All right. Thank you, everyone, for joining us this 20th Midcap Growth Conference. This morning in the virtual session, we're really happy to host Qichao Hu, the CEO of SES AI. They're a battery technology company, and Qi will get more into it. They have exposure to a lot of growing end markets. So again, I think this will be a pretty productive call. Qi's going to give a presentation, and then we'll follow that with Q&A. If you have a question, you can submit it in the chat box on your side, and then we'll make sure we get it asked. With that, I'd like to hand the call over to Qichao. Thank you, Sean. Hi, everyone. Our goal at SES AI is to accelerate energy transition through AI. We are uniquely positioned to do this because we have more than a decade of experience in the battery domain with data and domain knowledge, and especially developing one of the toughest battery chemistries, lithium metal. This capability of core material discovery and safety monitoring is really important for the development of next-gen battery technologies. We're seeing the energy transition now is requiring the integration of both AI software and hardware. We are taking our core capability, material discovery and battery safety monitoring, to address several large and fast-growing markets, including ESS and also drones. The ESS market is expected to be more than 10 x the size of EV. And both ESS and drones' battery market are expected to have very exciting growth. And our core IP is a platform called Molecular Universe that includes the material discovery and battery health management. And then we provide really exciting competitive advantage in energy storage cells and also materials business. And then Molecular Universe has also evolved into our own AI for science company. And this asset is really exciting. And there are private companies that have very exciting valuation in the space. From a finance perspective, last year, 2025, just in the first three quarters, we booked a revenue of $16.4 million. And then we provided a guidance of $22.5 million for the full- year 2025. And then we'll share more about Q4 and the entire 2025 and also 2026 guidance in the next month when we do Q4. So I like this quote from the Undersecretary of Science and the Genesis Mission. So we're linking the nation's most advanced facilities, data, and computing into one closed-loop system. And the key is really closed-loop system. And then I'll show you why. Here's our history. We started back in 2012 as a spinoff out of MIT. And then we've been focusing on one of the toughest battery chemistries, lithium metal. And then that's really helped train a lot of the data and domain expertise and model that now form the core of Molecular Universe. And we started this AI initiative back in 2017 as part of our joint development project with a car company. And then 2021, actually back in 2021, we built back then NDAA-compliant lithium metal cell manufacturing plant in Chungju, South Korea. It used materials produced in Korea, and also the cells were assembled in Korea. Then we signed the world's first EV A- sample with three car companies, GM, Hyundai, and Honda. Went public in 2022. Then in 2023, we opened up this electric foundry in Boston to further collect data and then have high-throughput molecule synthesis, electrolyte formulation screening, and cell screening to collect more high-quality data to train Molecular Universe. In 2024, we entered B-sample, and we started booking our first revenue. Last year, 2025, we launched Molecular Universe, and we acquired UZ as a launching pad for energy storage. Then we formed a joint venture with Hisun to commercialize materials discovered through Molecular Universe. We partner with Top Material to manufacture NDAA-compliant cells for drone applications. This year, 2026, we are establishing three separate business units to focus on ESS, drones, and materials. So this really allows us to focus on revenue generation while Molecular Universe effort, we focus on value generation. So the core, after more than 10 years of development in materials and battery safety, is our AI for science, this SaaS platform called Molecular Universe. And then these enable several three exciting business units. One is ESS. So we supply fully integrated hardware-software solutions with our battery health and safety management software powered by Molecular Universe for data centers and commercial industrial. Second is materials. We supply materials discovered through Molecular Universe platform to users and also cells. We supply high- energy and power- density and NDAA-compliant lithium metal and lithium-ion cells designed through Molecular Universe for drones and UAM robotics applications. So again, we have the three business units focusing on revenue generation, and the Molecular Universe focus on value generation. And recently in the AI for science space, there are several exciting companies that have raised capital at exciting valuation. So first for ESS, what is our goal here? Our goal is to supply this hardware and software integrated solution. And we have very good software through Molecular Universe and battery health monitoring capability. And then we acquired a company called UZ because they have very good hardware capability. They make everything from the small 5-10 KWh all the way to the 20-foot container 5-megawatt-hour hardware. So we provide this battery health management software, and we help their customers predict safety issues so that they can reduce maintenance and operation costs. And then in return, with their deployed capacity in over 60 countries and then more than 0.5 gigawatt-hours, we get data. We collect data from all these units in the field. These