Battery Revolution Clubhouse Recording - Battery Scale-up using EIS

0:00Transcript

0:00So yeah, and also before we start off the session, we did mention about Battery Den, which Mariam was part of in our first ever Battery Den about a month ago. So we really encourage everyone who's interested in a startup, if you're a startup, please do feel free to join us. Mariam has a great experience there. Maybe I'll have her share more about her experience after this as well. And we do encourage everyone to sign up for Battery Den. It will be a great time to network with our panelists. And feel free to reach out to Simon and myself if you wish to have more information on that. So without further ado, maybe I'll pass on to Simon to also share our Kickstarter session. Thank you so much, Catherine. Yeah, it's really exciting. As you can see in the top, Catherine mentioned earlier, it's our 23rd session. So it's fantastic, you know, how it's been going on and really grateful for all of you to stick around with us.

1:00And yeah, we kind of have different topics for each of these sessions. And we have a conversation starter. And today, it looks like we actually got two, which is even more amazing. Maybe just one quick word as well, as Catherine mentioned. So we kind of met Mariam through something called the Battery Den, which I don't know, it happened probably now about one half a month ago or so. And essentially, it's an event where startups can pitch their, you know, their really exciting innovation and the exciting work they're doing. And Mariam did a really fantastic job, and she maybe can also tell a bit about if she likes to experience with the battery itself. But yeah, today, she's going to talk about how to scale up viewing EIS. And I think for many of us who maybe are not familiar with it as much, I'm really excited. Maybe she can also share a bit of background on the technology itself. And yeah, with this, I will let you take it away and introduce yourself and the topic.

1:56Thank you, Catherine and Simon. And I do want to say one word about Battery Den, which is Battery Den is this, to me at least, speedfire way to gather a lot of feedback, but also in a room with people that understand the space and understand the challenges that the innovations are meant to address. So it's extremely valuable, even if you're just listening in, to actually look at the diversity of innovations that are coming out for startups. And then also get feedback from veterans in the industry. So I really enjoyed my time at Battery Den. And with that, I'd like to start by first sharing that last week, my team and I, we decided we really needed to adopt a goal setting framework for the company. So we set out, we did our research into all these different frameworks in order to adopt a way for us to set goals and track them over time. And that's because with the team growing, with more deployments happening, we started to face complexities with things like tracking performance, aligning the team, but also delivering effective communication.

3:20And so like many companies, we adopted a framework called the OKR framework, which advocates for essentially setting objectives and measuring them using key results that are measurable, but impactful, so that if you're meeting these key results, you've essentially met your big objective. And that's essentially the idea that I want to bring forth today. It might seem irrelevant to scale up using EIS, but if you're a battery developer, manufacturer, or manufacturer of electrochemical systems in general, you understand that there are a lot of complexities associated with scaling up your design. You've got all of these amazing systems and tools in the lab on a bench scale to gather data that would give you the level of visibility that you need to make decisions, but you're making those decisions on such a small cell level. So proof of concept. But now let's industrialize that. Let's move beyond that and let's go for scale. Now you're seeing a lot of different complexities. And the goal, the collective goal around the world is to, you know, double the energy density of batteries.

4:40And we need to meet certain safety. We need to meet certain performance and cost metrics. And we need to do that collectively. And it seems like in the space, a lot of what's happening, which is great, is actually that a lot of innovations are popping up. Materials innovations, design innovations, new ways that you can utilize the battery, new ways that you can actually gather data for insights when the battery is deployed. But one thing that the EU and it seems also quite a lot of organizations like the World Materials Forum actually pinpointed as one of the technical limitations to scale up were effective tools for characterization. So how do I actually understand what the effects of my new materials, my new components, my new designs, my new innovations are on the industrial scale performance of my batteries in a way that matters for scale? And that's the topic that I want to talk about today. Regardless of the technique used, I'm just most familiar with electrochemical impedance spectroscopy because that's all we do at Pulsenix.

6:00But the idea I want to discuss today is developing systems and processes in place to enable scale up. And how do we do that with things like techniques that add visibility or effective characterization throughout the life cycle? But also how do you then take that data and use analyses in order to bridge lab performance, field performance in a way that's effective, fast and cost efficient? Thanks so much. Thanks so much for that, Mariam. Shall we open up the floor for people to join on stage to discuss with us? Yeah, let's do it. I see a hand open. Hand up, sorry. Do I add them? Yep, I've done that. Okay, perfect. Hi, Kaushik. Thanks for joining us. Thank you, Kathleen. Thank you, Mar. Hi, would you like to share some thoughts or comments to the topic? Yes, I found Mar's introduction pretty intriguing. You have the technology, but not having the method or not having an understanding of the method which is needed to scale up is pretty interesting.

7:49You know, it leaves out a whole, it speaks effectively about implementation, which is a key challenge most entrepreneurs or technocrats face. And there are quite a few ways of doing that. Are we doing a deep dive into that? Are we just looking at an overview? How is this structured? Pardon my ignorance. Yeah, I'd love to have a discussion and then do a deep dive. I'd love to have a discussion first on the premise or the idea. I'd love to do a deep dive into how we do that through EIS. But I know there are a lot of different systems in place that we can do that for scale up. Yeah, thanks for that, Mariam. Maybe picking up on that just really quickly, if you're going to share with us a different type of systems that's currently being done so far and give us an overview on that, that would be great. Yeah, definitely. I mean, in a nutshell, when you're in the lab, you've got at your disposal some of the most advanced ways of characterizing your system.

9:03One of them being electrochemical impedance spectroscopy. So typically you'll use a potentio stat and that's just a tool that allows you to perform that kind of technique on your battery cell. You've got a cell in your lab. Usually it's running really low currents, really low voltages. Maybe it's ex situ. So maybe it's not even experiencing any power. And what EIS allows you to do is to extract time-resolved information about the battery. And what that really means is giving you the ability to isolate for contributors of performance loss or degradation modes in an individuated fashion. So finally you can say there's a characteristic to my electrodes and that characteristic is represented by a capacitance. And so what is my capacitance at the surface of my electrode? All right. All right. So now I've literally gone, instead of saying this is my entire response of my system or an average response, I can actually see into individually how each component is relating to performance. Because each component, I've essentially added an AC perturbation to that cell.

