Digging Deeper with Data with KP Kairies (CEO @ Accure)

And small changes in the way you treat your battery today can have a massive impact on what's happening in ten to fifteen years what's what's going on in the last ten years of battery cell chemistry?

And what's coming? Oh. And then boom. They just threw all of their experience into scaling it up and that was massive success.

What's what's the other curve ball that's coming down the way from a battery cell chemistry perspective. It's really the one technology that I believe has everything it takes to be wildly successful. Accelerating emerging technology like battery cells, the the control and the power is still in the in the manufacturers hand. Are you saying that you guys can change that using data?

Conturing enough? My brain's broken.

Hello everybody and welcome back to another episode of Moto, the podcast. Today, Quentin is talking to Dr. K. P.

Kyries. CEO of Accure battery intelligence. Over the conversation, they discuss the ins and outs of different cell technologies and how using data can help improve the lifetime and performance. Of a battery.

As always, if you are enjoying the podcast, please hit like and subscribe. It really means the world to us. Let's jump in.

K p. Do I talk about a cure then? How long you've been doing a cure?

For a couple of years. Officially, it's three inefficiently around five, I guess. What what's the idea? Why why did you start a cure?

You know, my background's in electrical engineering. I've been doing batteries for over fifteen years now. And in my last job, I was leading one of the largest research groups on stationary battery systems in the world. Maybe maybe even the largest one.

For a few years, And I was really frustrated that no one's using battery data in a meaningful way. And I was always hoping that some company with Naga Nadori and say, hey, university folks, do you wanna do a research project with us to get us on track? And no one knocked. And then I said, okay.

Seems like we have to do it ourselves. And you're a you're you're you're a doctor. Right? You you've got a PhD in something technical.

What's that? That's a battery related. Right? Exactly. Something technical. It's yeah. No. It's yes. It's it's a PhD on on lithium ion batteries.

For stationary use cases. And I in my PhD, I I I wrote like It's not the most technical dissertation that has ever been, but it actually was read quite a bit. It was like a manual, a one zero one, how batteries are gonna change everything about your business for grid operators and small utilities in Germany. And it turns out they already are.

They already are and it's still although batteries and stationary batteries have been around for a decade, back back in my uni days when I was at the lab leading that group, we were building a five megawatt hour system back in twenty twelve. You know, plan, build, whole development process, prequalification, operation, three lawsuits with the suppliers because like today, nothing worked. When you're building big batteries, so we've gone through the entire process. Yeah.

It's been a lot of fun and Interestingly, even though this has been around for for a decade, a lot of energy companies still don't really undue stamp what's going on and are not very well prepared for, you know, what's yet to come. I'm gonna set the scene here a little bit. We've done each other for a little while, and Acure and Moto, actually are quite similar, but do very different things. There's a venn diagram, where with two circles, which are probably touching now. Not quite overlapping, but there might be an overlap at some point in the future. And what we do at Moto is we look at how batches are performing in the markets, And what you guys do really, really well, is looking at how batteries are performing in the real world, the physical world. Do you wanna just talk for a little bit about what a cure does?

And what the journey's been like since you started this crazy adventure.

So if you think about lithium ion batteries, there's one really unique thing about them compared to all other energy assets like wind farms or solar farms or whatever, every lithium ion battery is basically an IoT device. Because it has to small chip the BMS that continuously measures voltage current temperature. Just to make sure you don't overcharge it, and push it into a thermal runaway or that you don't use it outside of the specified temperatures. So it's a really simple logic chip making sure that you don't abuse the battery.

And kind of as a waste product of that process, data is being generated voltage current temperature. And at a cure, we've built a platform that ingests data from millions and millions of batteries around the world.

Operational data. So really a time stamp for each day, thousands and thousands of data points. And then uses sophisticated physics based models, machine learning, artificial intelligence, this whole jazz to understand what's actually going on with a battery and to predict how it's gonna perform in the future. And the three categories that we offer is safety.

So is there a physical loss imminent? Is there thermal runaway on the horizon, performance, will the battery be able to make the money? You hope it'll make or that you're planning. It will we can talk about, you know, what's keeping it from from doing that in a minute, and the third one is lifetime.

You know, when you plan a project over fifteen years, it makes a huge difference if your battery will last for twelve or seventeen years.

And small changes in the way you treat your battery today can have a massive impact on what's happening in ten to fifteen years. And although that's typically not the most urgent problem for companies today because have so many other, you know, much more urgent problems. It's something that carries a lot of value for or companies that think long term that not just build a battery to flip it after six months but that really want to own the acid for maybe fifteen years. It's the the play here, if I got it right, is if you own or operate big assets, whether that's in electric vehicles, buses, stationary storage, whatever.

If you rather than being a black box, if you fully understand what's happening within them, and you're getting data, and you can understand that, you can there's loads of benefits, like performance enhancements. You can plan your maintenance. You can make sure you get the right up time. You got safety improvements.

