element of any electronic system. It’s where you store your software code and your parameters and anything that you would like to stay even when the power goes away. So, we all have systems in which the battery can run out or we shut it down, but when we power it back up, everything goes back to life.
So where is this stored? All this information is stored in non-volatile memory, meaning a memory that doesn’t need the power in order to retain the data. So, you mentioned flash memory, the technology of choice, even today in what we call NAND storage, like your phone storage is what we call NAND flash.
I’m not talking about this technology. We are now referring to technologies which are embedded within system on chips, on silicon. And this was previously served with embedded flash, which is a NOR type flash. Basically, flash technology is storing charge on some floating gates, so there is some capacitance which is charged and it retains there even if you shut the power down.
And that’s how flash works. As you mentioned, flash doesn’t scale well, so in geometries below 28 or 22 nanometer, flash reached its physical limit, actually. So physically you cannot implement embedded flash anymore and the industry had been looking for alternative technologies for a few decades now. And, we will talk today about one of these technologies that is now gaining really market and acceptance, which is resistive RAM.
What is Non-Volatile Memory?
Awesome. That’s a great introduction. So for people listening, we always talk about static RAM, SRAM, which holds the state, but it holds information without having to be refreshed. However, if you turn off the power to that SRAM, you’re going to lose everything in the SRAM. Like you switch off your computer, your OS is removed from the SRAM and all of that stuff, right?
On dynamic RAM, I think it’s more like you have to keep refreshing it, but of course, you take power out, that goes out too. Non-volatile memory that, Ilan, that you were mentioning is that it’s the stuff that stays all the time, on and it even if you turn off your phone, it holds information and that’s very important because your devices just can’t forget stuff all the time.
So we’ll definitely get to why E-flash stops scaling as we went to smaller nodes and what resistive RAM means. But I just want to like step back just a little bit and say like, exactly what is the role of non-volatile memory in a device? It could it could be an embedded device or a sensor or a power controller or an automotive chip, whatever it is. Like what is the role of this non-volatile memory? What does it actually hold?
Ilan: Yeah, that’s a very good question, Vik. So fundamentally, the base usage or the most wide usage is for any embedded MCU or microprocessor. It has its software. And if you want to have a really embedded, whether it’s an IoT device, a medical device, your hearing aid or whatever device that needs to be compact, low power, fully embedded, without additional external chips. You would like to have your code inside the same system on chip. You have an MCU, whether it’s one of any MCU architecture, it would need to run some firmware, which is software.
Like a microcontroller, right? Like MCU means a microcontroller unit and it has to have, it has to run some code. The question is, where does it get this code from?
Ilan: So, of course, if you want the code to be there, even when someone is replacing batteries or maybe the battery is out or even applications that don’t have power all the time. There are applications today which do energy harvesting and they just get energy from ambient sources and they’re not always connected to any power source. So, in any of these cases, you want the software stored.
So the main usage or the first usage, I would say, is to store your software. But then, in addition, just think about sensors. So you have some IoT device or medical device. It is logging maybe the temperature, maybe your heart beat, maybe some ambient data from some sensors and you want to have a logging of this data. So think about a sensor, it’s logging what happened and maybe after a while it is transmitting the information to some system, whether it’s an industrial sensor, medical sensor, any type of these devices.
So they always need in addition to the code storage that we discussed before, also some data logging storage. Moreover, most of these systems, these mixed signal embedded systems, they have also some analog parts. And for the analog parts, usually you would need to have some kind of trimming. So these are kind of normally one-time programmed data that you do upon production to just put all the analog values in place.
So you use trimming. That’s another usage for non-volatile memory. So you just want to to kind of burn these, this data into the chip so it retains throughout the lifetime of the chip. And last but not least, on security applications, there is always storage of some secret key. So encryption keys are normally stored on some kind of non-volatile memory on security applications.
Capacity and Endurance
Awesome. Yeah, that’s a good, that’s a great introduction because I think what most people don’t realize is take the example of the analog part you mentioned, my own background is in analog semiconductors. So it’s like, it’s amazing that when you’re making a billion of these little analog sensors or any analog power supply unit or a or a power distribution chip or whatever it is, as you manufacture a billion of them, right?