10:33And each AC or frequency signal reveals a chemical reaction that's dominant at that frequency. So you'll extract a unique response in a time-resolved fashion. And that really allows you to then see inside your battery cell and make decisions more specifically relating to materials or components that you would use to meet your performance goals. But now once you scale that beyond the cell level, you've met your performance goals at the cell level, you've got a prototype of the cell. Because of the nature of a battery or electrochemical system, any change in the battery design will add incremental changes in different areas or will manifest themselves in different areas of the performance. So maybe you increase the area of the cell. What that does is that it makes some reactions that weren't dominant at the cell level more dominant at the scaled-up level. And so on and so forth. And so if you don't have that same level or a system that allows you to track the same metrics you were tracking in the lab, but as you scale up, it becomes extremely difficult.

11:58It's like you're starting at ground zero to actually build upon that at scale. And that's why there are a lot of periods of trial and error and iterations and long testing times for scale. Thanks so much for that, Mariam. Kaushik, does that answer your question or do you have anything else to add? I understand completely where she's coming from as far as technology is concerned. But that answers one part of the question. Like you've figured out what you know and what you don't know using the AI as interface about the technical performance of the product that you're building. That is one part. Whereas there is a much larger ecosystem which is driven by the economics, the marketplace, and today the rate at which technical products are being cannibalized by superior versions of themselves. Also, the speed at which technology is becoming obsolete is another thing that we need to constantly account for while building a market space and looking at scale. So is that also a relevant part of the question that we're looking at answering tonight?

13:15Or are we looking at only the technological growth in batteries per se? It's definitely part of the equation. And the reason I say that is because at the end of the day, the only reason effective tools of characterization are really valuable is because it allows you to make decisions faster and that allows you to move faster. And that affects economics. So it all comes down, again, to economics and the goals set in order for us to reach our energy goals so that we can actually be economically efficient and service our needs. So it definitely is part of the equation. Do you have some insights into the interplay between design and economics that you'd like to share with the group? Thank you. Yes, I think that's a very critical part of the entire equation. I'm answering the question as to what the economics are going to be and how we're going to resolve those and how we're going to bridge the gap. Because a product is derived out of three requirements.

14:33One is the need. Second is the absence of the product, a particular segment. And third one is when it tags along with requirements which are modernized and therefore there is a need for this product, which is an associate of the other products. For instance, the requirement for batteries and automobiles is changing rapidly. The requirement for storage of renewable energy power devices is changing very rapidly, especially in smaller countries and islands where there is a limited exposure to sunlight or to winds and tidal energy. And how they store their power is changing. So one is defining the need very clearly. And the second one is do we fulfill the purpose commercially? We could have a great product, but if people can't afford it or if it is not priced right. And pricing is, while it is very much cost-driven, while we are inside our unit or within the place of business or where we are thinking where it can be made. But pricing is more a factor of market dynamics and perceptions and what the market can afford to pay for a product that you're building.

15:36You could get a whopping premium or you could be selling at a discount initially, depending on the sort of market that you're targeting. So that's something which you need to watch out for. The third one is, which I would love to believe is most critical, is the price of obsolescence. At what point are we going to cannibalize our own product or at what point is the market going to cannibalize it? Because I have a product which I've spent five years working on, but it's going to be in the marketplace for the next three years based on my current forecast. And after which things are changing, environment norms could change, government requirements could change, budgets could change. And these are things which are forecasted well in advance. And if I can tune a product strategy which is based on these external factors, then I've got something going which gives me direction. I may not be right, but probability is I'm headed in the right direction. So that forms a very critical part of the execution.

16:26To give you a quick background, we have built renewable devices in the last five years. We've been working in R&D and we went to market with our product and our need for funding. We've raised about $4 billion in the last six months for renewable storage devices that we have built in India. We're not dependent on lithium or cobalt or mechanism. That's a huge advantage. But we had to get our pricing right and we were too expensive for the marketplace initially. But we looked at the lifetime of the product and saw that we're going to be around for eight to nine years because policy requirements and demand supply is not going to be changing. Technology is not being obsolete because it's going to take some time for existing tech to catch up with where we are. And we can peter down the prices over a period of time and continue to make money. So we don't need to charge our customers up front for something. We'd rather grow the market base, increase the density of coverage of the number of people who are buying a product, and then recover the investments and get a decent margin.

17:21So that's how things play out. I think looking at that completely will give you a better idea than rather than just building a great product and then wondering what to do with it. So defining need, understanding market, and figuring out pricing. At what point do you really want to convert your thoughts into profits is very, very critical for how you approach them. I definitely agree with you there. And what I'm advocating for actually is developing systems in place in order to enable scale-up. And one of the factors or inputs there is what are my cost constraints? What is the need? What is the market need? And of course, the product roadmap, which obsolescence is a part of. So I love that high-level strategic kind of perspective coming into it. But those would definitely be constraints that would be placed within the system. And once you do have that, now you're developing an innovation. You need systems in place to track performance over time throughout the different life cycles of your development.

18:26And that goes beyond even obsolescence. That goes towards next-generation development. That goes towards tracking the same performance metrics in the field. So imagine having a big database of the same metrics that are really impactful to the performance of your system. And you're measuring that over time. And you're measuring that with different life cycles, with different materials, with different innovation, with different strategies in place. That becomes extremely powerful for both business, but also product development continuity. It makes it so that you've got a base to go off of. And yeah, that's the definitely high-level strategy that makes a lot of sense. And these are the inputs and the constraints that make sense to add to that sort of system. Maybe I can also... Or SMB, how you go. Yeah, I just want to chime in quickly. A big part of the... When we're talking about renewable deployments in general or batteries as energy storage, one thing that the developer has on working out for them is the typical project horizon.