You can maybe extend a few years, your lifetime of these assets. But it's all about collecting the data to start with. Mhmm. And that's what you guys do.

How how do you collect the data? Where does where does the data come from?

Standard engineering answer is it depends for large scale battery storage systems. It could either be a, like, all in one supplier like the fluance and powers of the world that offer hardware EPC ing and also have a data platform that we can connect with, or you might have a separate EMS provider sitting on top of a, you know, standard commodity asset And there's also, I think, a pretty confusing soft stack on top of battery assets that comes with a digital platform that also is a marketing service or that's only a digital layer. And so there's a few companies also here in the UK providing that part.

And so We're typically interacting with the EMS or digital layer What's the email? Sorry. Oh, it's the energy management system. So the part of the system that makes sure that the sixty containers on a site work in unison and that you have that the signals from the trader are routed to the right containers.

So if we think about big scale, utility scale storage, they they might have a load of containers in a field somewhere, forty fifty containers that say it's a fifty megawatt site. And on that site, there'll be one big computer somewhere, and that will be the site control, and I'll be figuring out how to you know, how do you balance between containers and all the complicated stuff. And you guys integrate with that thing and pull the data back to your servers and then do the analysis on that. Yeah.

And so we don't have to tap into the computer on-site because all this data is already online. There's very, very few exceptions where it's it takes work to actually get the data because everyone knows that this data is valuable although almost no one using it yet. So it's typically in a pretty easily accessible place. And then we set up an API, some, you know, cloud to cloud connection, pull the data, maybe once per day, maybe more often depending on the use case.

Run, you know, our models, and then we push back the information through dashboards, or through direct integration with a customer so they can, you know, close the loop. Could you just name drop some logos that you guys are working with? What companies? So we can have some idea of the scale.

You you can brag. Right? What sort of companies are you guys working with? What kind of fleets of batteries or that's for some scale and some color around what you guys are working on.

Yeah. Sure. So, unfortunately, some of the most exciting customers I'm not allowed to mention here, a lot of our customers say we're part of their competitive advantage and they don't want their competitors to know how much better they could be. That is a good line.

We need No. It's it's it's really convenient and mysterious.

But but I I I can name a few references.

And I mean, you know, all in all, As of today, we're mid July twenty twenty three. We're managing about, I think, two point six gigawatt hours of battery assets and by the end of the year, it'll be four to five gigawatt hours. And I know that you're in taxes right now.

You are right.

At least in the US. No. I am. You can see me well, that's I'm wearing I've gone to wearing, basically, oversized everything.

Yeah. And so well, t shirts are like triple x l now, so I can somehow breathe in the heat.

Yeah. I'm in Texas right now. I had a bagel this morning. That's how American I've become. It's good it's good stuff. Yeah. And so, you know, you know this better than me, actually, the the Ercotte market for batteries is crazy.

And we're extremely successful there, for example. So by the end of the year, we will have about a gigawatt hour, probably more of battery assets that we support only in Ercot.

And then there's the other American states, and obviously the UK, it's pretty strong for us. So one company, for example, that I'm allowed to mention is Total Energy, the French, you know, energy giant. We have a few more but also, you know, you mentioned commercial vehicles to Boeing Transport Authority to New York, transit authority with their electric buses. They trust our our services for safeguarding their electric buses.

The world market leader for electric ships. A company you've never heard of, but it's they're actually pretty cool from nowhere they're cold. I might have heard of them. They're called Corvus Energy.

Correct. I haven't heard of them. And they're, you know, they they're competing with Siemens and ABB and they're winning. And it's like two hundred people.

So it's it's really impressive. What's the split then? Like, how how much of your you could do it by customer or you could do it per megawatt or megawatt? Yeah.

How how much of the revenue that you guys generated comes from safety storage versus EVs versus others. Because it must be I'm thinking from from a company building perspective, They're very different customer profiles, and it must be quite tempting. Because your addressable market of battery stuff is so big. It's like, basically, the whole energy transition.

It must be quite tempting to do products for everybody. The issue with that is, as a business, you're gonna lose focus. Right? And everyone's gonna absolutely, you know, you've been off building something for one or two customers who feel really special rather than everybody.

The issue with that is, as a business, you're gonna lose focus. Right? And everyone's gonna Absolutely. You've been off building something for one or two customers who feel really special rather than everybody.

How how do you manage that? Yeah. No. You're absolutely right. And, I mean, The reason why a company like a cure can exist, the only reason is we have to be better than any in house solution a company could have and we have to actually be cheaper.

And we can be cheaper because we're so highly specialized that we're more efficient and that gives us some competitive advantage on that side. To your question, we have, for the last twelve months, had an extreme focus on large scale storage assets. Because from a business case perspective for us, It's the easiest way. It's a no brainer.

Right? If you look at Ercot, for example, forty percent of annual revenues occur in the best thirty days of a year, give or take. And so if you're batching That sounds like some tasty analysis. It sounds sounds like it came from Moto.

