Operator: Good day, and thank you for standing by. Welcome to the ECARX Q2 2026 Earnings Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded. I would now like to hand the conference over to your speaker today, Mark Hankinson. Please go ahead.
Mark Hankinson: Thank you, operator. Good morning, and welcome to ECARX's Second Quarter 2026 Earnings Conference Call. With me today from ECARX are our Founder and Chief Executive Officer, Ziyu Shen; Chief Operating Officer, Peter Cirino; and Chief Financial Officer, Dylan Jeng. Following their prepared remarks, they will all be available to answer your questions. Before we start, I would like to refer you to our forward-looking statements at the bottom of our earnings press release, which also apply to this call. Further information on specific risk factors that could cause actual results to differ materially can be found in our filings with the SEC. In addition, this call will include discussions of certain non-GAAP financial measures. A reconciliation of the non-GAAP financial measures to the GAAP financial measures can be found at the bottom of our earnings press release. With that, I'd like to hand the call over to our Founder and CEO, Ziyu Shen. Ziyu, please go ahead.
Ziyu Shen: Thank you, Mark. Hello, everyone, and thank you for joining us today. Last quarter, we outlined our vision to push the boundaries of automotive intelligence globally and how we are transforming into a truly global business, uniquely positioned to capitalize on the surging demand for higher-value software and physical AI. At our earnings in May, we said we expected a significant rebound in the market from Q2, both in terms of vehicle launches and shipments. The second quarter delivered exactly as expected. We delivered a strong financial result, and we continue to build momentum and make strong progress on our strategic objectives. The second quarter continued to be defined by disciplined execution and accelerating global momentum. Our top line revenue increased 45% year-over-year and up 71% from Q1. We reduced our operating expenses year-on-year despite the increased revenue. We grew gross margin to 19.8%, up from 10.8% this time last year. Most notably, we delivered our fourth consecutive quarter of positive adjusted EBITDA. I want to be clear about the market in which we achieved this. Demand in Chinese automotive has remained challenging through the first half of this year and global memory costs continue to be a significant factor. The growth in our revenue and profitability in this environment clearly demonstrates that the lean operating strategy we built through 2025 is doing exactly what we designed it to do. Throughout the quarter, we executed on our core priorities for the year with focus, accelerating our global strategy and investing in our R&D road map. That progress has strengthened our confidence going into the second half of 2026. First, our global expansion was accelerated during the quarter with more new models entering mass production, expanding the visibility and scale of our solutions. Our partnership with Volkswagen Group continues to make good progress. We are now putting the engineering, supply chain and support infrastructure in place across the Latin America region and remain on target for launch in 2027. Second, we continue to invest in our robust product and R&D road map. We expanded in 2 important ways this quarter. In June, we signed a definitive agreement to acquire the entire Flyme software business for approximately $266 million. This brings a highly strategic piece of our full stack ecosystem into the business. Flyme Auto is already deployed in more than 2 million vehicles and Flyme OS is already a core part of our Cloud Peak middleware used globally. So we are acquiring a mature platform that our own business already depends on. Flyme's momentum continues. It has recently been nominated to provide the software for a leading international luxury brand in China. In May, we entered into a strategic partnership with TPK Holdings to co-develop the ORCA LiDAR platform. As part of the process, ECARX will lead the systems integration capabilities and TPK will provide its manufacturing expertise. We are looking forward to seeing mass production begin in 2028. And lastly, we signed a share exchange agreement with Qualcomm Ventures last month. Qualcomm has been a strategic partner of ours for many years. We have collaborated deeply across multiple generations of solutions, including Zenith, which will be built on the upcoming Snapdragon Elite automotive platform. This agreement reflects another major milestone in our relationship to jointly develop highly specialized and integrated solutions. The progress we made during the quarter all points in the same direction. Our financial results are delivering even in difficult markets. We own more of our technology stack than we did before, allowing us to capture higher value opportunities. And we entered the second half of the year with continued confidence in our strategic and financial direction. I will now pass the call over to Peter Cirino to discuss our operational progress in more detail.
