Thank you for standing by, and welcome to the Enovix Corporation First Quarter 2026 Earnings Conference Call. Operator instructions were provided. As a reminder, today's program will be recorded. And now I'd like to introduce your host for today's program, Robert Lahey, Head of Investor Relations. Please go ahead, sir.

Thank you. Hello, everyone. Welcome to Enovix Corporation's First Quarter 2026 Financial Results Conference Call. With me today are President and Chief Executive Officer, Dr. Raj Talluri; and Chief Financial Officer, Ryan Benton. Raj and Ryan will provide remarks followed by Q&A. Before we begin, please note that today's conference call contains forward-looking statements that are subject to risks and uncertainties. These statements are based on current expectations and may differ materially from actual future results due to various factors. For a discussion of these risks, please refer to the disclosures in today's press release and our filings with the Securities and Exchange Commission. You can also find these materials on our website at ir.enovix.com. All statements made on this call are as of today, May 13, 2026, and we undertake no obligation to update them, except as required by law. Additionally, during the call, we may reference non-GAAP financial measures. You can find a reconciliation of these to the most directly comparable GAAP measures in the materials posted on our Investor Relations website. With that, I'll turn the call over to Raj.

Good afternoon, everyone, and thank you for joining us. This quarter marked another meaningful step in Enovix's transition towards commercialization and scale. We advanced across the areas we believe are most important for long-term value creation: customer engagement, commercial deployment of our silicon-anode batteries and manufacturing readiness. I'm very excited to share that in the smart eyewear market, we commenced commercial production of our A1 battery for our lead customer's reference platform and have multiple customers in the process of launching smart eyewear products. Initial shipments are underway with production expected to ramp through the second half of the year. We believe this validates our ability to manufacture our 100% silicon-anode architecture at commercial scale. On smartphones, we aligned with Honor on an updated qualification framework designed specifically for silicon-anode batteries. This framework, which includes revised specifications and testing protocols, better reflects real-world usage conditions. We are pleased to have also aligned with our second smartphone OEM on the view that they too will need to adapt a similar testing framework in order to get their products to market on a competitive timeline. Beyond these customer engagements, we're in active dialogue with several additional leading OEMs regarding silicon-anode battery qualification standards, and we are encouraged by the constructive cadence of our discussions with these OEMs as we work towards future qualifications and commercialization programs. Importantly, we believe we are doing the hard work now that will enable our future OEMs to roll out their silicon-anode solutions more rapidly. The principal structural mismatch in qualification has now been addressed to align with silicon-anode performance while maintaining and in some respects increasing qualification rigor. Cycle life testing at our lead customer for batteries that we shipped at the beginning of the year is progressing under the updated protocols with the results approaching required thresholds. The deep partnership and technical engagement we are seeing with multiple customers in the smartphone market reinforces our belief in the industry's interest in high-performance silicon-anode battery solutions. We're encouraged to see growing demand across our drone, defense and industrial applications, securing new customer design wins during Q1 2026 in each of these markets with deployments expected in 2027. Our global pipeline for products manufactured in Korea now exceeds $130 million, with the majority driven by rapidly expanding drone applications where demand for high-performance battery solutions continues to outpace the available supply. This creates an opportunity for an additional scaled high-performance supplier. We believe Enovix is positioned to emerge as that differentiated supplier in this rapidly expanding market. We continue to improve manufacturing execution at Fab2. Yields in most production zones are now nearing or exceeding 90%. Zone 1 dicing, a key throughput driver, is delivering step level yields of approximately 80%, demonstrating continued progress with our laser-based equipment. We recently appointed Steve Bakos as Senior Vice President of Worldwide Sales to support Samira Naraghi, our Chief Business Officer. Steve brings more than 35 years of global semiconductor sales leadership from companies such as Infineon, where he served as a Vice President of Corporate Account Sales for large global accounts, including Apple. This quarter, revenue was $7.6 million, driven by Korean military contractors above the high end of our guidance range and up 49% year-over-year. Non-GAAP gross margin was 26.3%. Now I'll walk through each of these areas in more detail, starting with manufacturing. On that front, I want to give you an update on our Zone 1 dicing, which is our current throughput bottleneck. Since I joined to improve the throughput of Zone 1, we have been working on a faster, cheaper way to dice our coated rolls. We have been making great progress. Last week, I received a video from our equipment vendor showing this in action. Rather than trying to describe it with another chart, let me just show you the actual process. What you saw was our mechanical dicing system processing silicon-anode strips directly from coated rolls. We're implementing a hybrid dicing configuration strategy that combines both laser and mechanical dicing approaches. I continue to be encouraged by the substantial progress our global operations and advanced manufacturing equipment teams are making in this novel area for silicon-anode batteries. As we mentioned on the last call, legacy smartphone qualification protocols were originally developed around graphite-based batteries and relied on a 0.7C discharge requirement. That standard can artificially stress silicon-anode cells at discharge rates far above real-world smartphone usage, which typically remains well below 0.2C. The consequence was important. Silicon life testing under this framework systematically understated silicon-anode batteries' longevity. We've aligned with Honor on a new silicon-anode specific qualification framework. The updated framework prioritizes a version of the 0.2C cycle test that commenced in Q1. This methodology better reflects real-world usage for silicon-anode batteries while enhancing the rigor and visibility into performance. We are seeing broader industry alignment around silicon-anode specific qualification standards. Our second smartphone OEM has joined our lead customer in removing the 0.7C test from their list of hard requirements as we're now progressing towards an updated framework similar to our lead customer. Discussions with several additional top OEMs are ongoing. We expect broad adoption of similar silicon-anode specific qualification approaches over time. With this framework now established, the plan with Honor is a targeted system-level deployment in the second half of the year to confirm in-field performance ahead of the broader commercial launch in 2027. Importantly, we also recently received the battery form factor for their next-generation device to support