It absolutely is. So everyone I mean, this is these are all customers already, but, yes, I I highly appreciate what what q and his team are doing here because, you know, we can take your analyses, go to a customer and say, look, what's your average error on state of charge estimation. You know, all these new LFP batteries coming in and everyone knows they're shitty at estimating how full they are. So if you have a fifteen percentage point error which we see all the time on the best day of the year, which has a revenue potential of seven to ten x of an average day. If you have ten percent error on that day, it's like a full day of being offline any other day of the year. And so although ten percent right? It doesn't sound that much.

But if it occurs on the wrong day, you can lose, I don't know, fifty, sixty thousand dollars in a single day.

And So you can basically get a full payback on our services in one day of the year, and then you have three sixty four other days where we're also doing good stuff that's basically free then.

So can we do a crash course then? You Of course. What's what's going on in the last ten years of battery cell chemistry and what's coming? I'm gonna put some some some threads into the conversation.

So there's a discussion between different chemistries from NMC and LFP, for example, what what's the key difference there on a high level? And the buying behavior has changed. You know, there's a lot of LFP systems being built now for storage -- Yeah. -- is that because the factories are have got slots?

Or is it because it's a better solution? And then what does the future look like? There's a lot of talk of solid state, sodium batteries, lots of other exotic materials.

What is the state of the market right now, please?

The state, it's it's very confusing for everyone.

I'm I'm trying to shed some light on a few things that that you said. So n m c l f p, what what does that mean? N m c nickel manganese cobalt, l f p, lithium iron phosphate. These are the cathode materials in lithium ion batteries.

If you have an n m c cell or an l f p cell, They share all properties except for the cathode material. So the separator is the same, the electrolyte is the same, the anode is the same. The form factor can be the same. So it's really like a drop in technology.

It's just a cathode material that changes. NMC was the technology of choice. And if you'd asked or people have asked that question five years ago when you asked experts, so called experts five years ago, which technology is gonna make it.

Three quarters of them said NMC is gonna be the only one, the only technology that takes the whole market because there's these synergies with e mobility, You can put NMC in cars. It has a high energy density, so it gives you a lot of mileage.

And because of the economy of scale, It's gonna be everywhere.

And that was a massive error. Right? Because energy storage systems, batteries were so successful that even stationary storage had its own economies of scale. Right? Econom of scale come into a saturation some point. And if you have enough volume in the market, you can scale several technologies at the same time and profit from these synergies.

And so LFP is a cheaper material. It doesn't need nickel. It doesn't need cobalt. So from a raw material perspective, it's cheaper. There's a few esoteric arguments about which technology is better for long lifetime or fast charging.

And the honest answer is there's maybe five hundred factors that have a huge impact on these topics and the catholic material is one of them. So just saying, oh, LFP is a long live material, that's bullshit.

In the right setup, it can be a long live material. Yes. But in the wrong setup, it can be just a shitty battery. So it's not the only point that's important.

Now the one thing that's clear is that LFP has a lower energy density, which means for the same amount of storage capacity, you need more volume, you have more weight.

In stationary storage, it doesn't matter a lot. Typically, floor space is not the main concern on an island like the UK, sometimes it is, but generally it's not top of everyone's mind. And therefore, LFP being a little bit cheaper and having drawbacks that don't matter so much, very quickly has become the technology of choice for stationary batteries. And this year we've onboarded fifteen to twenty large scale storage sites and Maybe one of them had NMC batteries still.

Maybe two, but all the new ones, all the big ones, it's all It's all LFP, and I think it's gonna stay that way for at least another two or three years. The narrative around that from what I've seen is that the supply chain issues from the pandemic meant that it was really tricky to get slots in production lines for LMC batteries.

And so the tier one suppliers, you know, Fluence and other other guys went out and said, okay. We'll we'll we'll get some slots for LFP back And before they knew it, they'd built the engineering around the LFP batteries, and that ended up being the battery of choice now because of the cost benefit. And when you talk when you're talking about supply chains, I mean, when you can design out the cobalt requirement, for example, that is just something that should be. I know it's a complicated question, but should be a good thing.

So how true is the fact that this was supply chain push to move from NMC to LFP? Or was it really that actually, LFP is a better solution from a cost basis anyway? So there's a few factors playing in it. So from a raw material perspective, I would say that most decision makers in these industry don't have ethics regarding their supply chain as their top concern.

Unfortunately, but I think that's just true. However, the volatility of cobalt and nickel prices are a big pain for everyone. You want to be you want to have predictable prices.

And by taking out, you know, the the three main volatile components of batteries are cobalt nickel and lithium.

And so NMC batteries have all three of them. LFP only has the lithium. And so that makes things more predictable. That's in itself independent of the absolute price, a more a better predictability is good. Then it also has a lower raw material cost, and l of p batteries have slightly lower requirements for production.

Facility. So you don't need a perfect clean room. It's like one level below that's still okay for LFP. And so it's much easier and faster to scale up not much, but it it it's qualitatively easier and faster to scale up LFP production plus China Chinese manufacturers have a lot of experience with that technology and recent improvements in some details of it have made it competitive.