Peter W. Cirino: Thank you, Ziyu. Good morning, everyone. At our earnings in May, we described Q1 as being a historically subdued quarter for both seasonal and quarter-specific reasons, and we guided to increasing momentum from Q2 onwards, both in terms of vehicle launches and shipment volumes. This has played out in Q2 as we expected. We achieved the launch and volume rebound we guided to in the first quarter, underscoring our strategy for our global commercial build-out through executing complex global programs across diverse vehicle lineups and markets. Shipments in the second quarter were approximately 550,000 units, an increase of 51% quarter-on-quarter that directly contributed to strong top line growth. Against same quarter last year, volumes were 2% lower. While year-on-year shipments were marginally down, overall revenue and revenue quality was significantly increased. Sales of goods revenue increased both quarter-on-quarter and year-over-year as shipments of our high-end solutions continue to grow and demand accelerates. Shipments of our high-end Antora solutions, in particular, increased 92% quarter-over-quarter and 52% year-over-year. Furthermore, shipments of our high-end performance AI-driven computing platform pikes increased 43% quarter-over-quarter and over 2,000% year-over-year. Antora and Pikes now make up 42% of shipments. These are the direct results of the decision we took during Q2 last year to begin phasing out our lower-margin legacy platform business and concentrate on high-end fully ECARX architected solutions. That decision moderated our unit volumes for a period, but is now improving the quality of what we sell. The second quarter is also where our 2026 model launches began landing at scale. And every launch pulls hardware volume and associated engineering revenue with it. Software revenue decreased from the same quarter last year by 42% due to lower sales volume, whereas Services revenues increased substantially from the same quarter last year by 21%, driven by new model launches. Services revenue comprised of reoccurring software fees, but the bulk of it today tracks the timing of design and development contracts and the vehicle launch cycles they support, so it can be lumpy by nature. We are pleased to see this improvement as we guided to expect 3 months ago. As launches accelerate, we anticipate software and services revenue will accelerate with them. To provide some context here, revenue in any given quarter is a function of 3 things. Those are vehicle model launch timing, shipments driven by end market demand and component pricing. We manage the first through operational discipline, the second through geographical and customer diversification and the third through pricing adjustments to structurally support top line revenue and protect profitability. But this business will show quarter-to-quarter variability, and we encourage you to look at the trailing 4 quarters rather than any single one. Turning to our customer base and growth strategy. During the quarter, we began mass production for 9 new models across 4 brands, of which the majority are using our next-generation Pikes or Antora series solutions. Of these new models, 4 are designated for markets outside of China, including Europe, Southeast Asia and South America. We are pleased to see this type of growth that further reinforces our strategy on transforming into a global company. We now have 12 million vehicles with our technology on the road. As Ziyu mentioned, our partnership with Volkswagen Group continues to drive forward during the quarter as we continue to build out engineering, supply chain and support infrastructure in the first region to support its expected launch in 2027. The program integrates our high-end Antora 1000 with Cloud Peak and Google Built-In for premium segment vehicles, alongside our cost-effective Antora 500 for entry-level segments. I want to again highlight the flexibility and scalability of the unique value proposition we are offering here, one portfolio of solutions that covers the full price ladder. Ziyu has already covered the strategic rationale of the pending Flyme acquisition. Flyme consists of 2 distinct but related pieces of software. The first is Flyme Auto, which is the application layer, which we use for the interface for products sold in China. In international markets, we use Google Built-In for this layer. The second piece is Flyme OS, which is the Android platform that we embed into our Cloud Peak middleware. This is the core of our software stack both in China and internationally. Let me turn to what this acquisition will change operationally in both of these markets. The first is road map control. A competitive advantage of ECARX is our ability to tightly integrate our product solutions across layers from silicon to sensors to software. Owning Flyme allows for deeper hardware and software integration and greater customization. That shortens the integration time lines for automakers, provides them with standardized, flexible solutions for diverse vehicle lineup and accelerates time to market. More importantly, this will also create a competitive moat, strengthening our ability to execute complex vehicle programs at scale. The second is a revenue stream that is not tied to hardware volume. Flyme generates revenues today from software licensing, from custom development work