Ilan: Yes, this is correct. And this is where again, one of the use cases. In this case, we are talking about a very small amount of data that needs to be retained throughout the lifetime, and it also is programmed only once or maybe very few times. So when we discuss non-volatile memory, it’s important to state one very important term, which is what we call endurance or programming cycles.
So, for example, if you program a software, then yes, you would need to change it, right? There are software versions, people do even what is called OTA, firmware update over the air today. So your system can update code even in the field or over the air. So all these requires reprogramming of of the data.
Unlike this streaming data or what we just mentioned, that is more like production data that you stored once. In any case, any type of non-volatile memory is able to reprogram data. Not as many times as you can write to SRAM or DRAM, which are the volatile technologies that you mentioned. That obviously you can read and write to them almost infinitely.
Exactly. Okay, this is great because you’re setting this non-volatile memory to be a kind of memory that you mostly write one time or a few times and read many, many times in its lifetime.
Ilan: Uh, yes, so any type of non-volatile memory will need to be reprogrammed. And one of the parameters that we look at when we look at NVM technology is the number of cycles, the endurance of this memory. So, some technologies can only work up to 10K, 10,000 reprogramming cycles.
Advanced non-volatile memory is expected to already allow in the order of 100K, 100,000 to a million reprogramming cycles. And this makes it a very good technology that allows a wide usage for many types of applications.
Awesome. That’s helpful to know because we will obviously talk about like endurance and reprogrammability when we actually talk about what resistive RAM is. One other important point that you mentioned, I think we should touch on is the capacity of this thing. When you talk about, you mentioned NAND in the beginning, you can you have, you can store terabytes on NAND flash, and in DRAM, you can store gigabytes, in SRAM, you can store hundreds of megabytes. So what is the range of non-volatile memory in terms of storage capacity?
Ilan: The capacity of the non-volatile memory depends on the application, obviously. So we mentioned embedded MCUs having some firmware. This is normally counted in kilobytes, but I would say tens, hundreds of kilobytes, to single megabytes. So maybe one, two or four megabytes is a large NVM capacity for embedded MCU. Maybe maybe four megabytes.
So, so we’re talking about single digit megabits, right? Like 1, 8, maybe 16 megabits, 32 megabits for code storage. Of course, all the other use cases like data logging or trimming that I mentioned are much smaller. We’re talking about kilobytes.
Now, in in more advanced MCUs, so maybe some in in lower geometries where people embed non-volatile memory with high performance microprocessors for any type of applications, automotive, industrial. It can reach to, I would say ranges of 16, 32 megabytes.
But we are definitely still not in the gigabit or gigabyte terrain that that is implemented in other type of technologies like the ones you mentioned, DRAM or NAND flash.
The Scaling Limit of Embedded Flash
Yeah, so that’s that’s also very important to understand about these non-volatile memories because you want to essentially, let’s say, reprogram them over the air. What you’re actually reprogramming is only a few megabits, bytes, or a few kilobytes, kilobytes. And when you drive your Tesla into your driveway and it wirelessly connects and updates its firmware, this is where it is actually stored. It is stored in a non-volatile memory, probably on the larger capacity side than you would have in a little sensor. I would assume an automotive non-volatile memory, NVM, with like embedded flash or whatever, is like a larger capacity than a sensor, which would or you for the trimming applications, it’s going to be really small because you only need to correct for like, you know, put in a code, that’s it. So it has to store very little amount of memory compared to what an automotive application would use, but it’s definitely not in the gigabits, gigabytes range and all that, right? So that’s that’s the that’s the whole thing here. So typically we’ve used E-flash for doing this. And now it seems like E-flash has hit some kind of a limitation as we go to smaller nodes. What what exactly is E-flash and why does it not scale to lower technology nodes say below 22 nanometers?