readiness for the next major product launch. Our commercial strategy centers on two complementary technology platforms that address large and, in some cases, rapidly expanding market opportunities. AI, short for artificial intelligence class, our flagship 100% silicon-anode platform, is targeted at smartphones and smart eyewear markets where volumetric energy density is a key requirement. Smartphones represent the largest battery market opportunity for Enovix. However, smart eyewear is emerging as one of the fastest-growing new device categories. We think that the smart eyewear battery market opportunity could exceed $1 billion by the end of the decade. More broadly, the AI platform is applicable to virtually any space-constrained device requiring high energy density and long cycle life, including future applications in wearables, computing, industrial handhelds, EVs and humanoid robotics. Previously, we acquired an established business producing graphite-anode-based products. These products are in production today, generating revenue in defense, drone and industrial markets through our Korea facility. We've been able to leverage these capabilities in combination with our silicon-anode technology know-how to create a high-performance MX silicon-enhanced platform. Our initial target market for MX represents more than $4 billion in opportunity, including approximately $2.4 billion in drones and $1.8 billion in defense technologies beyond drones. These applications prioritize performance and supply chain security with a greater focus on gravimetric energy density. Over the longer term, we believe the MX platform is also well positioned for adjacent markets, including robotics, eVTOL, healthcare devices, transportation, agriculture and broader industrial applications. The first product Enovix is launching on this platform is MX1, a ruggedized drone cell design requiring rapid discharge and high gravimetric energy density. I want to highlight something important here. These are not separate bets. They are mutually reinforcing platforms, sharing technology, supply chain capabilities and commercial infrastructure. We're increasingly seeing benefits flow in both directions with the AI platform leveraging Korea manufacturing strength and the MX platform benefiting from our silicon expertise and global commercial reach. Alongside qualification progress, our R&D efforts continue to advance the platform. This quarter, we produced the first engineering samples of AI2 for smart eyewear, delivering greater than 20% higher volumetric energy density compared to AI1. This represents a meaningful architecture-driven improvement, potentially enabling product categories that require significantly more power within highly constrained form factors. We've achieved this improvement through two primary drivers, reducing inactive material to improve packaging efficiency and increasing the cathode voltage. Together, these advances increase energy density within the same footprint and further demonstrate the advantages of our 100% active silicon-anode architecture. We believe this represents the first of many future advancements unlocking the full energy potential of 100% active silicon-anode architecture on the future AI product roadmap. Display-equipped smart eyewear is expected to become a rapidly growing battery market. And we believe increasing power requirements create a strong fit for our technology. Smart eyewear also represents an attractive initial commercialization opportunity for our silicon-anode platform. Qualification cycles are generally shorter, more flexible and durability requirements are lower, and the market is in the early adoption stage. Customer sampling of AI2 is planned for later this quarter. We have already received initial sampling orders and engagement commitments from several leading smart eyewear companies. The 20% energy density improvement achieved with AI2 is important not only for smart eyewear, but also because similar gains for future smartphone batteries could materially extend our technology advantage. The current AI1 smartphone battery delivers 935 watt-hours per liter and has been independently validated against graphite and silicon-anode alternatives. We believe this positions Enovix with a meaningful competitive advantage in high energy density mobile applications. Competing approaches remain largely focused on conventional graphite-based designs with incremental silicon additions. These architectures continue to face swelling constraints that limit long-term performance and energy density improvements. In contrast, our architecture is designed at 100% active silicon-anodes, which we believe provides a substantially higher long-term scaling opportunity. Now let's talk about our second platform, MX. This week, at the Michigan Defense Expo, we formally launched MX1-B01, a drone battery cell, delivering energy density of 360 watt-hours per kilogram positioning us competitively within the high-performance drone battery market. We achieved this performance through targeted silicon content enhancements, leveraging an already proven manufacturing platform. MX1 is designed for applications requiring excellent flight time, high discharge capability for power-intensive missions and a secure supply chain. We believe the product compares favorably with similar leading high-density solutions currently available in the market and offers a material cycle life advantage. We are manufacturing these cells from our South Korea factory, which has supported defense customers for years and our commercial focus is on drone manufacturers globally as well as their packaging partners. Following the Michigan Defense Expo, we plan to showcase MX1 at 11 additional conferences around the U.S. and Europe over the next two quarters as we continue building customer engagements and commercial pipeline activity. This slide shows how we see the MX platform evolving beyond the initial MX1 launch. Demand for high-performance drone battery supply continues to exceed currently available Western capacity, which we believe creates a meaningful opportunity for Enovix. These applications prioritize performance, reliability and supply chain security, supporting differentiated positioning relative to commoditized consumer battery markets. While drones are a key near-term focus, we've also established product offerings for subsea, munitions and industrial applications expanding the MX platform across multiple high-performance end markets. Our Korea and Malaysia manufacturing footprint directly addresses defense supply chain requirements backed by years of production history supporting major contractors and deployed programs. A key structural advantage for Enovix is vertical integration: because we own our manufacturing operations, we're not sharing economics with third-party contract manufacturers, which we believe supports both competitive pricing and attractive long-term unit economics as volume scales. As product competitiveness becomes increasingly established, the gating factor becomes commercial conversion, which is why we recently appointed Steve Bakos as the Senior Vice President of Worldwide Sales. He brings more than 35 years of global semiconductor and industrial sales leadership experience and will help build the commercial infrastructure required to support growth. Looking ahead, MX2 is targeted for 2027 with the goal of reaching 400 watt-hours per kilogram. Over time, we intend for MX to evolve into a broader platform strategy, spanning multiple product formats and defense and industrial end markets. Now I'll turn it over to Ryan to walk through our financial results. Ryan?