and from intelligent cockpit system delivery. Adding a licensable software asset will allow us to move up the automotive value chain and capture greater margin. The third is interoperability. Flyme OS, which is embedded in Cloud Peak, already spans vehicles, smartphones and wearable smart devices, which means the car connects seamlessly to these devices, which drivers already carry. What differentiates Flyme OS from current products is its speed and close integration with the rest of the stack, delivering a superior user experience. This fully integrated cross-domain ecosystem equips automakers with solutions they can deploy, whether that is Flyme Auto in China or Google Built-In outside of China across the lineup to differentiate their vehicles in an intensely competitive market. We will operate Flyme as an independent software division, which will preserve R&D continuity and ensure a seamless transition for existing customers. Existing operators of Flyme OS will continue to receive updates and user data remains in each operator's ownership. The second addition to our portfolio is our partnership with TPK to co-develop the ORCA LiDAR platform, making our formal entry into the LiDAR sector. Under that agreement, we will lead system integration, sensor fusion and global commercialization, drawing upon our relationships with international automakers and robotaxi operators. TPK will contribute optical design, engineering and high-volume precision manufacturing. Mass production is scheduled for 2028 at TPK's facility in Thailand, and we're excited about the additional options this will allow us to provide automakers as we continue to drive further hardware and software integration. Before I pass the call to Dylan, I want to leave you with one final thought. What these partnerships and solutions provide are critical to our broader strategy. When a global automaker asks us for a solution, we can answer with our own silicon heritage, our own computing platform and soon our own operating system and our own expanding sensor technology. Very few companies in the industry can offer this sort of closely integrated stack comprising silicon to software to sensors. With that, I will turn the call over to Dylan.
Dylan Jeng: Thank you, Peter, and hi, everyone. The second quarter performance is a clear demonstration of the operating leverage we have been building into this business. Revenue rebounded strongly as launches and volumes recovered after a historically weak Q1. Our cost structure continued to improve, and we delivered our fourth consecutive quarter of a positive adjusted EBITDA. We achieved this while managing a memory cost environment that has moved sharply against our industry. Starting with the top line. Total revenue was driven by 4 factors: growing demand outside of China, higher-value products, new model launches and the DDR memory price adjustment flowing through our pricing. Sales of goods revenue was $196 million, increasing 73% sequentially and 50% year-over-year. Software revenue was $0.7 million or a 42% decrease year-over-year due to lower sales volume. Service revenue was $28 million or a 21% increase year-over-year, driven by the new model launches in the quarter. The Chinese auto market remains challenging. However, as we guided in the first quarter, market conditions improved overall in the second quarter, particularly momentum related to the first quarter. Shipment volumes were up 51% quarter-to-quarter. I want to spend a moment on ASPs because arithmetic this quarter points directly at it. While volume was slightly lower year-over-year, revenue was up 45%. This was driven by 2 main factors. The first is the quality of the revenue with our high-end products increasingly accounting for a larger share of our shipments. Our Antora and Pikes solution both increased in volume year-over-year, resulting in combined 71% gains in unit shipments. This is the deliberate mix shift we began executing last year, and it is working as intended. The second is the memory cost. Higher global memory costs have structurally supported our top line revenue as those costs passed through into our pricing. Gross profit was $44.5 million with gross margin expanding to 19.8%, a significant improvement on the same quarter last year where margin was at 10.8%. And looking forward, our margin profile will continue to be influenced by global memory cost. While higher memory costs to drive higher revenue, we continue to expect that gross margin and the operating profitability may be negatively impacted by memory cost dynamics in the coming quarters. Our response to manage this impact is the one we have executed consistently. That means managing our supply chain, controlling our cost structure, maintaining pricing discipline and concentrating R&D on the higher impact solutions. Our new operating strategy continued to deliver substantial efficiency gains. Operating expenses actually declined 11% year-over-year, set that against the 45% revenue growth, and you have the cleanest single measure of how this business has been transformed over the past 12 months. A contributor to efficiency is the internal deployment of AI across our organization. This is changing the cost curve of the software development for us, with over 90% of our developers now use cloud code and other solutions in their workflow. This becomes structurally more