19:43So you're talking about, say, you're deploying an emergency backup system or with any electrochemical process, generally speaking. The lifetime is, you're talking 10 years, 15 years. And so you do have amortization working in your favor. And so the initial cost spread out over those years typically would do have a significant impact on the project's IRR. But what we've seen creeping up mostly in the later years especially is operating costs, maintenance, who carries that, the burden of maintaining these systems and ensuring that you've got component longevity and so forth. And a big disconnect between accelerated lifecycle tests that you do in a lab or in a pilot phase before you've sold your commercial systems, there's just a disconnect between what you see there in doing those accelerated tests, which you have to say get in within a year or two or three or possibly five, like Kovchuk was saying. And the real-life performance once your system is in the field. And a big part of future-proofing and making sure that economics work out is once I put a system out in the field, I want to make sure that I'm promising a performance I can actually deliver.

21:14And not only that, but I'm promising that performance and delivering it at an economic point that makes sense for me as a developer or a deployer of this electrochemical technology, but also for the customer if they're using it. And what we've seen is there is a bit of a dissonance between the tools available to you when you're in your scale-up or piloting phase and what you do on the monitoring side. It almost appears to be that on the monitoring side, it's always an after-the-fact thought. Okay, now we've got the system, now I'm deploying it, let me do some monitoring so at least I know, okay, now a system needs to be replaced or something's going on. So it's more of a binary choice at that point rather than a continuous visibility like you would have in the piloting or scale-up phase. And I think that warrants a bit of a discussion in and of itself. Thank you for sharing this, Asim and Miriam as well.

22:19I'm just curious, kind of as you said, based on what you were saying, right? I mean, EAS, I think, can be used in different ways, right? You just mentioned testing and reward applications. Maybe give a bit of an insight from this week. I was fortunate to attend a design sprint from the German environmental ministry in Germany because they kind of work on a battery passport concept for EV batteries. So the idea is to track all the important parameters of the batteries. The scope is still to be determined. It can go all the way from raw materials, but then also in the production, but then also in the world applications all the way down to potential second life use and recycling. And it's kind of connected because in the European Union, there's some regulation coming up from 2024. You have like labeling, so which includes to say where the materials are coming from. And that's also, you have to essentially publicize your CO2 footprint of your battery of different steps, et cetera.

23:18And then in 2026, there's this kind of battery passport, which I just mentioned, which especially kind of should also include performance data. And the idea is to kind of, you know, kind of bridge these gaps of, like, you know, data access and things like that. So I'm just really curious in your standpoint, you know, looking at the title as well of, you know, scale up using EAS. Like, do you see it mainly on, you know, on the production side or like, you know, accelerated testing, things like this? Or do you actually think about in-application? And then if in-application, you know, how do, you know, how low can you reduce the cost? That's one thing I'm kind of, you know, have heard as an argument to not use EAS a lot in-application is that it's too expensive to have it in all applications, right? Like, could it be scooter batteries, et cetera, cars, et cetera. So I'm just curious kind of where do you see the biggest potential and also how to overcome, like, you know, the cost issue and things like that and complexity.

24:11I think you hit the nail right on the head, Simon. This is, you're mentioning the EU's, there's a similar EU report that was put out by the Hydrogen Council. And they were essentially, they laid out the main three points exactly how you said it, right? You want to be able to understand the performance and the lifecycle assessment, not just of your system, but of every component within your system. How these changes can be, how can they be made to a system in development? And then once in deployment, how do you interact with that system in order to make sure that it's still economic? And the big reason why we're focusing on EAS is the way we view it, is we view it as a process continuation tool. So it is something that most electrochemical researchers in the lab have been exposed to. They understand how to use it. They understand how to interpret the data. And they use it as a part of going from a single cell design to maybe a larger single cell design with components, operating point, and some definition of optimality that fits their application.

25:25Where we're betting our future, essentially, isn't saying, well, if you already had this amazingly powerful tool in the lab, and you already know how to use it, you don't need to train people on it. All you have to do now is say, well, if I got the same tool working in the field, will that be beneficial in terms of addressing some of the concerns and pain points that most of the green technology and the clean tech bodies are saying are some of the major issues and the pitfalls that a lot of these companies see when they're scaling up? And we think the answer is yes. Because you're coming from a point where you have data sets on your stacks, on your cells that you know is where you want to be operating. And once you deploy that in the field, now it becomes a question of remote condition monitoring and control. And if EIS is able to give me that information in the field, I'm no longer relying on average measurements of different parameters, be it temperature, pressure, voltages, currents.

26:39I'm actually extracting components that directly relate to the characteristics of the cell and its multiple components inside of a stack and inside of a cell. And mapping those back to the original data set that I have makes it a lot more powerful in terms of my ability to predict failures, my ability to predict maintenance scheduling, and my ability to optimize my controller. Because a lot of the times we forget that we deploy these electrochemical systems with controllers in the field. And now you have this controller that is essentially waiting for you to give it a command saying, well, I can do whatever you want me to do with these batteries. Just tell me what you'd like me to do. Should I shut off a battery pack? Should I switch the power demand to another one? And these are all things that you can actually derive from an EIA-based analysis and have a controller act on it. And that has a direct impact on the operating cost and on the life cycle or economic life cycle of any electrochemical-based product.

27:46That could be a battery system. That could be an oxidation system. The non-linearity of these systems is pervasive across any and all of the electrochemical plays. So that's where we think EIS could be very helpful. And Simon, if I've also understood correctly, you mentioned EIS could be extremely expensive when used beyond the cell level. And is it because of the tools available to perform EIS at the moment? Yeah, this would be kind of my follow-up question because I agree with you. I mean, as on what you said, but I think I see tension. I mean, you have seen all the researchers from Vorak University, right, to kind of do – I think they published this. And, you know, you can also buy tools not doing it. So you can do state-of-health predictions, you know, on EIS for, you know, application automotive batteries, et cetera. So they did this on this and leave batteries, you know, and they could do a test in three minutes instead of, you know, hours of charge, discharge and things.

28:50So I think there's definitely, like, you know, room for these kind of EIS systems. The question is just, like, you know, dropping in cost. I mean, we have to look at it ourselves. You know, we are developing an open-source battery cycle at the moment. So, you know, one idea was can we put EIS in there? And I think we potentially could at some point. The big question is just, like, you know, what's the price we could get at? Because if you look at a commercial EIS system, right, I mean, it starts with, like, a few thousand bucks or so, right? Like, and I think for many of these in-field applications, it could be too expensive. So I'm just kind of curious, like, you know, where do you see, you know, the price point could go? I mean, are you going to finance it because you're going to have a subscription for software so you can kind of reduce the hardware cost, let's say, because you're going to incorporate it into your software?