are real-world data that we can use to further train Molecular Universe, so it's a positive cycle, and then in materials, so users of Molecular Universe, we have discovered six breakthroughs through Molecular Universe, and then these are listed here. They include electrolyte to improve cycle life and storage for EV application, include electrolyte to improve cycle life and power density for drones application, and improve low temperature cycle life, power density for heavy-duty trucking, better cycle life, longer runtime for consumer electronics, better cycle life, and better low temperature performance for energy storage applications, and also better cycle life in the storage for consumer electronics. These are six breakthroughs that users of Molecular Universe have discovered, and now these are being tested at 40+ customers. Molecular Universe is the first and only AI for science platforms that actually has made breakthroughs that has actually discovered new materials. Unlike other AI for science platforms, those are tools and have yet to make discoveries. We actually have discovered new materials. So we partner with Hisun in the joint venture to produce these materials on a commercial scale. And we have users that like Molecular Universe and then both from a material discovery platform and also just want to buy the materials from us. And these are more details of the six breakthroughs I mentioned earlier. For drones, drones is actually really, really exciting. So our value proposition includes two. One is high energy density and high power density. Actually, we first started supplying drones, batteries, lithium metal batteries back in 2018. And then we focused on EV applications. So all of our lines were occupied for EV cell production. But we built NDAA-compliant line and then cells in 2021 in Korea. So now we are converting that line to produce cells for drone applications. So we have the benefit of high energy density up to 500 watts per kg and high- power density up to 10C continuous, as well as these cells are NDAA-compliant in our own factory. And in addition to our own NDAA-compliant capacity, we also have third-party contract manufacturing capacity. So you can see in this chart on the left, the current lithium-ion for drones really just does not provide enough energy density or power density, let alone most of them are not NDAA-compliant. So drones is a really exciting market for us. And then finally, just on Molecular Universe, this value proposition, again, this is the first AI for science platform built by battery professionals for battery professionals. So this is a domain-specific AI for science platform, whereas other and when you talk about AI for science, you really need three things. You need model, you need data, and then you need domain expertise. Most of the other AI for science platforms out there coming out of developed by AI companies have models, but they don't have data and they don't have domain expertise. So those are not very useful tools, not very capable tools, whereas we have all three. That's why our AI for science platforms have discovered these six materials with more in the pipeline. So this is a really exciting capability that we are both offering to other companies to help their development and also using in-house to commercialize the discoveries that we make. And then in summary, so now we are focused on ESS and then drones applications. We have a very experienced team. And also as a company, we have both the software and the hardware technology for these two applications. And then in particular on drones, we've had NDAA-compliant line since 2021. It was for EV applications. And then now we're converting for drones applications. And then we continue to grow UZ's hardware business and then deliver on this exciting six materials that we have discovered with more in the pipeline. Our goal is to leverage this Molecular Universe core platform to accelerate the product development and focus on this CapEx-like business model for ESS, drones, and materials. That's a quick overview. Okay, Qi, that's great. Thank you. Specifically, when you think about the pivot that the company has made last year, ESS, drones, materials for 2026, if you could pick one of the three, which of the three is probably most exciting to you in terms of market opportunity? So I think all three are important. And then from a so I would say military drones and data centers, these two are the most exciting opportunities for us. If we touch a little bit on we'll start with the military or the drone side, not necessarily all military. But if you think about 2025, do you have active revenue in that space? Or what activities were you doing in testing cells with drone companies, like pilot cells with them? And how do you think that sort of proceeds from here in terms of timing and cadence? So in drones with customers, typically it takes about one to two years of testing before you start to get sizable revenue. And then by end of 2024, we started this pivot. So 2025, we focus on converting lines from producing EV cells to producing UAM and then drone cells and also testing with customers. We do expect to recognize some drones revenue in the fourth quarter. We'll discuss that later. And then so in 2025, with all these qualifications, and we expect to get pretty sizable revenue from these drones in 2026. And you had a slide up here that had your capacity on the drone side. I think it was like it said 1 million of NDAA-compliant cells and then 10 million cells through a third party. So are the cells in South Korea, are those also NDAA-compliant? Or what's the pathway to having higher capacity on the compliance side? So that one million NDAA-compliant