valuable as our software footprint expands with the addition of Flyme. On a sequential basis, we realized improvements across almost every key metric, revenue up, cost down, profitability increased. The only exception is the adjusted EBITDA, which remained positive but was down from $4 million last quarter to $0.5 million in Q2 and improved annually by $30.2 million. To explain in more detail, last quarter's adjusted EBITDA of $4 million included that $14 million of partial monetization of our shareholdings in SiEngine, which was a onetime item. And there was no similar onetime item this quarter. We are very pleased to have delivered our fourth consecutive quarters of positive EBITDA, which is a testament both to the recovery in the market that we guided to at Q1 and our robust cost discipline. Our confidence going into the second half of the year rests on 3 things. The first is the launch cadence Peter described it, which leaned heavily towards the second half of this year. And the second is the order backlog underpinning those programs. The third is the historical seasonality of our business, where the second half was consistently carried the largest shares of annual revenue. With that confidence in mind, we are reaffirming our full year 2026 revenue guidance of $1 billion to $1.1 billion. In summary, the second quarter delivered the rebounds that we guided to in April. Our cost structure continues to improve, and we have added materially to the strategic assets of this business to drive growth. We remain focused on the disciplined execution and creating long-term value for our shareholders. With that, I will hand back to Ziyu for his closing remarks.
Ziyu Shen: Thank you, Dylan. As you've heard today, we have made meaningful progress across our strategic priorities for 2026. This positions us for growth in the near and long term. In the first half of the year, we entered into an agreement to expand our capabilities with the addition of the Flyme business portfolio. We extended our global reach with the Volkswagen commercial build-out, and we delivered a strong financial result on both top line and gross profit against a challenging backdrop. And I'd now like to open the call for questions. Operator, please open the line.
Operator: [Operator Instructions] We will now take our first question from the line of Wei Huang from Deutsche Bank.
Huang Wei: So first, I would like to ask a bit about our gross margin. You have guided in 1Q that -- and in this quarter that our memory prices are going to be pressuring on gross margin this year. But 2Q hardware gross margin was quite strong, actually at 15%, even though memory prices also increased quite a bit this quarter as well. What is the outlook for the second half of the year?
Dylan Jeng: Yes. Thanks. We have reiterated our comments about the margins at sets of earnings. And obviously, the higher memory cost to support structurally the higher revenue as the balance is memory cost with our customers. However, this sort of a pass-through does come at the understandably lower margin. And we do the right things, and it doesn't really indicate any negative about this business. And in terms of the Q2, you're right about this was a strong performance, and that was really driven by a few things. And first, with any cost increase in components and some of the -- which that we already have in stock. And there are timing discrepancy between the purchase and purchasing and the passing through. And also the selling higher value products than we did last year. So Pikes and Antora is up significantly, which support both revenues and revenue quality. And very importantly, we also did a great job on managing costs in the business, and we reduced our operating costs year-over-year despite growing revenue by 45%. So I think with all the elements and the reasons that really helped during the second quarter.
Huang Wei: Just a follow-up on that. So can I assume that our memory purchase inventory to be depleted and I guess the memory price hike is going to hit us more in the third quarter and fourth quarter?
Mark Hankinson: Wei, I'm sorry. The question wasn't very clear. Would you mind repeating it, please?
Huang Wei: Yes, no problem. You stated that one of the factors that led to the strong hardware gross margin this quarter was the memory that we had in stock. So I assume as this gets depleted, our memory is going to be -- our margin is going to be more under pressure in the third quarter and fourth quarter.
Dylan Jeng: Yes, it is. And we are very working closely our supply chain teams and also working very closely with the marketplace. So we'll continue to manage that going forward, which we do anticipate.
Ziyu Shen: Yes. So this is Ziyu speaking, sorry. I'll jump in here. So I will say our supply chain team had a great job. So we built a very strong strategic partnership with [ XT ] and also Samsung. So we are our partner with them. So from a memory supply point of view, we are very leading. And we had a very strong inventory and future pipeline. Also principally, I want to say confidently that most of -- I think most of the increase actually pass over to the customer. So no impact on our gross margin, that's for sure. But we will strongly maintain our good supply chain operation to sustainably support our customers. That's our very strong advantage in market. That's all clear?