Thanks, Raj. Our first quarter results reflect disciplined financial execution alongside continued commercialization investment. First quarter revenue was $7.6 million, above the high end of our guidance range and up 49% year-over-year. These are largely batteries deployed in active programs with repeat demand. Non-GAAP gross margin was 26.3%, our sixth consecutive quarter of positive gross profit on both a GAAP and non-GAAP basis. Non-GAAP operating expenses were $30.8 million, reflecting continued investment in customer qualification completion, research and product development and smart eyewear production readiness. Non-GAAP loss from operations was $28.8 million, better than the guidance range of $29 million to $32 million. Non-GAAP net loss per share was $0.14 at the better end of the guidance range despite higher interest expense from the 2030 convertible notes issued last year in Q3. Adjusted EBITDA was negative $20.3 million, roughly flat year-over-year. We ended the quarter with approximately $582.7 million in cash, cash equivalents, restricted cash and marketable securities. We believe this provides substantial liquidity to execute on our operating plan to support commercial scale up and to pursue strategic opportunities from a position of strength. Free cash flow was an outflow of $36.3 million, increased from the same period a year ago, primarily driven by timing-related items, including the semiannual interest payment on the 2030 convertible notes and working capital movements primarily higher inventory levels in Korea to support planned shipments. Capital expenditure payments were $3.2 million in Q1, below guidance due to the timing delay of certain payments, the majority of which we expect to be paid in Q2. Turning to Q2 2026 guidance. Revenue is expected in the range of $8 million to $9 million, reflecting continued growth in defense and industrial shipments and initial smart eyewear revenue as deliveries to our lead customer begin. Non-GAAP loss from operations is expected between $29 million and $32 million. Non-GAAP net loss per share between $0.13 and $0.17, and capital expenditure payments are projected in the range of $9 million to $13 million, which includes the aforementioned deferred payments as well as initial payments for the investment to support capacity expansion in Korea. Last quarter, we approved a share repurchase authorization to provide additional capital allocation flexibility. We have not made any purchases under that program. Our capital deployment priorities remain unchanged: qualification completion, scaling smart eyewear and defense production capabilities and selectively pursuing strategic opportunities with a high bar on strategic fit and return. And with that, I think we're ready to take questions.