cells, those will be produced in our South Korea facility. And then if there's a demand for more, we can increase that. So for example, that line, the current bottleneck is with a stacker, and then that produces one million cells a year. And then if we have demand for two or three million cells a year, we just double and then triple the stacking machines. We don't need to increase the whole line. We just increase that particular machine that is the bottleneck until another step becomes the new bottleneck. So yes, we can definitely increase the capacity there. And practically, the commitments there then for a stacking machine are relatively small. So you're talking about being able to add capacity quickly. Yeah. How do we think about pricing in the drone space and that 1 million of capacity on the NDAA-compliant side? What's the potential for that? That depends a lot on the particular cell format and the customer. In general, we're seeing NDAA-compliant cells about 2 to 3x more than the current non-NDAA-compliant cells, generally now, in the market. Okay. And then again, how should we think about gross margin profiles there for the drone side? We can share more in the Q4 when we get cell form. But in general, the gross margin is similar. The cost is higher, but then the customer is also willing to pay a bit more. It's in the 10%-20% range. They have the bigger piece of capacity that's not NDAA-compliant, right? It's in main cells. So can you say if you think about the opportunities that you have in the drones space and you split them, are you also seeing opportunities for people that don't care about NDAA-compliant cells? Yeah. So that's for other countries, non-defense use, for agriculture use. But for now, we're seeing about half, actually more than half, that really want the NDAA-compliant. Okay. And have you made any comments or can you make comments about the pipeline there, maybe how many different companies you've been testing cells with as you think about shots on goal for conversion next year? So since we focus on people that want NDAA-compliant cells, we really want to focus on the larger accounts, not lots of small accounts, but really the large ones. So we're focusing on about 100 key accounts. And then out of those, really about 20-30 large ones. And those primarily want the NDAA-compliant cells. Okay. Great. And then if we look at ESS and UZ, can you kind of give us a little bit more color on that acquisition and how that's positioned in the ESS market on data centers and C&I, what that product looks like? And you gave a pretty big size of the market in terms of gigawatts. But is that the addressable market for you, or are you a subsegment of that broader market? Because does that gigawatt include front-of-the-meter utility storage? So both front-of-the-meter and behind-the-meter, there is a strong need for an operating system. The softwares for both are very fragmented. And then in particular, in data centers, our end goal is we want to create this closed loop. So we have Molecular Universe that discovers new materials, and then we make the materials. And then we have companies contract make cells with these materials. And then we put these cells through UZ in a container. And then we provide this container and this battery health monitoring system as an energy solution to data centers. And then data centers will power Molecular Universe. So that's the closed loop we're trying to build. And in terms of UZ, UZ used to be a customer of ours before we acquired them. And then we provided battery health management. And there's a lot of BMS and EMS out there, but then no one goes super deep. And then no one can provide as accurate and high-confidence estimation of the state of health, state of charge as our tool can. And then that makes a big difference in terms of the longevity and the health of the battery, but more importantly, the economics of the battery operation. So if you can actually drain the battery even more to get more economics out of it, you won't be able to do that. And then this software allows us to do that. And then why did we decide to acquire them? To take our software to the next level, we really need more real-world data to train this model. And then if you were to collect data ourselves or from fragmented ESS operators, it will take a long time. So by acquiring a company, we get a massive amount of data. And then the accuracy of the Molecular Universe model actually had a step jump in the accuracy because we have this influx of data. So by having this data, now we can actually provide even more accurate for data center applications and also commercial and industrial applications. And also from the front of the meter because most of these are LFP chemistries and then just different use cases. So by having this data, we can apply this software tool to all the use cases. So UZ is predominantly LFP. That is the area where you have the LFP technology? Yes. For now, they're all LFP. Yeah. Okay. And I think, can you discuss sort of how many deployments UZ has to date and just kind of the scale of UZ? I think it's pretty broad in terms of geography, but are you trying to hone that down into certain growth markets for the next couple of years? So for now, it's Middle East, Europe, and they're trying to get into the US. And so for UZ, the biggest growth market is the US. But for us, our goal is not to sell this software to UZ, right? Our goal is to train this software by using UZ so that we can sell this to everybody. That's our goal. UZ. Right. So I guess, and we'll touch on that, but UZ, to get into the US, what are the hurdles there? Is it around cell supply, where the cells are coming from, or what's going