Ilan: That’s a good question. So embedded flash, basically flash memory, is storing some charge in an insulated floating gate. So it’s like a transistor in which the the gate has some charge stored or a floating gate, and it changes the conductance or the threshold of the transistor. That’s the way the data is stored, so you can have charge stored on that floating gate or taken away, and it changes the behavior of the transistor. Now, the amount of charge that you store there, when the technology scales to low geometries like 28 nanometer, 22, and I’m not even talking about single digit nanometer technologies that are available today. The the dimensions of this transistor and the dimension of this so-called floating gate are so tiny and the amount of charge that is stored there is just moving physically to almost single electrons.
A very small amount of charge, and this is just a physical limit. So, and of course, holding this charge over there without any leakage, without the charge going away by all kinds of physical mechanisms is becoming really impossible in the smallest geometries, and that’s why flash,
Okay, yeah, yeah. That’s very helpful. Actually, I wanted to compare this to what like for example, DRAM looks like. A modern DRAM actually uses the equivalent of let’s say a 10 nanometer transistor. But the the charge that is stored is never stored within the the transistor itself because there is a capacitor on top of the DRAM that’s actually stores the charge. And the whole problem of DRAM and capacity and DRAM density is how to get as much capacity by digging a deep trench in the in the silicon and connecting a transistor to it and how to maximize that capacitance without having to rely on the transistor to hold that charge. But what you’re saying is like in the in the flash world, the transistor has to hold the charge. And then you make the transistor so small that you can hold basically a few electrons of charge in it and that’s about it because it’s so small. It physically stops scaling. You you need to hold more charge than that for which you need larger transistors, period. And that’s why flash does not scale beyond 22 nanometers. So, there’s no flash memory in FinFET transistors or below.
Ilan: Yeah, this is a very good point because you mentioned all kinds of vertical capacitors and trench into the silicon that are done in in discrete DRAM chips. When we are talking about the embedded world, it’s very important to note that the whole idea of embedding a system on chip with memories and with analog is that you would like to keep a plain CMOS technology that allows to integrate on the same chip with the same basic technology features. You would like to have analog components, you would like to have digital components, you would like to have SRAM because there will always be SRAM and you would like to have an NVM, but you will never invest a special trench capacitor or you will never invest a lot of special technology because it will make the whole system on chip too expensive. So another key parameter that we look when selecting a non-volatile technology, I mentioned before endurance and the performance of the memory, is how much does it interfere, how much does it add on top of the baseline CMOS technology? In the end of the day, semiconductors is very cost sensitive. So it doesn’t seems like that, but just imagine how much is the selling price for such an embedded MCU. If one would need to add too many processing layers, which eventually translate to money and to time of manufacturing the silicon wafer, that will make the whole system on chip too expensive. Imagine there is a large system on chip that has everything I mentioned, digital analog, and it has one NVM component, one NVM code storage. If you add more layers of processing, special structures, you pay this additional cost over the whole area of the chip, which is making the whole system on chip more expensive. So a key parameter in adding any technology is how many masks. We count basically in in semiconductor processing, we count masks. So how many masks does the technology add on top of the plain CMOS, plain vanilla technology? And that’s another key element because in the end of the day, this means how much are you paying premium for having a system on chip with embedded NVM. So when you mentioned, for example, in in NAND flash, that’s a totally different technology with 3D integration and many levels. And when you mentioned, for example, discrete DRAM with trench capacitor, these are expensive technologies, but they’re optimized for a single use case, which is an NAND flash chip or a DRAM chip. In embedded, you need to have a plain vanilla technology with the minimum additional layers to support all the type of memories that you need.
Yes, okay, it’s very important. Because you’re saying that, look, the the cost of these microcontroller chips that require these non-volatile memories is already very low. Now, if you want to add like a whole lot of masks into the semiconductor process just because you have to implement this non-volatile technology, it better be very inexpensive because otherwise, you can’t have the the whole bill of materials for the chip go up just because you want to add an NVM materials inside. So, non-volatile memories should come with as few masks as possible. So, with that being said, how many masks does a typical NVM embedded flash process add?