29:32Or I'm just kind of curious if you can say anything on where do you see the cost can drop? Yeah, that's another great follow-up question, Simon. And I'll answer this in two different ways. The first one is, let's say you're buying a large, say, high-power potential stat from one of the main vendors. That system could run you anywhere between fully specced out to do stack measurements, maybe $50,000 to $100,000. The question is, when we're talking about price, whether something is expensive or affordable, it has to always be tied to the value that you're extracting out of it. So if I'm paying $100,000 a year, but I'm saving $300,000, then I know what my payback is. And if I'm paying $100,000 and getting back a million, then okay, that's definitely more affordable. The big dissonance that you'll find right now with the existing business model is that you cannot use EIS as a process continuation tool. If you have to deploy 100 units in the field and with every unit you have to spend $100,000, chances are you may not have an economically competitive product.

30:57But also, when you're deploying in the field, you don't really need all the features and all the software and all the techniques that you would want to have be made available to you in the lab. You'd most likely be interested in monitoring one or two or maybe three parameters, and your reporting rate is going to be way lower. Because presumably, by the time you've deployed, you've deployed a system that you have a higher confidence in than the one you're working with in an R&D setup. So what we found, and that's the second part of how I'm going to answer this question, what we found the business model to be the most suitable as a process continuation tool is, again, something that you just, exactly like you were mentioning, it's a hardware as a service. So it's a hardware and software as a service. So it's a monthly recurring fee where the cost per unit in the field drops down to less than $1,000 a unit, almost 80% discount with volume than what you'd get for, say, 10 units in a lab-based application.

32:07And a big reason why we're able to do this is we recognize that your needs are not going to be the same. So we don't have to be giving you the same resolution in the field. We are focused on giving you a system that helps you achieve the goal that you're after, which is typically to monitor a few parameters that you know have a direct relationship to your performance. And so it's all about tying the business model with the technical requirement for the different applications, whether you're in scale-up or whether you're already deployed in the field. Does that make sense? Yeah, that's great. Thank you for sharing. I mean, there's some, I'm curious really, I mean, maybe at some point we can do another chat offline, but I think it would be really interesting to see how you do this price point. I think it's, yeah, it's exciting. So, yeah. Yeah, I loved that point as well. And it can be argued that $1,000 extra is still expensive.

33:12But given how you use it and derive value from it would actually either justify or negate the price point, right? So the value that we found can be extracted for battery systems in the field is related to fire safety, number one. And number two is actually optimization or predictive maintenance visibility. And those two points, at the very least the fire safety point, is actually quite foundational to your product and to deploying it in the field. And why we're proponents of EIS for making those type of decisions is just the resolution that you can gather from this sort of technique. In terms of the resolution you gather from how your system is behaving in an individuated fashion. Each of your components is giving you a signature telling you this is how I'm behaving. This is my threshold. This is my benchmark at the moment. And also giving you a unique signature that will inform its state of health at all points in time. But because it's such an individuated sort of variable that you're getting, your mathematical models and your analysis will be more time resolved.

34:47It will be more predictive in nature. And it will allow you to actually make decisions more reliably. So it's just in the resolution of it. And it allows you to then optimize for certain things that are foundational to your product beyond just optimization of performance. I have a follow-up to that, if I may add to what Mar said and what SM just mentioned. I mean, I think SM has opened up a whole new thought process here, which is pretty significant in terms of mission. As we see them. However, would those be an option of optimizing the cash flow per unit of customer that we have? And are we looking at a definite, for these models, are we looking at something which has got a definite lifetime, which has got a definite turnaround time with the customer? Are we looking at products which are really long-term? And then you have the, because once you place a product out and then over a period of time, the value is going to deteriorate.

36:06It's definitely not going to have a premium unless it's something of that sort of value. And in the renewable space or the battery space, are we looking at something which, you know, could we lose out value by not optimizing it up front or instead of, you know, capital, how do I say? Instead of measuring the capital intensity, we just work on optimizing the cash flows for customers. Do you think we have a chance to lose it out there or are we covered in some way using the model that you have? Thank you. That's a great point. And I think it speaks more to the nature of deploying battery systems or clean tech systems in general, more so than it speaks to how our EIS monitoring would be embedded into that. Typically, when you're deploying these systems, you'll have a long time horizon where you're guaranteed, if you're guaranteeing performance, you're guaranteed your payments for a long period of time. And the question really then becomes is when you're optimizing your cash flow from the perspective of the buyer.

37:20In a lot of cases, they do prefer to have very little cost up front and a higher recurring annual or monthly expenditure. And mainly that's driven by, depending on where your operation is, it's just driven by pure accounting. If you have a large upfront capital investment, the government dictates how fast you can write off the depreciation on it. And so that essentially is deferring your tax benefits for some time. Whereas if you're buying it at a minimal upfront cost, but putting the burden on your recurring operating costs over that, say, 10-year period, then from your customer's perspective, that has a huge tax benefit. Because they can realize that those savings immediately as soon as they're paying them. So this is a big part of just the economics of deploying these systems in general. Now, the question then becomes if you're looking at it from your perspective, in the later years, if you're doing it as some sort of an annuity and you're having a fixed payment over a 10-year period, you've presumably built in those, you know, the time value of money into it so that the payment you're getting in year 8, 9, and 10 are still relevant and meaningful to you today.

38:54However, what will be out of your control is the amount of the maintenance dollar that you're going to be spending on those later years. And unless, which we've never seen, unless somebody spent, say, 10 years in, like, really trying to hone in on how those systems perform over time, you're going into these contracts with an estimate or an idea of how your system is going to perform in the later years. And so the way we view it is that any measure that you can take in order to lessen the chance of you getting a surprise in the later years is an insurance policy on your contract. And there are different ways of how you price those. And what we try and do is we price it so that at the base minimum, you're capturing a percentage of that insurance value for those later years. Because otherwise, that could have the potential of risking your entire 10-year project to be an economic dud, essentially. So for me, sorry, Catherine, please go ahead.