Operator instructions were provided. Please note that this call is being recorded. Before we go to live questions, we're going to read the two most highly voted questions submitted by shareholders ahead of this call during the call registration. The first question is, "Previously, management has discussed multiple pathways to achieve final smartphone qualification targets. Could you elaborate on which of these pathways currently appears most promising?"

Yes. Thank you for the question, and thank you all for listening. Of the pathways we discussed, as I mentioned in the prepared remarks, we have now aligned with Honor on a combination of different pathways that we could use. The 0.7C legacy test requirement, which was developed mainly for graphite batteries, is not a must-have requirement and has been removed as a gating item. We're now working on a slightly different 0.2C test, which better reflects the real-world usage of the smartphone, and that's been prioritized now. Not just with Honor, but many of our other lead customers also agreed to the same thing. In general, I feel like the smartphone market now realizes that as silicon batteries become more and more popular, they should change the requirements that have been used mostly for graphite. So it's a really great result. I'm really pleased with my team being able to convince them. The 0.2C test is more than halfway done at our customer, and it's continuing to progress and we're tracking it.

The second question is for your AI2 platform. When will the samples be submitted to customers so that testing can begin?

Yes. As I mentioned in the remarks, we have engineering samples now internally, and they look really good in terms of the 20% energy density increase versus AI1, a great achievement by the R&D team, harnessing the full potential of silicon. These will continue to get better over time. We expect to sample this quarter to our customers and quite a few people have actually expressed interest in that and we have a few sample batches out now.

Operator instructions were provided. Our first live question will come from Mark Shooter with William Blair.

My first question is just focusing in on Honor a bit. I saw in the press release that you have some field testing looking at second half, right, is when you're targeting. So I'm wondering if you could give us an updated understanding of what unit volumes may be for that field testing with Honor. And if successful, do you have a better view on what a ramp would look like in '27?

Yes. So, again, as I mentioned, we have a test methodology aligned with Honor, and we're in the middle of doing that. The next big step is to actually put the battery inside an existing phone for which we made this custom-size cell. It's hard to predict the number of units, and they'll be small because it will be an initial test just to make sure that everything is smooth and a limited launch maybe. But the real volume will be in 2027. More importantly, we got the size for that particular battery that's going to be launched in 2027, and we're now working on how to make that battery after the field testing is complete.

Okay. Great. Switching gears into drones, which is a very interesting opportunity. And congrats on the silicon carbon composite and that mixed graphite cell. A 360 watt-hours per kilogram, that's going to put you well in the running against the current peer set. So I'm wondering if you could speak to maybe some of the customers that you're engaged with in sampling and say you win all that business today everyone you're engaged with. Do you have an idea of what those qual times look like and your current unit volume and revenue capacity?

Yes. Firstly, thank you for that comment on the drone cell. We are also super proud of the engineering team that came up with it in a very short order, particularly because we have a well-established manufacturing facility and we were able to quickly make that. The product actually is extremely competitive with what's in the market today and it's made entirely within our factories. It's not contract manufactured. There's a lot of interest. I just got a call from our sales team who's at a drone show right now about the tremendous interest they're seeing. Because this is also an NDAA-compliant cell, which is actually a big deal for many of our customers. The go-to-market works this way: typically, there are companies who take this and put them in packs and integrate the BMS, and that system goes into multiple drone makers. So it's hard to call the volumes right now, but the market is growing really fast so we think that qualification times will be shorter than things like smartphones because there's such demand. Also, cycle life requirements aren't as long in some of these markets. Our cell goes to 300 cycles, but even shorter cycles are acceptable in some of these markets.