Ilan: Yes, so important to to mention. So, embedded flash adds another, I would say, polysilicon layer, a floating gate layer on top of the transistor. In some cases, there is split gate technologies. There are all kind of flavors of that. But they add a lot of processing and this processing is done in what we call the front end of line. So, in the transistor, very close to the silicon substrate where you actually implement your transistor. And for making that, we’re talking about in the range of 10 additional masks, maybe eight to 10 additional masks just to process these additional features required for embedded flash.
Okay, that’s what percentage of the total mask count would you think that is? A typical CMOS process has, would you say, like 70, 80 masks?
Ilan: Uh, even less. I mean, a typical CMOS will have even less masks. And so yeah, this can add 25, 30% add-on and it can become very significant in the cost.
That’s a problem. That’s that’s what we’re actually trying to solve here. And when you’re trying to go to smaller transistor nodes, anyway, you can’t store enough charge and you can’t add so many layers if you have to add 30% more layers, that’s not a cost optimal feature. So, I think this is a good time to get into, we’ve set the stage exactly as to what is non-volatile memory, why is it used, where is it used, what are its features, and why what is required from a non-volatile memory in terms of cost, simplicity of implementation, capacity and everything. So, I think we have set the foundation of non-volatile memory and where it lies. I think the next interesting step is to go into what Weebit Nano actually does, which is resistive RAM. Like, this is a nice time to get into what that is and how it solves many of these problems that we just talked about.
Introducing Resistive RAM (ReRAM)
Ilan: So, indeed, resistive RAM is a technology that is aiming at replacing embedded flash. It has several key benefits related to our previous discussion. First of all, it is very low cost. It is adding just two additional masks to the technology. It is embedded only in the what we call the back end of line. So, we actually build the structures between the metal layers above way above the silicon substrate and transistors. Now, why is this important? We mentioned before all kind of mixed signal chips, analog chips that require some non-volatile memory. These technologies have passed through a lot of optimizations of their unique analog characteristics. For example, some technologies have BCD transistors, which are high voltage transistors. And when you invested so much in an analog technology to have the the perfect analog property,
One second, I just wanted to that is a very, that’s a very interesting point that you mentioned here. That’s because I think people need to understand where resistive RAM sits inside the stack of a silicon chip. What you were saying is that back end of line means that it is sitting like very way on top of the chip, where if you think the transistors are way below, the substrate is way below. This is all the way on the top. So it’s actually processed almost very much at the end of the chip processing. And so it’s very decoupled from what is happening in the transistor world downstairs. And so you also mentioned BCD technology. So for those who have not heard of this, it stands for BCMOS, CMOS and DMOS. And this is a specific technology that is good for high voltage transistors. It’s used a lot in automotive applications, because it can withstand a lot of voltage, like tens of volts. So these transistors are unique to what they do. Like they are used in like power conversion circuits and things like that. You you don’t want to go and put a a storage and non-volatile memory storage technology that messes with the transistor structure. So that would make the whole BCD transistor, which is so well optimized for power performance or like whatever that is supposed to do at high voltage performance, it’s going to mess up because the non-volatile memory is like interfering with that. So the best place to put it is like really far away and process it really after all the silicon is done at the back end of line right on top of the chip, right? So that is what resistive RAM does with just one or two added masks, which makes it very inexpensive.
Ilan: That’s so, that’s a right on. So, what is resistive RAM? So basically, as opposed to flash that we said is storing charge, resistive RAM, and the type of resistive RAM that Weebit Nano is doing, which is called OxRAM. So oxygen vacancy based resistive RAM, is actually implementing a resistive element, which you can change between two states, a low resistive state and a high resistive state by applying a voltage. And so I will try to explain it with this demo that we have here. Let’s see if it works.
Yeah, if you’re listening to this, this is a good time to switch to YouTube because Ilan has like a a nice 3D model of a the internal structure of a resistive RAM cell and I’m just describing what I’m looking at here, which changes between its low resistance and high resistance states.