And now you have like the step function from LFP is kind of like NMC's ugly cousin.

But then there were a few improvements and all of a sudden it was good enough for most people. And then boom, they just threw all of their experience into scaling it up and that was massive success. So what so what happens next then? The world got caught out a little bit.

Although, I certainly did, if you've asked me in twenty eighteen, same as you, what was gonna win out, it would be NMC. Biggest car manufacturers were betting on NMC. Pretty much all the station storage was being designed in that way. And for maturing technology, it felt like the safer option.

Right? Because people knew how to deal with it -- Mhmm. -- to some extent.

But LFP courts out. So so what what what what's what's the other curve ball that's coming down the way from a battery cell chemistry perspective? So there's a few contenders out and I think I would put my money only on one.

So I don't really believe in fast or in dramatic breakthroughs when it comes to flow batteries. Maybe I just haven't seen the one that's really promising, but I've seen ten years' worth of announcements that never materialized, so I'm a little bit cautious when it comes to redox flow batteries. Sodium mine is really the one technology that I believe has everything it takes to be wildly successful It follows the logic going from NMC to LFP and from LFP to sodium ion. Lower raw material costs slightly or significantly lower energy density.

However, knowing where LFP was coming from and how they, you know, once it became sexy, it improved if we can assume a similar learning curve for sodium ion and knowing that it's backed by some of the most relevant battery companies on the planet, you know, CATL among others. If they identified sodium ion to be strategic value for them, they're gonna make it successful.

It's not a question of if but when.

And so I'm pretty confident that sodium iron will play a small role in the next two to three years and a major role in the next five to ten years. What is it? I know that the words of sodium iron are less sexy than the others. Right?

Sodium could be one of the least sexy words to use in this context, I think. So what what is it? And I know that the density is lower, so you need a lot more space. There's some thermal characteristics which are quite different.

So so I'd like to explain what it is and then how you treat these batteries differently.

So the main benefit of sodium ion batteries is they don't need lithium.

So lithium supply has become a bottleneck. It hasn't stopped the battery industry from growing and it's not as dire as some people claim, there's enough lithium and with recycling, there's you know, we're not even close to the end of the curve yet. However, I mean sodium is like in its truest sense abundantly available. Right? It's It's basically a table salt.

Building a battery technology based on that is extremely attractive from a raw material perspective.

It's similar enough to a lithium ion battery that a lot of the experience we've made with producing and operating these systems can be transferred.

It's not a drop in technology, so the gap from l f p to sodium ion is larger than the one from n m c to l f p. So we're gonna have to make more changes to the production process, more changes to the way we operate these batteries and so on.

But again, since the battery industry these days already is so massive and there's so much more to come it's pretty easy to mobilize funding for the next level technology.

I think what sodium ion has to offer is a very reasonable package saying, hey, lower energy density but also probably lower safety risk at lower costs, for stationary storage and maybe even for entry level cars, it could be a pretty attractive option. And it's it's the form the form factor to to the same. So if you think about the difference between battery cells and flow batteries, they share the word battery, but pretty much everything about them is different. Right?

You've got -- That's right. -- liquid moving around and tanks and pumps and and pressure. And and then cells are these smaller things that we all know, basically, big duracell batteries, but in different shapes. Yeah.

What about the the this new type of battery, sodium ion, and where does that fit in? It's very, very similar to the lithium ion batteries as we know them today.

Interestingly the form factor After ten years of massive adoption in automotive and stationery storage in power tools like lithium ion batteries, they're used everywhere.

Until today, manufacturers still can't agree on what's the best form factor.

So and the same manufacturers over the past five years have reengineered the entire tech stack saying, oh, we did a drum the first time, but now we're using cylindrical cells because we finally found that they're the answer. And then three years later, they scrap it again and start building prismatic cells. It's Yeah. Just -- -- ongoing.

-- clarify that for anyone who's listening. So you get a couple of types. You get cylindrical cells, which look like big duracell AA's. They are cylinders, and there's some benefits in that big that there aren't any right angles.

So when you're manufacturing, you could roll things, and rolling things from a manufacturing perspective, has got lots of advantages. Also, in electrical systems, right angles aren't nice. We don't like right angles, they tend to cause problems.

However, there is another type of cell which is called prismatic, which is essentially think about a half an a four sheet of paper, and a whole bundle of them put together. It's about that size. It looks a bit like a car battery, but a thin one. You put lots of them together.

And they they they vary in size and quantity per rack and all those things. And there are benefits to that too from a there's some safety features you can put in and some other manufacturing manufacturing differences. And I guess the point here is not one of those is winning. Right?

If you are just arguing to get as much material into one space like you do in a car, then square ish kind of shapes make sense. But if you really, really care about the other things, then the cylinders make sense.

And for stationary batteries, am I right in thinking they're pretty much all prismatic square ish shape.