And maybe I'll chime in. Mark, you asked about volumes. Again, we talked about it in the CapEx forecast in prepared remarks: we're already spending dollars to add equipment to one of the existing buildings in the non-sim facility. One of the great strengths that we have is we have multiple empty buildings in that facility as well. So we have numerous stages of additional expansion capacity there, and we're going to do that in a methodical way as demand presents itself.

Your next question will come from Colin Rusch with Oppenheimer.

Could you talk a little bit about the mix of silicon material that you have in the new drone batteries? And the pathway from the 360 watt-hours per kilogram to 400 watt-hours per kilogram, how much of that is being driven by increased thickness or a different form factor or increased concentration of silicon in the anode?

Great question, Colin. For this cell, we made something that meets an existing requirement for a cell in the market today that many customers wanted us to provide as a fully NDAA-compliant solution made within our factories. We achieved that performance quickly and it performs very well. The cycle life is very good and the capacity holds. We have about 60% silicon in that cell, but we now believe we can increase that percentage because in this market some amount of swelling is acceptable since the cell is packaged inside a drone and you can apply pressure around it. It's not as space-constrained as a smartphone. Also, discharge rate, pulse discharge and number of cycles are variables we can tune. We mentioned 400 watt-hours per kilogram as an achievable target. I believe we can go higher by making the right trade-offs between cycle life, discharge rate and allowable swelling. We've been working on silicon for a long time, so we know what electrolytes work well with silicon. We have graphite and manufacturing know-how in Nonsan. That factory has been supplying defense for a long time, so our product quality is at a very high bar and the factory is qualified for that. That's why I think you'll see a pretty competitive roadmap from us for this market very quickly.

Great. And then looking at the laser cutting, I just want to understand the cadence of the learning cycle on yield improvement and how we should think about the engineering work that you're doing and how quickly you can implement that to start getting a little bit better output on the overall facility?

As I mentioned, from last quarter to this quarter we've improved our yield across multiple zones, well into the 90% range now in most zones. Laser yields are around 80% but have improved quite a bit from last time. As you saw in the video, we've been working on this for a while. Today, I thought it was a good time to show a combination of laser and mechanical dicing that cuts much faster and cleaner. When you laser dice, there are challenges we've been addressing around yield and throughput and cost. So when I started, we've been working on a different approach and you saw the mechanical dicing now. We have enough lasers and enough throughput to meet demand for this year. Our plan is to get the mechanical dicer online this year so for next year's demand we can use a mechanical dicer with some combination of laser finishing. So really exciting results and improved throughput.

Your next question will come from Jeff Osborne with TD Cowen.

Maybe just a quick two questions, but one quick follow-up on Colin's. Can you get to 90% yields without that machinery intact? Or do you need to add the mechanical dicing to get there? And is that more a 2027 event getting to 90%? And second, Ryan, any update on the M&A pipeline? I know you've been searching for a couple of quarters now. Also, Raj, a clarification: you mentioned providing packs to Honor to put in a phone in small quantities in the second half of the year, but then you mentioned something about getting a second design. Is there a second SKU that they've given you? Or is the SKU they gave you what they intend to produce in '27? I'm trying to get a sense of whether your relationship is deepening and they're giving you a glimpse of what they intend to commercialize after the first product launch.

Maybe I'll take the first part about yields. We are capable of getting to 90% yield. But yield goes hand in hand with throughput. The mechanical dicer operates much faster and is more economical for throughput. Ultimately it's about cost, and we believe the mechanical approach is the most economical way for some subprocess steps to operate.

It's exactly the latter. We actually have a SKU that will launch in 2027, which is a fairly large deal and shows the relationships we have with them. The packs to Honor in the second half of the year will be small volume — just to test and make sure system-level integration works and to support a limited initial launch. That will be small, on the order of thousands of units.

And on the M&A pipeline, I'm pleased with the pipeline; we have multiple opportunities we're pursuing. As we've said repeatedly, we'll be disciplined. It has to meet strategic fit and price discipline. We've looked at quite a few opportunities and passed when the price tag didn't make sense. We're excited about some opportunities we're pursuing, but we will remain disciplined stewards of capital and ensure any transaction makes strategic and financial sense.