Ilan: Resistive RAM is actually a resistor that can switch between two states, a high resistive state and a low resistive state. What we do, we take an insulating layer, like a capacitor that has a very high resistance. And by applying electrical voltage and current, we actually create these oxygen vacancies, that are, that you can see here, that are actually forming this filament that is a conductive filament. So with oxygen vacancies, charge can go through and this is becoming conductive. Conductive mean lower resistance. So when the memory has this filament built, the resistance is low and we are storing the value of one. Then by applying voltage in the opposite direction, what we can do is actually dissolve this filament. So in this side, in the high resistive state, we dissolve the filament, it goes back to an insulating layer that has a high resistance. So, current can no longer go through or the resistance is high. So we have these two states, either the filament is created and we are in the low resistive state, or the filament is dissolved and we are in the high resistive state. So these two states, we then read them with some, sense amplifier, which is a typical circuit that is used to read any type of memory, whether it’s flash, DRAM, SRAM. So we can actually sense whether the filament is there and then we read a one, or the filament is dissolved and we read a zero. So that’s basically the way resistive RAM is programmed by applying voltage that will create the filament or voltage in the opposite direction that will dissolve it. And it is read by reading or sensing the resistance, whether it’s a high resistance or a low resistance.
Why ReRAM, Why Now?
Okay, this is fascinating. Usually these are not the kind of devices that we see inside a typical CMOS technology process. So what I, what I find fascinating about this is that clearly you can’t use any kind of a dielectric material that creates oxygen vacancies, right? There’s like, there should be something that is different. Now, I know Weebit Nano would probably have its proprietary materials, but would you be able to broadly explain what, what is the property that causes these oxygen vacancies? What kind of materials do this and how is it compatible with the CMOS process flow?
Ilan: Yeah, so we’ll not dive into the whole physics of resistive RAM, but basically, I mentioned that we have an insulating layer, but on top or bottom of it, and there are two electrodes. So there is a bottom electrode that is built from some kind of inert material, and there is a top electrode that is built from what we call a scavenging, layer. So the scavenging layer is the layer that can actually, kind of store oxygen or give it back. So according to the, electrical field, this scavenging layer either will take away oxygen from, the dielectric material or will, give back oxygen, to it. So by applying the electrical fields in opposite directions, we can either have, oxygen vacancies inserted into the material or oxygen going away and creating vacancies or the opposite, contribute back, oxygen, to, dissolve the vacancies.
Awesome. Okay. That’s, that’s unique. Like, so this, this material is actually, compatible with CMOS foundries?
Ilan: So, maybe the most important and overwhelming property of resistive RAM is that to generate it, we at Weebit Nano use only fab friendly material. So we use materials that are available in any CMOS technology. We don’t use any rare earth material. We don’t use any expensive and we don’t use any contaminating materials. So some technology that maybe we’ll not discuss in this podcast, like magnetic RAM,
Yeah, that’s resistive RAM itself is like not an idea that is like new, right? You mentioned Webit Nano has been doing this for 10 years. But why now? What is what has changed now about resistive RAM that we should, people have been trying this since Panasonic have have tried it, Samsung has tried it. What’s what’s the need now? What what has changed?
Ilan: So, I think it’s a it’s a combination of many elements. One is what we mentioned before, that, you know, flash is hitting the physical limit. And cost is, we know that the cost pressure on semiconductor is always growing. There is the Moore’s Law, and there is a lot of pressure to reduce cost. Moreover, I think one of the driving application, driving is a good word here, is automotive. You mentioned it before. So, we at Webit, were able to demonstrate that this technology is not only low cost and has a very good performance, but it can actually serve this very challenging market of automotive. So, we all know that in our cars today, everything is electronic. There are, you know, hundreds of chips, and many of them require non-volatile. Today, resistive RAM technology can serve the automotive market, and I think it’s a major breakthrough.