40:05Oh, no, Kaushik, I'm just going to reset the room a little bit. So maybe you want to ask a question first? Thank you. Yeah. It boils down to me for the question of at what point am I going to call the insurance or what point am I going to budget for maintenance? And what we also do, you know, so we do extrapolation, as you know, because we don't have 10-year, 15-year results at the moment. We do accelerated testing. But what we know is at the moment, data point is saying three years, three and a half years before the first round of maintenance kicks in. And then if we don't have in-house maintenance for this, we can tie up with an agency and say that, you know, 42 months, 45 months, you could have your first round of businesses coming in. Therefore, we could tie this up with your discount or cash flows and take it forward from there. And if it doesn't work, then we find a way to call the option.

40:59That's the economics of the thing. What I'm trying to figure out with you is this boils down to two options. One is either you offer the customer a lease option or you ask them to pay up front and they do the financing at their end, especially for large volume, non-government private sector projects where, you know, people are reluctant to put in that kind of money up front. And paper performance is preferred, but they don't have that option. So they opt for financing schemes. But as a supplier, as a manufacturer, especially in the value chain where I believe that I do have value and the customer is going to see that over a period of time. I would like to have my cake and eat it too. And how do I strike a balance between that? Or is that still an open-ended question? Thanks so much, Kaushik. Isan, if I may, I actually have a follow-up question to that. And maybe would it help to answer both at the same time?

41:54Go ahead, Catherine. Yeah, go for it. Okay, great. Thanks. So, yeah, I was just going to build up on that. And I have two questions with regards to the, you know, the question on price and market as well. So I think pretty much, Isan, if you're going for the hot weather service model, it depends on two things. One is the applications that we are looking at. If we're looking at automotors versus scooters, I'm not sure which one do you see more commonly right now. I would imagine automotors, but please let me know your thoughts to that. Because unless the scooters are going for a leasing model, like a battery as a service model, and if they're not ultimately the battery users, predictive maintenance is probably not too much of an interest. But I would love to really hear your thoughts as to, you know, where do you see EIS being used more commonly right now in terms of application? And also whether that's dependent on the company's choice of business model.

42:58Typically, do you see that more in leasing model where the companies are battery owners? And do you see that a lot or do you see that a lot more in companies that sell the whole vehicle together with the battery itself? I would love to hear your thoughts on that. And also maybe one last question. If you could share more on the technology side as to how much more benefit it brings as compared to, you know, using EMS or other types of systems to track performance. That would be great. I know Maram shared a little bit more on that for the previous session where she pitched about one and a half months ago. If we could, for the benefit of the audience here, to share a little bit more on that. That would be great. Thanks. Sure. Great questions, Catherine. So I'll first start by answering the first one relating to the automotive industry in general. If you think about the big two or three players or car manufacturers that deploy electric vehicles, they obviously have their own BMSs, but they also do EIS on their stacks.

44:11And these are EIS systems that are highly customized to their performance, to what they want to get out of it in terms of maintaining their performance and also in terms of collecting data for their next iteration. So as from where we stand, we don't see that as the biggest application for the industrial side of EIS. What we think is an area where you pretty much lack a lot of visibility that you would need in order to economically deploy is in stationary applications. So if you're thinking about energy storage, you're thinking about other electrochemical processes, green hydro production, oxidation, electroplating, electroinning. These are all stationary industrial electrochemical processes, including fuel cells for the automotive, even like long haul trucks and so forth. So this is where we think the big impact really is. Maybe with the, we've seen this, especially in the last year. The only caveat regarding my comment on, on the automotive industry, I would say is, is in long haul trucking. And there are different business models that we've deployed.

45:43And one of the more successful ones that we've seen with some of our customers is the one where they, they are responsible for electro, electrifying an existing fleet. So they're not really manufacturing the truck, but they take an existing truck and retrofit it with their systems in order to electrify it. And, and, and in, in these situations, we've also found the, the industrial level type of EIS to be extremely useful. So that's, that's relating to the EIS on, on, on, you know, automotive versus deploying it on a truck versus deploying it on a scooter. Because at the end of the day, you're deploying it to sit to capture a high of the total value of, of the, of the unit that you're trying to monitor or trying to optimize. And if that unit cost in and of itself is not that high, I mean, if talking about a scooter, then it most likely is not going to be, uh, where you get the biggest bang for your buck.

46:46Now, regarding the second, your second question, which was about, uh, the technology and the impact, uh, the economic, uh, deploying it at the industrial, uh, scale. So, um, the example I like to bring forward is, uh, from a mine that we were, um, uh, in, in Brazil, that's, that's, uh, uh, focused essentially on extracting zinc. Um, so they, uh, they, uh, they have a huge facility and they electro in zinc and a big part of, uh, the mining process is they have to use ANFO, which is just explosive so that they can, uh, create the, the, the, the, the ground and extract some of the, some of the minerals. And, and part of that is you end up with, um, a high nitrogen, uh, contamination in, in the, in the tailings spawns or, or on the soil itself. And so what the customer had was beside it, besides their electro zinc, electro winning plants is they also had another plant where they're electro oxidizing the ammonia that came out of, uh, came out of the ANFO.

47:57And what we were able to show, uh, on their systems is by running on just by retrofitting two of their small scale. And we're talking about a hundred liters an hour flow. It's not very small flow rate. Uh, we were able to show, uh, that we're, that on a, on a monthly basis, just on electricity savings alone, by being able to modulate, uh, the amount of power that's being, uh, demanded by the electrochemical system, you're able to, to, uh, to save over 180,000 us dollars a month. And so, and that's just an example of, uh, being able to identify some of the characteristics and, and, and, and the, and the mechanisms by which the electrochemical cell changes and degrades over time. Um, and, and, and taking a proactive approach to addressing some of these changes, uh, and essentially writing the curve down, making sure that at every single point you're getting the biggest bang for your buck until you no longer do that. Um, and then you, you stop it and, and you perform your maintenance.