Yeah. At least, what we see in the last two to three years, two massive trends that we've seen is the chemistry went from NMC and sometimes even NCA, so the technology that Tesla had been leading for many years in their cars sometimes was also used in stationary.

Panasonic was the main supplier of that. But so the typical NMC cell that is still installed in hundreds of large scale storage system was the LG ES, JH three, JP three cells, they've been sold hundreds of millions of times. Also have major recalls with General Motors, the APS fire in twenty nineteen, Yeah. We could try to just talk about that.

That's an interesting thing that happened, and you will know a lot about it. So LG sold a ton of these cells They're installed everywhere. There's lots of there's lots of systems across Europe, utility scale that had these cells installed. And they're prismatic ones, weren't they?

Pouch out. Sorry. Yeah. So pouch is in a soft pack without a rigid container. It's basically like piece of plastic and then it's vacuumed.

And they had some issues. Right? And it cost LG a lot of money. So so what happened there?

Yeah.

So several things happened, but to put it short is LGBT had some systematic production errors that they were not aware of at the time and they only found out after there were a number of fires in different applications. So among others, there was probably everyone in the large scale storage space knows the incident at the Arizona at a p s in twenty nineteen, that was a fluent system. And it's so well known because unfortunately there was first responders that were hurt during the response And so that has triggered a lot of changes in codes and in how first responders work with batteries. Which was overdue but still it's a very, very unfortunate event. And then there was the major recall of Chevy bolts with LG Also with Hyundai Kona and Kia, Hyundai and Kia both use the same battery packs. And so I think there were recalls just in automotive for over three billion dollars.

Just think about that. That's three billion dollars that you didn't expect to spend on something. To be a fly on the wall in that boardroom and be be that poor person who has to stand up and say, we're gonna find three billion dollars from somewhere. Absolutely nightmare.

And that person would say, hey. Compared to all other options, this is our best one, by the way. Yeah. Or I remember their share price drops in, like, twenty five percent.

But, I mean, the yeah. The interesting thing it's from I wasn't an asset owner. I spoke to lots customers who were, but across the board, it felt like they did the right thing. They were transparent about it.

General Motors also were transparent about it. And whether they did the the the recalls. And then the UK systems, I I don't think from a UK perspective. But the grid scale systems out there that had issues Lg came and swapped them all.

So people who built big grid scale assets found that three years into their lifetime of doing frequency response, LG was gonna come into a place or they're selling. It's actually pretty good. Right? It means you get brand new sales.

So hats off to LG for doing that. Absolutely. And, you know, LG really to their credit for many years has been the leading force in pushing battery technology forward.

So right now, I would say most of the innovations or the most relevant innovations are coming from China, from c a t l, BYD and others, but for the largest part of the two thousand tens, it was really LG and Samsung who were leading the industry with their technologies.

And also to put it into perspective, there were maybe a few dozen incidents compared to hundreds of millions of cells that they sold over the same amount of time and from an engineering perspective, they're impressive still. But of course, it's a danger to the industry if we have more of these incidents as an emerging industry it can get you in a bad situation really fast and so transparency and productivity is really of the essence here. Hundred percent. Could you could you help us with a little crib sheet of which manufacturers are doing what? So we know we know CATL are going hard at LFP, and they're also investing big in Sodymian.

What what what groups are the other manufacturers in? And is Is it broken down by China, Korea, America, Europe, or is it more complicated than that? I I would say you could still break it down by country in terms of country where the company comes from, not what it produced. Because the production facilities are becoming international.

For example, LG has a huge facility in Poland. There's in Hungary, there's being production facilities in Europe being built right now.

CATL has a factory in Germany.

And from a technology perspective, So CATL has a world market share of over thirty percent now, which is insane. Yes. It might be thirty five or even slightly above that by now.

So, you know, since we're it in a setting here where most listeners will probably be interested in large scale storage, probably fifty percent of the offers you'll get when you are looking for a new big battery, fifty percent of that will be CATL batteries. They might be integrated by CATL themselves in a DC container, or they might be in a Sunbro system, or they might be in the Canadian solar system, or in whatever the integrator is called system, but it's always the same as battery cell, which is really interesting.

BYD is the same. And then you have up and coming companies from China, Heathium, Goshen, EV, as well, they're all not up and coming. They're massive companies with a decade of experience, but for Europeans, they sound new. Because they're entering the European market just now.

I can't believe that. CAT have thirty, thirty five percent of global. Yeah. I mean, what an incredible business success for CATL.

And also, it was only a few years ago that it looked like BYD was was was miles ahead. And not far behind that, Samsung gonna something SDI and LG chem were saying they're gonna win the market. It's good about the LG thing actually. It's pretty funny about twenty twenty or twenty twenty one.

I remember being at one of the big shows in Europe, Munich, or something like that. And LG had just had the recall, and half of Samsung's stand was about safety. And, you know, all the pamphlets are about safety. And LG and Samsung were just going so hard at each other.

And it's almost like they missed the Chinese thing.

It did. So They missed LFP.

They missed LFP.