Your next question will come from Ruplu Bhattacharya with Bank of America.

Raj, I wanted to ask the first question on smart eyewear batteries. I think the press release says that you expect 50,000 units in 2026. How should we think that scales as we head into next year? And how should we think about the revenue progression from smart eyewear over the next few years?

Yes. The 50,000 is this year. With the way yields and throughput are progressing, we're able to manufacture the cell and deliver to the customer; they're testing them and it looks good. That's a great result. The smart eyewear market is growing rapidly and could be in the millions of units next year. It's hard to predict exactly, but we have samples with many customers now because our battery meaningfully improves product usability: many glasses today don't last the whole day. That's why we decided to launch AI2 in that space first — the market asked for more energy density. The first product will launch this year. AI2 sampling will occur this summer and AI2 production will be targeted for next year. We expect scales into the millions next year.

Okay. For a follow-up, can I ask Ryan: as you ramp smartphones later this year and next year, how does that impact gross margin? I think some OpEx might move into COGS. So can you help us understand how to think about gross margin progression as smartphones become a bigger part of the mix?

Certainly. As we ramp smart eyewear and smartphones, you'll see some costs that are currently operating expenses move above the line into cost of goods sold. The majority of our cost of sales is materials, so gross margin will be driven primarily by the bill of materials cost and our ability to drive those material costs down. Direct labor, variable overhead and some fixed overhead are smaller components; as we reach meaningful volumes, those become a smaller percentage of costs. So the path to improving gross margin is through bill of materials reductions and manufacturing efficiencies as volume scales.

Your next question will come from Derek Soderberg with Cantor Fitzgerald.

I was wondering if there's any way you can segment that $130 million Korea pipeline for drones and defense opportunities: how much of that is legacy Routejade and how much of that is drones?

This is a future-looking pipeline, representing new designs we're working on with customers. Some of it is continuation of the defense business that Routejade had, but the majority is new work that we are winning. Drones account for over 60% of that pipeline.

Got it. That's helpful. And then curious on the NDAA-compliance piece: how unique is that? A few others have that, but not many. Might it be difficult for others to achieve over the next couple of years? And within drones, what kind of drones are you getting interest from — a wide variety from sea drones to air drones, heavy and light. Where do you think you can build a nice business in drones?

NDAA compliance is not easy to achieve because there are multiple elements: the origin of sourced components, required percentages from approved countries and where the cells are manufactured. For us, we manufacture in Nonsan, South Korea, which is a non-FEOC country, and the majority of our materials are not sourced from disallowed countries, so it's a big advantage for us. Because we own our factories and supply chain, we can meet those requirements. In terms of drones, we're seeing demand across training, public surveillance, inspection, public safety and multiple other markets. What varies is the discharge rate and number of cycles. The first product we made targets 300 cycles with high discharge capability, but we have the technology and factory to make different products optimized for higher energy density or different cycle life. We'll build a roadmap addressing various parts of the drone market as demand evolves. It's an opportunity that grew very fast.

And I'll add, we have a long history supplying subsea batteries out of Korea, so that's a continued strength for us as well.

Your next question will come from Alek Valero with Loop Capital.

Yes, just on smartphones, what impact is memory cost inflation having on your lower-end phone volumes?

We're not shipping much smartphone volume right now, so there isn't a material impact at present. We do see that total smartphone unit shipments this year may be lower, and memory cost inflation could affect lower-end phone volumes. Hopefully that will normalize over the next couple of years as we approach higher volumes, and the impact will be less pronounced.

Okay. And another clarification question. Did you say that the 0.7C testing requirement was removed or you're hoping to remove it? If removed, what impact does that have on your smartphone qualification timeline?

Our customers agreed to remove the 0.7C test as a must-do requirement. They've agreed to use a variation of lower-rate tests like 0.1C and 0.2C, which better reflect actual phone usage. That helps our qualification effort. In terms of timeline, lower-rate tests take longer to run because they're less accelerated than 0.7C. So initial qualification may take a bit longer in calendar time, but it's more representative and avoids overstressing silicon cells. Once the methodology is established and everyone understands the trends, the process should normalize and not impede future qualifications.

Your next question will come from William Peterson with JPMorgan.