49:03So by, by virtue of just writing that, that curve, we were able to save almost $200,000 a month. Thanks so much for that. Isam, uh, we're closing into halfway in, or actually we're past halfway, uh, in, uh, the, the halfway mark of the session. So we have Milo's and Hans have also joined us on stage. Maybe I will, um, give them the chance to speak. Milo's, would you like to, uh, add your comment or thoughts? Yes, uh, probably I would more talk or ask the. Generally, what we are trying to do. We are trying to speed up the process of the commercialization of the. Uh, battery technologies, uh, faster and to make it more clear. So there are many tools, uh, which we can deploy, including the, uh, EIS, uh, the battery, battery is complex, um, machine, electrochemical machine, uh, in a small scale. So, um, uh, we need to bring the clarity to the battery, uh, battery, like the technologies for the components of the battery, but also clarity to the, uh, to the process of the commercialization.

50:31full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full full AI is helping us to do it.

51:04So my question is, what kind of role AI will play in this characterization of the battery components and also speed up the process of the commercialization? And I have a follow-up question that dovetails perfectly into that as well. That's a great question. And that's actually our philosophy, actually, also at Paul Senex. And the way that we approach that is on the premise of let's perform an extremely powerful, non-disruptive characterization technique that is effective for characterizing the state of health of my system and my components as my system is operating. And that gives you so much data. Every five seconds, you're getting a data point on your electrolyte and your electrode health. Now, once you've got that, that's where AI and big data analysis tools or techniques come into play. Because it is extremely difficult to take all these different data points and try to interpret them at an individual level. But once you've got enough data, you can then use big data analysis tools, AI being one of them, to trend and to correlate different variables within cells, within each stack, within your system, while it's in operation.

52:43And that allows you to, A, commercialize much faster, make decisions from a design and operational level, but also allows you to, like Asam had mentioned, utilize this level of resolution to develop models that would allow you to control your system and optimize it for some sort of optimization and goal you're looking for, whether that be energy costs, performance, or whether that be safety. I just want to add a little bit of that because Milos did make a great point regarding the focus is on scale-up. And, you know, AI can be a valuable tool that helps you in that process. The way we view it and a lot of our customers view it is, in the scale-up process, I think the best use of AI is to think of it as an assistant to the decision-maker, the decision-maker being the stack engineer or the chief research scientist responsible for deployment. You don't want to have a black box that tells you, you know, this is what we recommend you be operating at and say in terms of voltages or current density.

53:58You want to know why. You want to know why the AI came to that conclusion or how did it arrive at that recommendation that it's making. And unless you know the reason, we found that most would be a little bit skeptical and hesitant to adopt it without understanding the reason behind something. And so in the system to you and what it can do much, much faster, much more efficiently than a human is in analyzing a huge data set and deriving statistical significance, significances between certain parameters that are of interest to you. And this is exactly the philosophy that we adopted at VolsaniX. So when you're, for example, looking at your stack and you have, say, 100 cells within your stack, it's going to be extremely excruciating time consuming if you have to go measure the cell voltage, for example, the EIS response at every cell, the data. And then if you're, and then you have to understand how each of these individual cells in the form of seven statistically significant mathematical functions that we think someone who's in the space, you understand the characteristics or the operating profile with the response you're getting.

56:10For the deployment or for the scale up to make that decision and reiterate on it. So you're essentially assigning a quantitative measure on every experiment and every step you take along. Yes. So basically, you need to, those data which you are collecting from the technology tools like EIS and many others, you need to plug them somewhere to AI. And AI needs to analyze those data because of a human being, we are not able, we are not capable of analyzing so much data together. But the pros, what we are trying to do, in my opinion, we are trying to speed up the process from the idea to commercialization of the, in that case, battery. So we are, we have now the tools to get available. We just need to, we just need to deploy them and we need combine them together with, I believe, AI, which will enable us to speed up the process of the commercialization, satisfying some ego of some researcher. I'm interested only to invest to the company which has the chance to, yeah, so my question that touches on this topic is really where you all see yourselves at in the stack, in the long term of the value that's being delivered by, I mean, essentially, if I've understood correctly, what you're doing is you're providing, the atomic unit of value that you're providing is just data on a very opaque new piece of technology.

58:26But then when you think about the value that you can extract out of that data, that's where the AI piece comes in. The more data that you have, the less siloed it is, the more valuable it is. I especially appreciated the point that you made earlier about, you know, being able to track supply chain things and how those interact, the battery chemistry and any data that can be captured at the back end, how that is affecting performance long term, but also use cases in the end. And so, you know, really data integration seems like one of the highest value uses of the data where you have valuable data throughout the lifecycle of the electrochemicals, you know, from mine all the way to recycling. And so my question is, from a business model standpoint, are you thinking about things like Airbus, for example, and they, in a sense, are a hardware as a service business. They provide aircraft to a variety of airlines, but they have recently partnered with Palantir to provide Skywise software on top of the aircraft that then does a whole host of diagnostics on every single system in an aircraft that provides value to the airline operator that helps them reduce their maintenance costs, helps them increase their profit margins, reduce downtime, all these things.

1:00:09Are you planning to really try and just focus narrowly on a lower part of the stack where you're providing the data related to maintenance? Or do you have time to move to the stack? Stack and try and provide a more holistic, smart decisions based on the data within their own company, but also incorporate data from across industries, across customers. That's a loaded question. So the thing I want to just point out is remember that EIS has been used as a diagnostic tool. You didn't really need, historically speaking, you didn't need to analyze sets and sets of EIS data because right from the get-go, the data of itself contains value to you. Now, being able to perform EIS in the field, on the stack, or in R&D during your scalar process on the stack, also in and of... is value. So, the second part of your question relating to providing the software and the analysis, we do that. With the caveat of we can... ...somebody else's data for two reasons.