Okay. I wanna talk talk talk more about the data. So you guys get loads and loads of data from lots of battery do. And you guys probably see stuff that nobody else sees.

So can I ask you to share some little tidbits of knowledge you've picked up ideally really saucy bits of knowledge that only you know? Yeah. Happy to. You're also, you know, a startup CEO.

So as a young company, I would say it's more difficult to get the attention of customers, of potential customers. You have to provide a hundred fifty percent to be taken seriously. And so for us, what has worked incredibly well is We pitch to customers and say, look, here are ten examples of things that we found with companies like you. However, if you give us the chance to look at a piece of your data in a limited pilot project, give us six months worth of data, we can all but guarantee that we're gonna find something that will surprise you that you never thought possible and you will need to change immediately.

And in one hundred percent of the cases truly, that we started working with a company and they shared some data with us, we were able to identify stuff that they didn't know about but that was extremely relevant to them. And so this is the way that we get customers. We showed them something that they completely didn't expect and they're like, oh shit, gotta do something about it. It cured, please help.

So, yeah, that that that's the way how we operate. And so what what are some of the things that we found? Just before you get there though -- Yeah. -- you're you're a startup, but you're almost grown up now.

Right?

Oh, yeah. You got, like, a hundred people ish. You operating across lots of different contents.

You're almost a big company.

Yeah. Yeah. Yeah.

Yes.

Actually, I don't like the word startup that much because it doesn't communicate credibility or trustworthiness, and these are the two things that we actually sell.

We sell peace of mind and we sell being experts available at hand when really difficult situations come up and a start up might be too you know, cloudy. No. But we're about seventy people now and we operate in Europe and we have an ink in the US. But actually, we have customers on six continents, so it's pretty broad already.

But, yeah, I jumped in. Sorry. Carrie, you're gonna you're gonna share all of the the golden nuggets. Well, all of them we need another few hours, but it's maybe like a best off.

And actually, anyone who's interested, we recently published a few examples in energy storage news. So you can just Google it if you want a little bit more detail later on.

We've been doing quite a bit of commissioning support with companies.

And you know, the commissioning phase is so crucial for all parties involved because the EPC will try to get the project over the finish line because then they get the biggest chunk of payment.

Not only then. And so they're highly incentivized to get site acceptance no matter how And the owner operator is also highly incentivized because only when the battery's fully commissioned, they can start making money.

However, in the process of commissioning and construction and all these things, a shocking amount of things happened, whether it's from battery cell quality that was delivered, that's not to specification, BMS, errors, imbalancing, and all these things And so we've been doing a lot of work with companies during that phase.

So basically, Either with the EPC who might pay a six figure amount of US dollars in a month, sometimes in a week if they're late with side acceptance. And so they if it doesn't work, they really need to find out fast how to fix it, and data is the obvious answer to that question. You have eighty containers, some don't work, no one knows what's going on, look at the data. Right?

But typically they're not doing it yet, which is crazy. But okay. And then on the other side, you have companies that might have a twenty five million dollars investment.

And now after one week of testing, without detailed looks into the actual batteries. It's like hey, this container, it switches on. Yes.

And if we press discharge, some electrons go out, okay? Please sign here.

And what we found in these analyses, just one example, is we did an analysis on the internal resistance of all the cells and what you would expect or hope for is that they should be pretty close to each other because it's like a chain. The weakest link kinda defines performance of the overall chain. And so if one or few batteries have a very different internal resistance, it drags down the entire the entire site. And so we were looking for these outliers. Right? So to see if everything is good.

And what we found is that very clearly, this side was using batteries from three different production batches. So there were like three distinct groups with very different internal resistances, and they would just mix them match.

Like Wow. Can you can you say who the supplier was?

No. I can't.

No. It was acronym. That that that gets it down for a hundred percent of the markets at ninety nine percent of the market. No. So it I mean, once we saw the results and we started to ask ourselves, okay, how the heck could this happen?

But the answer is pretty easy, right?

Most of the batteries that are being produced right now come from production lines that were not up and running two years ago. So they're all in the early phase of ramp up, relatively phase. Think about this. Yes.

This this build out of battery manufacturing capacity Yes. You you only have to speak to see Elon Musk sleeping on the floor of the factory to know that it's -- Yeah. -- really hard to build manufacturing at scale fast it is.

And the other thing, so cell quality, I mean, understandably there's a certain spread in it and what manufacturers typically do to get around that problem is they do end of line testing and then they group cells that are similar to each other and then they pack them up and send them to a site. And apparently, that site got three different groups of these cells, whereas it should have been three times the same group. And then whether the internal resistance is five percent that way or that way as long as it's all the same, it's okay. It's not perfect but it's okay. This is so I've got a mosquito fly.

Join this bloody recording, I've got there's actually two mosquitoes, but one's chilling on the chair over there. So have always wondered this. Acceptance Testing. So you you buy a battery, you spend fifty million quid on this thing, and you get factory acceptance testing, FA, and site acceptance testing, SAT.

