Kyle Pilkington: Great. The next question we got from investors ahead of the call is what are the key challenges to scale lithium-metal anode in production when it comes to notching and stacking?
Qichao Hu: Yes. We actually covered this in previous earnings calls. So the material itself, lithium, and then especially lithium on copper is quite thin and then quite weak. And if you use conventional processes like laser or metal dye punching, you actually don’t get very good results. And then you end up with lots of issues like powders or tearing and also making large with very thin lithium foil is also very difficult. And then you end up with wrinkles and then we actually worked with several partners and tried different techniques from extrusion, lamination to disk coating. And, eventually, we actually settled on one process. And then in the early days, we used to have the vendors make it for us, but then we realized it was hard to control the quality. So then we took this in house, and then we’re able to improve the quality and the product a lot faster.
Kyle Pilkington: Great. And the final question we’ll take from the pre-submitted questions from investors is, how will you use artificial intelligence to advance battery technology?
Qichao Hu: So for us, it’s really two parts. One, is to ensure safety. And then anytime you have a new battery technology, especially one that has very high energy density, and then especially your B-sample and then C-sample and then very soon commercial, the consideration around safety becomes quite different. And then we’re not talking about just paper safety, but we’re talking about, can you put this battery inside a car or inside an aircraft? And then in actual usage, is it safe? And then also what happens when the worst happens? So the use of AI basically by collecting the manufacturing quality data, because also quality is safety, a lot of the issues that happen in quality will directly result in incidents. And also collecting actual live vehicle data, then we can actually predict the incident and then have a very accurate monitoring of the battery health.
And then we can actually predict the incidence before that happens. So this becomes really important, and then it’s really important to give OEMs the confidence that, yes, it’s a new battery technology. Yes, it has higher energy density than other battery technologies, but it’s actually safe because we can make it safe with the use of this Avatar AI. And then the other part is using this for future material development, and then that’s when we try to improve performance, cycle life, energy density, safety, and then for future road map, material development. And this AI really allows us, to screen a lot more candidates much faster.
Kyle Pilkington: Excellent. Thanks. I’ll turn it back to the operator. Operator, are there any additional questions in queue at this stage?
Operator: Yes. The first question comes from the line of Shawn Severson with Water Tower Research. Please proceed.
Shawn Severson: Good afternoon, everyone. Try this again. Can you hear me now?
Qichao Hu: I can hear you.
Shawn Severson: Okay. Good. Alright. Just making sure. Qichao, I wanted to talk about the B-sample push in 2024. And how does that reflect itself in terms of milestones or events over the next 12 to 18 months that are going to be able to provide kind of progress reports and updates. So just trying to understand the news flow of data that comes out of this as you go through this the B-sample push.
Qichao Hu: Yeah. The B-sample itself and entering B-sample itself is a big milestone. It represents a new chemistry lithium-metal is no longer in R&D or just early stage engineering development, but it’s actually in B-sample. And then we are considering a lot from the perspective of the final vehicles. So some of the milestones would include, for example, we are setting up the line, and then we will begin operating the B-sample lines. And then we expect second half of this year, we will have the B-sample lines running. And then in the first half of 2025, we will have data from the sales coming off the B-sample lines. And then these cells will likely be different from the original designs in terms of product designs. For example, some will have high nickel cathodes for the premium vehicles, some will have LFP cathodes for the economy cars and then also benchmark cost of LFP lithium-metal with high nickel lithium ion just showing that that these LFP lithium-metal B-sample cells can actually be low cost, but also achieve very high energy density.
And then yes, so also in the first half of 2025, then the quality, the number of quality control points, the incident prediction and then integrating the B-sample line with Avatar just showing that the number of cells that we built and the number of quality checkpoints per cell and then the product of the two, how this increased number of amount of data that we will have access to, how that will help train Avatar and make Avatar more accurate. So this B-sample line and Avatar are actually tied together.
Shawn Severson: That was kind of my next question was about, Avatar and safety as well. How differentiated is this 100% safety goal? When you look at some of the other technologies and batteries out there and we had conversations with OEMs, I mean, is this something that is very unique, you think, to SES that this goal and to be a realistic goal to achieve versus what’s out there today for options for OEMs?
Qichao Hu: Yeah. So the goal, obviously, is a goal shared by almost all OEMs near 100% safety guarantee. And then this is really important, because when you have a car or an aircraft out there, you need to make sure it’s safe. So what’s unique about SES is that all of this data are actually for lithium-metal. And then Avatar AI any AI that use is actually quite dependent on the quality of the data. And then no one else has the amount of lithium-metal data that we do. And the reason — we actually don’t care about the OEM JDAs. The reason that we build these B-sample lines is to collect more data, is to generate more data. So for us, the B-sample lines are like a mine of data. So no one else has these B-sample lines, and then no one else has the quality and the quantity of lithium-metal cell data that we do so that we can use all these data to train this Avatar.