to get you into the US with UZ? Yeah. So with UZ in particular, the cells currently are not NDAA compliant, and they need to switch to NDAA compliant cell vendors. And then we are discussing with a few. That's the key bottleneck. But for us, SES overall, once we train this software using UZ's cell data, because it doesn't matter if it's NDAA compliant or non-NDAA compliant LFP cells, you have the same data. And then once we have this software, then we can apply this not just to UZ products, but also the other products that are NDAA-compliant. Okay. So if you were to look forward a couple of years, UZ, if we think about hardware versus a software services business model, how should we think about that? In other words, hardware revenue is still going to be the dominant thing just because the ASPs on the hardware side are higher, but it's also very much lower margin. But how should we think about splitting those two and then sort of margin profiles between the two? So three years from now, our goal is to sell this hardware-software integrated, not separate, so that we have this big box. And then inside this big box, you have one battery chemistry for UPS, one battery chemistry for BBU, multiple battery chemistry that are managed by this software. But we will sell the big box. But will there be? there be? Is there a recurring revenue piece from the software side, or is it all going to be upfront? Yes. So upfront, the cost of the entire box. And then recurring, you have software updates, maintenance, and then subscription model. How should we think about? Are you at the spot where you can think about what the recurring piece would look like practically on a revenue stream? Yeah. I think that we can provide more details later. Okay. All right. And then you mentioned data centers. So are you actively engaging with data center operators, and particularly any hyperscalers on the UZ side? We are, yes. Okay. And that's for—is it primarily for—in other words, you're seeing some data centers talk about very large deployments of battery systems. I guess I'm just trying to understand if it's more of a niche application that UZ is fitting in, or if it's those broader deployments. So it's actually a broader deployment of energy solution to the data center. So if you look at a data center today, right, and the design is outdated, there are hyperscalers working on next-gen designs that will completely replace the current design. But the current design, you have a diesel generator, you have a UPS, you have a BBU battery backup unit inside. Just a UPS and a BBU alone, these two are very different chemistries. They could be both LFP, but very different designs. One is rated for maybe 1C, the UPS, and another is rated for 30C, 40C. So these are very different chemistries, very fragmented. And then for each one, there's no way to predict how much capacity you can get out of a one-year down the road to two-year down the road. And for example, the BBUs inside the data centers, basically once a month, engineers go and then collect very rudimentary data. And then there's no AI prediction. It's funny. You build these AI data centers, but the hardware inside actually don't really use any AI. It's really outdated software. So by having this new software, then we can help both the UPS and the BBU. And then where UZ comes in currently, one of their products is a 5 megawatt-hour, 20-foot container that they can use for the UPS. Not the BBU yet, but we, SES, we can source both the 20-foot container for UPS and BBU chemistry and then manage both through this software. Okay. Yeah. That's great. On the materials business, it's slightly different, but can you kind of talk about the capacity you have with Hisun and what that looks like from, I guess, these different materials? Maybe you have different ASPs, right? But in theory, what does that capacity with Hisun look like for you from a revenue standpoint? Yeah. So in terms of materials, we have users of Molecular Universe that want to use the platform to make discovery, and then that's it. And then they go do their own things. Or they want us to provide Molecular Universe as a service and help them discover materials. And then once we discover, they actually want to buy materials from us. So then in that case, Hisun currently has an annual capacity of 150,000 tons a year. And then most of the electrolyte, you do have market price. And then that goes up and down based on the cost of carbonate, hydroxide, the raw material costs. And then with new materials, you can add a premium on top of that. So we're definitely not constrained by capacity. And then so far, these six materials being qualified with 40 companies, it's a small fraction of their capacity. Yeah. And Molecular Universe, you're sort of like it's more of a SaaS model. So how do we think about the ramp-up and deployment of that over the next few years? I mean, how interested are you? How interested are potential customers on deploying that into their systems? And then how do you price that for them? What's the revenue stream from that look like over time? Molecular Universe, and that's why we say it's value generation, because this entire space, it's still trying to figure out the business model as a standalone business unit. This is why we have drones, ESS, and materials as business units, but not the Molecular Universe because it's so core. And then in terms of how to think about the commercial opportunities, so we have both a general platform, Molecular Universe, where we basically offer this platform to users in academia and then in industry for low cost. But then in return, we get data. And then that data helps make this tool really sophisticated so that we can