And you should really spend a lot of money and time in it for in all other aspects of engineering or power generation -- Yeah. -- you take this stuff really, really seriously because if you pick up any issues there and you usually do, then you can commercially it's way cheaper to get them dealt with, and, you know, you can fix them at source. One thing I have noticed in the last few years is that the world of battery acceptance testing, generally speaking, the the manufacturer has got a lot more knowledge than the buyer.

Oh, yes. And that is that is quite if you look at, you know, gas turbine sales, for example, You can the bio is usually pretty well informed, but you can buy in that knowledge. But but such an emerging industry in a new accelerating emerging technology like battery cells, the the control and the power is still in the manufacturers hands. So Are you saying that you guys can change that using data?

Yes.

This is one of the, I would say, no brainer use cases that we're doing, when you look at a site acceptance form, like what are they testing for?

Okay. You have like hold point testing, black start capability, a shake down where you maybe just rip out a few communications cables and see if anything goes crazy. But it's crazy to think that none of the testing concerns the actual batteries.

So just looking at whether the battery cell quality you got delivered is to specification is not part of standard site acceptance testing. Why? We could just change this. Seventy percent of the value of the site. And it's not so difficult and it's not so expensive to run a few additional tests with a specialized company. I mean everyone's used to having owners engineers, go over the site and check if you know some of the cables are installed as they should be and check if the water piping is installed correct.

But no one yet or, I mean, our customers do, but many companies still don't really pay a lot of attention to the actual battery cells which is yeah. I think this will change very soon but Right now, it's not as standard yet. Yeah. We've got a lot a a lot of work to do there.

You gotta approach these things like it was your money. Right? So, generally, you work for these funds or, I know, a bank that's lending against it. It's other people's money.

But if it was your money, it was your ten million quid, you would go through everything, tooth and nail. One of the things that always surprises me is the amount of assets that are built. They have a usable capacity in couple hours or kilowatt hours. But it's never been tested.

So the manufacturer said, hey, I tell you what, you've got a one hour system. We're gonna test the capacity. We'll we'll we'll test half of it, and you can just double it. Or we'll test the crusty at half power for two hours.

And look, you gotta wow. And, of course, there's so many things that can go wrong when you're at full fat capacity for a step like a step wave. But still, we're seeing sites come online that have never been put through a whole power a full power discharge for an hour. And you put them into traded markets and they fall over.

Yeah. They all trip like thermally, you know, produce too much heat and then it gets hot and they switch off But it is starting to get better. It is starting to get much, much better. And I think asset owners are incredibly sophisticated and in the know now in a way that they weren't a few years ago.

What what would you advise people? So we talked a lot there about site acceptance testing or testing your site before it comes online. So what would you advise people to do? What the tests are really important?

The whole testing that's being done right now in terms of mechanical, electrical. I mean, this is all very important. Right? If the transformer isn't connected properly, decide won't work and it's also a major safety hazard.

So all of these things, they're still needed. Right? I don't wanna say let's just use data and get rid of the rest. No.

What I'm saying is that currently the tests are treating battery sites like like traditional assets like a gas turbine or like a wind turbine, not acknowledging that there are some additional complexities coming from the electrochemical part of the site. It's not just about loose nuts and disconnected water pipes, but also the electrochemistry going on. And what's really cool is that that can be solved really fast. It's it doesn't add Time to the commissioning process, it adds a little bit of costs for the services, but every single time we've worked with a customer, they saved several x the amount of money that they paid us for investigating.

Right, either because they found some shortcomings of the site where they got reimbursement from their supplier or they were able to go to the market faster because some of the problems they had been working on for weeks and weeks. Could be solved in a day all of a sudden because we were like, hey, you're aware that that module is just broken. You're not gonna balance it healthy, right, never like, oh, yeah, that's kinda what we were trying. So let's just replace it.

And then, you know, you have all the information you can make more choices.

Regarding factory acceptance testing, No clear standards. And again, there's a huge it's a seller's market right now. So there's a limited leverage that buyers have and the manufacturers really don't want to share more information than they they have to. What does that mean?

A seller's market? Oh, right now if you have a if you have a battery production line that can pop out batteries at an acceptable quality you can pick your buyers. So when there's more sellers than buyers, the buyers can be picky and can put some more requirements on on on the deal, but I mean, CATL is in a very, very good position right now. And if they don't wanna share information with you, they're just gonna move on to the next buyer.

Yeah. So if you are if you are gonna spend you're gonna spend this hundred million dollars on a battery. And in the tender the tender documents, you go out to CETL, BYD, all of those, and you say, right. What's really important to me?

I want a really great quality battery, but I also want access to all the data around the temperatures and the voltage years and how these cells behaving. And the supplier, CATL, or I'm not not picking on CATL here, but they're just the biggest, so it's easy. They can say, look, if you wanna buy from us, we'll double give you that. But trust us, we'll look after the batteries in your service level agreement.

We'll definitely give you access to all that data. You don't need all that data. And if you wanna slot in our production line, you're gonna you're gonna pay for know. Yeah.