1:02:05One is data privacy issues. And two, in most cases, the applications wouldn't lend themselves to, say, having an application system, for example, to a battery stack, or even from a battery stack operated with certain application parameters to a different battery stack under different operational parameters. And... ...and... ...comers typically are very hesitant when... ...once they know that their data is being used for... ...somebody else's... ... ...so... ...using it with a direct competitor, just from a high-level... ...a high-level view of it. So, the... ...the biggest draw to why you'd want to do EIS in the lab and then in the field is it's providing you with a way to continuously ensure that you're at the operating point that you promised your customer to be at. And the alternatives are quite limited. You... There's so much... ...to you... ...on the R&D side, but the question is, what can you take out of those and... ...and take it to the field and get the same level of resolution so that you're able to provide the same level of performance...

1:03:39...now that you've gone from a controlled environment to a highly uncontrolled environment. ... Sorry, Hansa. Go ahead, Mara, please. I think you also were touching on a great point, which is... ...the... ...towards how you can use EIS beyond just the scale-up process, right? And our thesis is, it would be a great disservice if you don't use EIS for decision-making in the field. And what I mean by that is plan T. So, first being, it could be a manual decision that you're making. So, on monitoring or safety or... ...replacing or resourcing your... ...your deployments. Another being... ...using the EIS data to... ...actually control your system autonomously. And... ...because of the level of resolution and dividuation, you're actually able to do that in a more reliable fashion... ...in a way that will give you a bigger delta, a more meaningful or significant delta result. So, energy efficiency of 13.7% using EIS versus 0.5% using average data points that you... ...that are available today to use, like DC voltages, currents, temperatures, pressures...

1:05:21...and using AI on top of that as a layer to analyze and add controls. So, the delta is quite large. And that's because of the resolution of the EIS data that you can gather. So, if you can couple that with AI and build it into a controller, that is the ultimate optimization strategy. Thanks so much for that, Maren and EIS. Hans, do you have any follow-up questions to add to that? Or does it...does it help with... Oh, that's very helpful. Yep. Thank you. I appreciated both of your answers very much. Thank you, Hans, so much for that question. Great. Thanks so much, everyone. We're about 20 minutes left to the session. So, if you have any questions for Maren and EIS, do feel free to raise your hand. Oh, I see someone. Can we just let you... Hi. I have a question. What do you think about the sodium-ion batteries? Does it have any future in the near term? Like, I don't know, one or two years, for example, CATL is coming up with some technology for sodium-ion batteries.

1:06:55And then, also, I think in UK, there is one company who is also developing. What do you think about that? Yeah. So, we're not aware of any electrification, like, automotives or any even stationary application that uses sodium-ion batteries. So, they were of great interest in the academic fields about 10 years ago, primarily driven by lithium-ion batteries, environmental impact. So, the biggest, I would say, right now, the problem with the sodium-ion battery in terms of, say, the cost per kilowatt hour of capacity is that we haven't seen reliable published data that compares to the kilowatt hour of... ...the cost per kilowatt hour of capacity, but academics seem to think that it should be similar or maybe a little bit above what you typically expect from lithium-ion batteries. It definitely has the benefit of being a safer technology. So, the sodium is abundant, unlike lithium. And theoretically, it does have a good efficiency overall, a federatic efficiency. Having said that, we haven't seen much of it being commercialized, and it's very hard to...

1:08:45A lot of these systems in the R&D phase will show a lot of promise, but then a lot of issues pop up when you're manufacturing larger electrodes and you're putting them in stacks. So, it's too hard to call it, and too early to call it, I would say, is my view. Okay, thanks. Thanks so much for that, Suma. I just wanted to chime in and maybe one last question for Isam and Mariam, if that's okay. Before I do that, I'd like to give a plug to you guys. For anyone in the audience, please feel free to follow Isam and Mariam on their profile. I see that they're relatively new to Clubhouse, so it'd be great if everyone can give them a follow and also check out their website and also their LinkedIn profiles. They've both been great in sharing a lot of insights and information on this topic. So, really thankful for them to do that. Actually, I do see two more people who wish to join on stage right now, so I'm going to let you guys up.

1:09:59Hi, Arjun. Welcome to the stage. Welcome to the stage. Yeah, hi, Catherine. I have just one question to Isam. That when you actually sell this platform as a service, and you're looking at real-time connectivity from the field for large applications, how do you establish the connectivity? Is it through a communication port like RS-485 or something similar? Or what exactly is the communication stack to your server? That's the only question. Thank you. Yeah, thanks, Arjun. So, in the field, if there is a centralized controller present, typically you do the connection over Ethernet IP. So, that would have roughly a 100 megabits per second baud rate. So, it's pretty good. It's actually an overkill. But it's been standardized enough that most controllers will be able to support it. Another option, like you said, RS. So, if you're using a serial connection, that's also an option depending on the amount of data that you have and you're reporting. But typically, we would do it over Ethernet IP in the presence of a centralized controller.

1:11:12Now, the interesting one is if you don't have a centralized controller and the reason why you wouldn't have it is you're trying to reduce your cost. And you've deployed the system with enough confidence in it that you really don't need to be doing much on site. And you're deploying it under a remote condition monitoring. So, you have a centralized controller, but that's not physically close to where your deployment is. It could be at your headquarters or in a different facility that's maybe closer to the deployment, but not quite at the deployment. The other options there you'd have in terms of connectivity is you'd have to rely on either cellular, if there is cellular connectivity, or satellite communication. We've utilized a company called Kepler. I don't know if you've heard of them, but the way they establish connectivity is using nanosatellites that thousands of them that are orbiting the planet. And it's actually a cost-effective way of doing the communication. So, you're going to be able to do that.

1:12:15Thank you. Thank you. Thanks so much for that, Arjun. Rahu, would you go next? Thank you, Catherine. I appreciate the opportunity to connect. I come to you to the forum from India. I have the opportunity to connect with Simon offline. We are an e-based recycling firm doing e-based recycling for more than a decade now. One of the recent additions that we have done to our sort of processes is that we have added multi-metal recycling as well as lithium battery recycling to our sort of repertoire. We carry a very specific hub and spoke model, wherein the hubs are centered around India for us. We have around 10 working spokes. And the two hubs that we have are based within the Delhi NCR region. I appreciate the opportunity being provided here by the forum and be glad to connect with essentially all of you, whosoever is interested in battery recycling and wishes to connect basically from a battery repurposing and recycling perspective. Thank you. Thanks, Raul. Kashyyyk or Milo, do you have something on this one?