For AI2, the 20% increase in energy density using the next-generation platform, can you speak about the trade-offs, including cycle life we should consider? I realize this is initially for eyewear, but assuming AI1 will be the focus for your initial and second smartphone customer, do you have a plan to sample AI2 for smartphones next year? Or is this longer dated? I'm trying to get a sense for what needs to be solved for next-gen to be used in smartphones and how to think about your smartphone roadmap beyond AI1.

Yes, absolutely. The 20% improvement in smart glasses came from increasing cathode voltage and reducing inactive materials to improve packaging efficiency. We'll apply those improvements to smartphone batteries and you should expect similar energy density improvements over time. There are other improvements we're pursuing on packaging efficiency and electrolyte optimization. We've been using 100% silicon architecture but haven't yet unlocked its full potential because we've been addressing associated challenges like cycle life, accelerated testing frameworks and fast charge. As we work with customers to establish the right qualification methodologies for silicon, we'll continue to improve energy density and roll those advances into smartphones next year.

I'll add the finance perspective: some of these roadmap improvements not only increase energy density but also reduce bill of materials and manufacturing cost, which is beneficial for margins.

One other point: our batteries swell very little. Many smartphone OEMs include design allowances for end-of-life battery swelling. If we can take advantage of that space, we can increase energy density further. Once we reach qualification, we'll work with customers to utilize that opportunity.

On the timeline for shifting to mechanical dicing, are there any new challenges we should be mindful of? I'm wondering about particles, mechanical stress or other technical issues you need to overcome. Why wasn't mechanical dicing the primary path relative to laser dicing?

Mechanical dicing is not trivial. The roll-to-roll process requires development of appropriate binders, coating processes and electrolyte compatibility so the material can withstand mechanical dicing without damage. We've invested significant R&D and know-how to get to this point. There are still issues to finalize — we need to fully integrate the dicing into the anode and cathode processes and validate full cells. We have enough laser capacity to meet demand this year and will take the rest of this year to finish mechanical dicer integration. Next year we expect to ramp mechanical dicing for higher throughput and lower cost.

On the testing you mentioned — 70 to 75 tests, two life cycle tests and one below-freezing power test — do you have insight on expectations? I think you're more confident about cycle life, but what about the freezing power test? That's not something we heard much about previously.

The below-freezing power test addresses a corner use case: silicon behaves differently than graphite at extremely low temperatures. If a phone with a silicon battery is very cold and then sees a high power draw, there are challenges in how much current the battery can deliver. We're working closely with customers to understand the specific use cases, how much current is drawn and which markets this matters for. Like the 0.7C issue, silicon requires different consideration than graphite. We're confident we can address this with qualification work and system-level integration testing.

Your next question will come from Auguste Richard with Northland Capital.

You mentioned the $130 million of pipeline for military projects. Is that an annual number or a lifetime opportunity, and how much capacity in Korea do you have to support that?

That's a total pipeline reflecting new designs we're discussing with customers. It's not an annual number; these are lifetimes of those designs and can take one to two years to launch. We have enough capacity today, but we are adding capacity now. From our acquisition we gained almost 300,000 square feet of factory with buildings, power and dry rooms. We're expanding that facility to increase capacity in line with demand and qualification timelines.

On a relative basis, the expansion is standard equipment investments — it's mainly blocking and tackling to add capacity.

Got it. And you mentioned the eyewear customer is a reference design. Is that a chipset company or an OEM reference design?

Due to confidentiality I can't provide more detail on the specific partner. A bit of color: eyewear is a style and fashion-driven market; many tech reference platforms are used by fashion brands for final products. Securing a reference platform with a leading technology provider helps enable go-to-market through fashion brands and broader channels.

There are no further questions at this time. With that, I'd like to turn it over to Dr. Raj Talluri for closing remarks.

Yes. Thank you, everyone, and thank you for all the questions. Over the past year, much of the discussion has been on validation of the technology and its commercial readiness. We believe this quarter we provided additional evidence and the conversation can increasingly shift towards disciplined execution against commercialized milestones. The markets to watch over the coming quarters are clear: the continued progress in qualification, targeted system-level deployments, initial smartphone production ramp and conversion of this drone pipeline into revenue. Now we view these as tangible operational milestones, and we expect to demonstrate progress methodically over time. Thank you all for your support.