Huang Wei: Understood. Yes, very clear. And then the second question is on our high-end and current Pikes, you mentioned it increased quite a bit sequentially as well. Did you have a number for what percent of our volume was in the first quarter? Because you said 2Q was 42%. I want to do a comparison [ Q-o-Q ] basis.
Mark Hankinson: Yes. I don't know if we have that number at hand. We can come back to you on that. I think we did talk about it at the Q1s, but let us confirm. I don't think we have that number at hand.
Peter W. Cirino: Wei, maybe I'll just make some comments. I mean we see very good traction on these 2 product lines, and they continue to roll out across multiple customers for us in China and in the global market. And I think there are 2 strong lighthouse project programs for the organization. So as I mentioned in my comments, Antora saw a 52% increase year-on-year. And I think on a year-on-year basis and on a quarter-on-quarter basis, I do suspect we'll continue to see increases in that platform. Pikes was just launched last year. So we saw over a 2,000% growth year-on-year. And again, I'm very confident it will continue to grow. We're offering a great user experience to our customers on those 2 platforms. And I think they are solid performers for us in the market and show exceptional technology leadership.
Huang Wei: Well noted. And I assume our improved product mix is also one of the reasons our ASP has increased to roughly around [ USD 360 ] in the second quarter. Do you have an idea on what's a reasonable level to assume for the third quarter and fourth quarter, assuming with the new product launches and those models upgrading to these new platforms, how much higher can this go?
Mark Hankinson: Wei, I'm sorry, the question again was quite muffled. Would you mind repeating it?
Huang Wei: Yes, no problem. I wanted to ask about our ASP outlook for 3Q and 4Q since it reached around [ USD 360 ] in the second quarter due to, I assume, a higher shipments of Antora and Pikes. And during the new model launches and the old models upgrading their chips to the newer platform, how much higher can this ASP go?
Peter W. Cirino: Yes. So Wei, I don't think we'll announce a specific number on that, but that's a trend that you should see from us, I think, and you see from most of the industry who is investing in new platforms and delivering the high-end user experience, you'll see that as just an industry trend, which will feel as a tailwind. So as our older products roll off and these newer higher-performance products roll on, and the customer experience is actually a net decrease in their vehicle architecture cost because more functions go on to these platforms, more of the vehicle becomes software-defined. It actually enables the automaker to deploy additional features inside their vehicle environment. But with a higher performance computer in the car and higher performance software, we definitely will -- we should continue to see, I would say, an increase in ASP.
Huang Wei: Well noted. And my last question is on our software license as well as the service sector business. So gross margin for these 2 actually declined sequentially for the second quarter. Software license went to almost 0 breakeven and service gross margin declined as well. Do you have anything to highlight that contributed to this or just normal business seasonality?
Mark Hankinson: Wei, perhaps I comment on that. I mean the software line item in our financials is one that gets a lot of attention because it moves up and down by a big percentage every quarter, but it's a very, very small number. And the way that we think about that is that's generally around, say, $1 million to $2 million a quarter, except when there's a significant event in the quarter. We saw that in Q1 2025, for instance. But we encourage people not to focus too much on the movement within software. I think it's important to understand as well that a lot of what people might think of as software comes into our services line item as well. I think the movements around margin on those are just general business dynamics as the quarters move over. It's principally driven by new product launches and being specified on platform.
Operator: [Operator Instructions] There are no further questions at this time. I would now like to turn the conference back to Mark Hankinson for closing remarks.
Mark Hankinson: Thanks very much, operator, and thanks for joining today. The second quarter clearly reflected strong execution. We saw this demonstrated through our financial performance, and we saw it in progress against our strategic objectives. ECARX is positioned to become a leading global supplier of innovative next-generation solutions for OEMs, and we look forward to providing more updates on our progress in the second half of the year. So thank you very much. And with that, we'll conclude the call.
Operator: This concludes today's conference call. Thank you for participating. You may now disconnect.