actually sell this to the top-tier battery companies. And the way we commercialize this is the top-tier battery companies are very behind in AI. Then, for example, a company that has 20,000 R&D scientists has a huge advantage over a company that has, say, 1,000 R&D scientists. For now, it's a competition of manpower and the resource. Molecular Universe kind of levels the playing field. A company with 1,000 R&D scientists can actually make a similar level of breakthroughs as the one with 20,000 R&D scientists. This is actually quite big. The way we commercialize that, there are several tiers. We charge a subscription fee to use just the software piece. We charge a fee for development service. For some of the companies, they will give us a specific task. For example, those six materials we discover, these come out of specific development service projects. One company wants a better low-temperature battery for ESS, and then we would develop that. Another wants a higher voltage battery for cell phones and then works better in cold places. We'll also develop that. And then once we have that material developed, either we charge a royalty, a licensed royalty on the IP, and then they will go make the material themselves, or we would have this Hisun JV manufacture the materials and then make margin through the margin of the products. I guess I'd probably like you're a smaller company compared to some of the larger battery companies that are out there, and you're going after sort of three distinct businesses like drones, ESS, and then this material side. So over the last year and a half or two years, how have you restructured the organization to go after the opportunities there? On the drone side, obviously, there's a big funnel of a high number of customers. They want to sample cells. They want to get cells to qualify. Then you have to get into these programs, so I'm just curious of how you've built the organization to address the scale that you would be at. So I think it's about finding specific opportunities, and then we don't compete with the bigger players. We're not trying to capture a small market share in the big market, trying to compete with them. We're trying to focus on these specific markets and then really win there. For example, drones, not a lot of the almost none of the big cell producers is competing in drones, the pouch cells, because previously, it was seen as a small market. And then also the requirements on NDAA compliance, most of the big cell producers were focused on EV and then ESS hardware. That's one really exciting opportunity for us. Second, even for ESS, the hardware is a commodity. LFP cells for ESS now is about $40 per KWh. So we're definitely not competing with CATL, LG there. But the software piece, managing and then predicting the life and the state of health of these batteries, that is a very big but fragmented space. So that's an area that we have a lot of opportunity. And then in terms of R&D, again, it's hard to overestimate how powerful this AFSIS platform is. And last year at the Battery World, we talked about this Molecular Universe in a box. One of those boxes can replace about three scientists. And then if you have 1,000, 3,000 scientists, we will build a larger cluster. The amount of R&D work that can be done and the acceleration that can be done by AFSIS is really fast. I'll give you one example. Currently, human scientists, the success rate is about 40%. What does that mean? So a scientist, after nine months, a year, will come up with a breakthrough that has a 40% chance of success being validated in the pilot line and the commercial. Now, AFSIS, you still have the same 40%, but we do high-throughput screening. You get to know that in about a day. So that saves a lot of the time and the resource. And we saw that in the drug discovery industry, and then now the battery and the energy industry are gradually learning to also adopt this tool. Great. And then from the financial side, you're still burning a little bit of cash. So how do we think about what level of revenue? I know these are between the three business models, but how should we think about the pathway to getting to break even from a cash standpoint from here? Do you think that's possible in 2026 or 2027? Or what type of revenue level do you need to do that? Yeah. I think it's definitely possible in the next one to two years. And then we'll have more detailed guidance when we do Q4. But roughly, we're pretty confident that we can at least double the revenue this year compared to 2025. Okay. Great. I think we're kind of running up on the end of our time, Qi. So if you would like to kind of leave us with a couple of thoughts, what you think are kind of high-level, high-important thoughts that you want people to have coming out of this meeting, that'd be great. I think the battery industry as a whole, it's going through a really exciting transition. For example, five years ago, 2021, the focus was on EV, but now, I mean, name one battery company outside of China that has not pivoted away from EV since 2024. So now the focus is on ESS, drones, and the markets change, but the core fundamental technology remains the same, and then so for us to remain competitive and also sustain and fuel this growth, we focus on this really fundamental technology platform that we developed with more than 10 years of domain expertise and then data on Molecular Universe. And now we're seeing that we can actually commercialize this platform through these three business units that we have. So I think the battery space and also this intersection between battery and AI is a very exciting space. Great. Thank you, Qi. Thank you, everyone, for joining. Okay. Thank you, everyone.