So, that’s nice. You mentioned that the competitors to this technology are MRAM, magneto resistive RAM, and ferroelectric RAM, both of which are alternative materials and not really well suited to be integrated into a a CMOS fab. So, those are those are like technologies are ruled out, but then there’s also I think another alternative technology called like phase change memory, which, you know, alters its physical crystal structure with voltages. So, like this is a resistive RAM is an other application where automotive is driving a lot of requirement here, especially as we go to smaller technology nodes, where flash stops working, resistive RAM is really well suited to automotive. And so, has this technology been qualified already for automotive? Because I know that there’s this standard called the AEC-Q100, which is like a very hard automotive standard to pass. Does Webit Nano already pass this standard?
Ilan: Yes. So, indeed, our technology is now qualified for ACQ100 automotive. And we are continuing this with additional customers. So, we recently announced two I would say leading IDMs that are both active in the automotive space that license our technologies for automotive platforms.
More Than Just a Bit Cell
So, the way that this resistive RAM technology works from Webit Nano is that you guys have the secret recipe that explains how resistive RAM can be made in a fab, and a fab will license it from you. Is that how this whole thing works?
Ilan: Yes, but that’s where when you asked me why resistive RAM is now becoming available. It’s because it’s not all about the secret sauce. So, the secret sauce, for sure, it was developed throughout decades, and we have now all the recipe for the the exact dielectric structure and the exact electrodes that are needed to to, you know, do the scavenging of the oxygen and all of that. And we really have, you know, tens of PhDs in Webit working on this part along with our partners, as I mentioned. But that’s not enough. If you are an embedded, I would say, MCU, let’s say, a system on chip architect. What you need is a full-blown memory module that you can just commission a a read command, a write command, that’s it. All the rest is done. You don’t work with the bit cell, you don’t work with physics. You you need to write a 32 or 64-bit word onto the memory and be able to read it back when you need it. And that’s where Webit Nano is today the only fully vertically integrated and fully independent IP provider. Because we at Webit, we have all the pillars that are needed to build such a technology. So, yes, you need the bit cell that I showed and the physics around it and the secret sauce. But around it, you need to build the design and the layout of the full memory array. You need to design the the and layout all the peripheral circuits, like the the decoders and the drivers and the sense amplifiers that I mentioned before, which are basically of analog nature, that are able to sense these exact resistances. And for programming the resistive RAM, we need to apply, as I mentioned, electrical field, which is actually voltages and currents that needs to be controlled very precisely. So, for that, you need a whole analog design and layout team. And then on top of it, the programming sequence of resistive RAM is done using a proprietary smart algorithm. So, not only that we did an analog capabilities, we also need a very strong algorithm team, we need a very strong digital design and verification team. So, basically, what I’m telling you is that the resistive RAM module that our customer actually embeds into their SOC is comprised not only of the physical bit cell, it has a lot of analog design, memory architecture, layout, digital design, algorithms, and and even software is developed at Webit in order to have a full offering of embedded resistive RAM IP.
Yeah, it sounds easy, but actually what is unique at Webit Nano is that all these teams are sitting together. So when there is a discussion of, let’s say, a performance that needs to be met or automotive, let’s say, meeting automotive qualification, you mentioned before, it requires meeting very high temperature profiles and high endurance. And in order to meet this, you cannot do it by just the bit cell. You cannot do it by just some analog design. You need a fully multi-disciplinar approach that takes the best of the device physics and the process, the best of the analog design, the best of the digital and the algorithm, and wraps everything together into this product. And and that’s the main strength of Webit Nano and that’s the main the probably the only way that you can have any NVM technology be commercially viable. It was done before in flash memory. It is mostly done in other technologies. But for resistive RAM today, we are doing it and all these capabilities exist inside the team at Webit.
Performance and Future Applications
So you mentioned endurance was one of the factors that you have to really design in as well. It’s not as easy as it seems. What in the rough ranges is the endurance of resistive RAM? Like what and how does it compare to like, let’s say, flash memory or any other alternatives if you have the numbers off the top of your head? I’m trying to place like, is it more endurance, more endurance or less endurance or where does it fit?