You're right. It's good it's good for manufacturers. I'm not sure it's good for the ecosystem. It isn't.

And actually, you know, the one company that has been doing that hardcore is Tesla. Tesla doesn't give access to the operational data of their assets to the people who own the asset, which is crazy. I mean, after German law, they would be the rightful owners of that battery and, you know, Tesla holding back that. But then again, right, it reminds me of a little it's like the TC versus Mac thing.

Right? Uh-huh. You asked and you said to Steve Jobs. Right? You need to start telling us the the temperature of your cause and your graphics card.

On all of your Macs. You'd say, no. Forget it. I'm gonna I'm gonna keep it ready for you.

I'm gonna give you a beautiful user experience. Yeah. There's no way you're gonna have that. And it kinda feels the same way.

I'm aware that we've gone well over time here, and you might have something more important to do in fact you probably do. I wanna ask you the two last questions. The first one, this is the chance to plug something. So any big deal, sale, or some announcement now is your chance.

I don't know what I'm gonna ask you the the most important one, which is what is your contrarian view. But firstly, what's your plug?

Yeah. So we we have a nice announcement that came out in in several charges over the past year. We have a very successful cooperation with HDI Global, which is one of the leading insurance groups in Europe, and they're active in over a hundred countries.

And they are recommending us as a risk reducing measure for battery systems.

And they are giving better insurance conditions that can include other obligations, and can also include lower premiums.

And, you know Wow. Congrats, man. That is huge.

Yeah. Of course, it's a case by case decision in every case like insurance, no two contracts are the same, but for us it's a really nice it's a nice backing by such a conservative and risk avoiding stakeholder to, you know, to get their trust and we're the only company in the world who has achieved something like that so far. What an endorsement. To enhance your credibility.

You guys must be doing something right. Very well done to who I'm starting that deal. Okay. So what's what's your contrarian view? This might be a little bit battery nerdy, but maybe also interesting for the folks that stuck so long in the podcast. So to the last twenty five percent that are still listening, Listen the ship is pretty niche. Don't worry.

So and that actually goes back all the way to why I started company with a few colleagues because the contrarian view is testing a battery, like physically testing a battery in a lab.

Will get you worse results when predicting the future than doing nothing at all. So more information will lead to worse decisions.

Contari and enough? My brain's broken.

Why? It's okay. So you know, the traditional way to predict battery behavior is you get fifty to a hundred cells in a lab and then you test them under different conditions.

And you know, one you test while it's very hot, one very cold, one with a high sea rate, full cycles, part cycles, and then you bring all the information together in a model, and then you can say, well, if I add another twenty full cycles, looking into the lookup table for my testing, this will degrade the battery by zero point zero two percent or whatever.

And if I increase the temperature by ten degrees, this will kick up the aging rate by whatever, you know, always referring to the results from the lab test. Okay. And it makes a lot of sense.

However, if we look if we take one step back and we look at the setup that we're just trusting, where do these cells come from that you test in a lab? You go to your supplier and say, hey, I wanna buy fifty million quid of batteries from you, and I would need just a hundred specimen to test them in my lap before, so I could just send a few over. And see why this is just Wholes are they gonna send you? They're gonna free test the shit out of five hundred batteries and send you the best hundred.

Really making sure that there's nothing out of the ordinary that you could find there. And now you're doing all these tests and you're thinking, man, I really butter what it is. This is like Volkswagen diesel cheating scandal scandalous.

Well, you know, you can't force them to give you realistic batteries. It's they're handing you what they want you to see. Right? No. I can I can imagine I can imagine everything you're saying is true? Yeah.

And now if you're doing all these tests on this cream of the cream split of the battery production, you're priming yourself to expecting that all the others will be just the same because that's how a model works like. You build a model and then you assume it's it's valid.

But in reality, the battery that gets you in the news because it burned down or during a heat wave and taxes It just didn't work, and now you had a blackout, all these nasty things.

They are caused by the batteries that you will never get for testing. These are the outliers that kind of, oh shit, we need another ten megawatt hours for that customer, but we don't have to AAA grades.

Yeah, just fill it up with whatever cuts.

And so that's why at a cure, we believe Building our models based on tests makes it worse. Instead, we're looking at each individual cell behavior and will building the models on the fly so that we also include the weird ones.

That is absolutely fascinating. My brain's broken.

Okay. There's a healthy dose of cynicism in there, but I think measured and appropriate.

So k p, I just wanna say a massive thank you for coming on a pod. We're gonna have to have you on again because we didn't cover eighty percent of what we're gonna talk about. There's so much stuff we talk about. But I wanna say from from our side, from me to you, it's been great watching your journey.

Glad to be your friend and seeing a cure go from strength to strength. And we're gonna have you on again soon. That was brilliant. So if you listen to this -- Thank you.

Thank you so much. If you're listening to this, hit all the good buttons, like, subscribe, all that stuff. Yeah. It really matters.

Cheers guys.