1:13:46Otherwise, I have another quick one for you. Ask Faris, Simon. No problem. All right, Raul. I'd be happy to connect. We're doing a lot of work on structuring e-based in the islands and it would be great to understand your views offline. So I'd be happy to connect. Go ahead, Simon. Thank you, Kashyyyk. I appreciate the support extended. I had like one question just to ask you about the question. I would like to ask you about the question. I would like to ask you about the question. I'd like one question just maybe also probably going to the end now. But I'm just still trying to understand kind of, you know, what, because you mentioned many really exciting use cases today of EAS, you know, different applications and things. I'm just trying to understand like, you know, what scale because, you know, it's also in the title, right? Like of scale, how to scale and I think also how to apply that scale. So because you mentioned your system, right, you can maybe go below $1,000, you know, dollars, which is really exciting.

1:14:46And then I'm just going to say, so what's like, do you think is the minimum battery size where it kind of makes sense? And if you have a system, like how many batteries can, like, you know, what kind of capacity can you like, you know, do with one system? Let's say if you want impedance measurement, is it like, you know, one pack? Are we talking like, you know, for kilowatt hours? Are we talking about megawatt hours? I'm just really curious to understand a bit better what kind of sizes, you know, if you think about scale, et cetera, what's like the minimum, what's the maximum, et cetera, and how many systems do you need to make this happen? Great question, Simon. So in terms of the spec of our hardware, we have a 5 kilowatt unit and a 20 kilowatt unit, and you can stack them in any way you want. So there's no limit to what, say you're going from a basic unit to like a kilowatt unit to a megawatt hour.

1:15:44It's just a number of, it's a matter of scaling up the number of units that you'd have in terms of deployment. In terms of the cutoff, I think it very much depends on the application and what is it that you're trying to get out of the system. So if it's monitoring that it will, then it, and if you're deploying, you know, a piece of hardware with every 20 kilowatt unit, then it becomes the unit economics degree, 20 kilowatt unit, 20 kilowatt hour unit, and going at it from that perspective. For the case in Brazil, this was a 450 kilowatt hour system, and it was stacked by having one big bath where the tailing spawns water will get in and then be split amongst different sub-baths and each with its own electrolyzer unit. And so scaling up the electrochemical system is, if the company is doing it right, it should be highly modular. And we just attach onto that with the same level of modularity. And Simon, I do agree, the title is scale-up using EIS.

1:17:07So maybe allow me to walk you through a use case for scale-up. So you've got a battery cell, you've been doing EIS using a potential stat, you understand the performance limitations, the mechanisms of your cell, and you've got a performance capacity that you've hit with your cell level. Now, once you venture up, the potential stat loses its ability to service you beyond that because of two reasons. First, it's technical limitations that allow you to run industrial power levels and perform EIS on your cell or stack at those levels. And two, the cost. So again, you've got your bigger systems and being able to perform EIS on the smaller power levels and then try to extrapolate or try to stack. That is quite cost inefficient. So how we do it, at least at Falsanix, is our hardware removes those technical limitations. So you're able to perform EIS at industrial power levels. So think higher current, higher voltages, stack level. Now you've got your cell. Let's change the design so that we're changing parameters that would allow us to scale.

1:18:31So I'm changing the size of my cell. I'm changing the operating profile or conditions. And I'm stacking them together. Now, as a researcher scaling up, what I need to track are the same parameters across the life cycle. That's the thesis that we're coming in with. And the three most important parameters really that you care about for scale up or should care about is what's the resistance of my electrolyte over time? How is my electrode changing over time? So that's represented by the capacitance of the interface of my electrode. And third is how much power consumption are each of my cells within my stack consuming over time? And if I'm able to track that in situ, so in real time, as my system is operating, gather those data parameters and then use statistical tools to correlate each of these variables to each other and to themselves, then I'm able to actually extract or isolate for the sources of performance loss within my system. And then also track the effects of using new materials or using new components or using different operating profiles on the actual performance of my system at scale.

1:19:54So add those power levels that I would want to deploy in the field. Before you've ever gone into the field, you can use that in order to ensure reliability. As in, I know my numbers like the back of my hand because I've got visibility over six months, five seconds, second intervals. And I have big data analysis tools to allow me to trend that over time. So that's just using EIS for scale, how we would use it or as a use case for scaling up from cell to stack, from stack to bigger stack, from bigger stack to reactor level and using the same metrics and the same tools across all of these life cycles. And then it's a choice whether you want to take that into the field with you, if it's cost beneficial for you or not, or if you just want to correlate these individual data points to the measurements you actually do take in the field and then use that as your metric.

1:21:00Fantastic. Thank you very much. Thank you very much. Thank you very much. Thank you very much. Thank you very much. Great. Thank you so much. I really appreciate my time for today. It's great to have a deeper insight and dive into what you guys do and also the technology in general. It's very interesting as well because from a battery pack manufacturer perspective, to be able to use something like EIS to convince our customers in terms of longevity and performance of our batteries. It's always something that is of interest, I believe, for manufacturers. So thank you so much for sharing on that. And, yeah. Yeah. To finish the session, I'll give you a final plug to Battery Den. Especially startup founder, particularly startup founder, or if you know anyone in this space or any female founders in this space, especially, we would really love to encourage them to join us, you know, to give the chance and the platform to encourage them to network with our panelists.

1:23:57We have a very established group of panelists and they all bring in tech expertise and also from the investment side as well and to give you some feedback and also potential network in the battery space. So they are all knowledgeable in the space and you know you will not be speaking to your investors who do not know anything about batteries. I really encourage anyone in the battery space or if you do know any friends who are in this space to join us. So the sign up is really simple just follow myself or Simon on LinkedIn and we can send you the sign up form link and we'll probably finalize all the applications by next week. So really looking forward to you know receiving your applications as well. And without further ado and with that maybe we'll end up the session. So please everyone give Mariam and Isam a big clap and a shout out and also follow them on LinkedIn at Clubhouse. Thank you guys for your time early morning on Saturday.