Ilan: I would say, typically, a flash is 10K cycles. It depends, of course, there are some technologies that are more, but I would say 10K cycles typically. And for our resistive RAM, we have some products that are 10K. We already demonstrated and qualified 100K cycles, which is opening the market for additional applications, automotive and others. And we are working on extending it beyond the the 100K into the million cycles in the near future.
Okay, awesome. So it’s going to be quite performant in spite of the treacherous conditions that automotive applications want from, I don’t know, from driving cars in Alaska to the Sahara Desert. You have to support all temperatures and so it’s usually a very challenging qualification to pass. And especially the life’s the lifetime of memory used in automotive is also should be pretty high. So all of these things should fit broadly into that umbrella. Where do you see resistive RAM going next? So what’s next? So, you know, we already spoke about microcontrollers. What’s the next big thing?
Ilan: Yes, that that’s a very good question, Vik. So we all know that the next big thing or the existing big thing is AI. So, resistive RAM, or I would say in the academic, naming is called memristors. That’s kind of how the academy wants to to name it. Had been, studied in, multiple academic researches, for what we call in-memory compute or IMC. Webit is also very active in this field, probably a good topic that we can discuss in one of the the next podcasts. But basically the idea is that resistive RAM, a a different type of resistive RAM array can actually create AI calculation inside the memory by just combining resistive elements. And this had been demonstrated in the academy and commercializing it, I would say, is one of the next big things for resistive RAM.
An Innovation Story
Amazing. That’s that’s would be very interesting to hear. Analog in-memory compute has been around as a concept for a long time. And as we have like the AI era coming up and GPUs sucking like 1,000 to 2,000 watts per GPU, maybe the future lies somewhere where we can do it differently. But that is a big topic and a topic for another day. All right, Ilan, before we wrap up this, I want to hear like, what is the one story of innovation that you faced or came across while developing this technology within Webit Nano? That’s a good story to tell for the whole audience.
Ilan: Yeah, I can think about such a story. We mentioned analog trimming before, right? So, embedded system on chip mix and chips will always have some analog, but I also mentioned that we have our own piece of analog within our memory. So obviously we need also to trim that part. So we came up with the idea, hey, we have non-volatile memory and we need to do trimming. So let’s just take some part of the non-volatile memory, just some additional row, store our own, data, our own trimming values there, read them and we can just use it to trim our own system. And then we when we went to to our lead customers and we explained to them and they asked us, how do you do this? And we said, yeah, we have our own mechanism that we use our own non-volatile memory to trim our system. They say, why? We want to use that as well. So can you grant us access? And we would like to trim our own system and it it just saves us having additional NVM, for trimming our system. And today it became a standard feature in all our modules in which we let the end customer get his own, user data, his own trimming data, right out of the NVM, and this is done automatically and became a lead feature.
That’s amazing. So the key takeaway I’m getting from all of this is that it is important that you solve problems. It doesn’t have to be the customer’s problems. Even if you solve your own problems that you come up sometimes, you just have like a new product on your hands or a new problem you solved for someone else. And you can turn around and sell your own solution to your problem to solve someone else’s problem and it’s like a perfectly viable business model.
Ilan: I guess that’s what you call engineering, right?
Yeah, that’s that’s amazing. It’s a nice story. Awesome. Thanks Ilan for coming on Semi Doped. It’s glad to have you here. I know we don’t often talk about things like resistive RAM, but that’s what Semi Doped is for that we sometimes try to go outside of the normal narrative of AI, semiconductors, memory, optics, you know, we talk about all this stuff, but every once in a while we like to throw a curveball and have someone like you explain what resistive RAM is to us. So thank you for coming on the podcast.
Ilan: Thank you for having me, Vik.
All right, so that’s the end of the podcast. So please check us out on all the podcast platforms if you haven’t already. Apple really likes it if you give us five stars, they push us up the technology rankings and it’s always a competition as who gets to the top. So if you can go press the like five star review thing that really helps us out. And as usual, check us out on our daily newsletter, which is daily.semidoped.com and my newsletter is Vic’s newsletter.com. And Austin, who was not on this episode, but is usually on the podcast with me, also has chipstrat.com. All right, see you on the next one.
