Peter W. Cirino
Analyst · Zacks Small-Cap Research
Thank you, Ziyu. Good morning, everyone. As we noted last quarter, our 2026 model launch cadence leaned heavily towards the second half of the year, and every launch pulls hardware volume and associated engineering revenue with it. The third quarter delivered on that expectation with new launches landing at scale and shipments continuing to build. Shipments in the third quarter were approximately 721,000 units, an increase of 31% quarter-over-quarter and 8% year-over-year. We continue to see the quality of our revenue improve as our customers increasingly opt for our higher-end performance products such as Antora, which grew 20% year-on-year. While shipments of pikes was relatively flat year-over-year, this does not indicate a reduction in sales, rather a project-based shift in demand from our partners. As I said last quarter, this business will continue to show quarter-to-quarter variability, and we continue to encourage you to look at the trailing 4 quarters rather than any single one. This quarter, partner programs using our entry-level solutions added significant volume across our portfolio, while our high-end solutions continue to grow in absolute terms. Over time, we expect our mix to keep moving forward toward higher-value, fully ECARX architected solutions. The key here is that the underlying demand remains healthy. We do not foresee any negative impact on customer demand beyond normal seasonality and our competitive position and program execution remains strong. Our customer base continues to broaden alongside our shipment growth. During the quarter, we began mass production for 21 new models across 5 brands, 6 of which are designated for markets outside of China, such as Korea. Our technology is now deployed in more than 12 million vehicles across 18 OEMs and 28 brands and 159 markets, with more than 1.4 million new vehicle deployments in the first 9 months of this year alone. That scale matters. Every vehicle on the road shows that our solutions perform reliably across different brands, platforms and markets. And that track record is what opens the door to new partners. We are seeing this clearly in our global partnerships. Our partnership with Volkswagen Group remains firmly on track as we build out the engineering, supply chain and support infrastructure across Latin America ahead of the expected launch in 2027. We are also in the final stages of signing a partnership with a leading European luxury automaker to establish a joint cockpit AI innovation lab, which we expect to close in the near future. Using Flyme OS, the lab will develop next-generation AI cockpits customized for the automakers' vehicles in China. As our first partnership with a luxury automaker outside of the Geely ecosystem, this innovation lab serves as an important proof of concept. It shows how we can co-develop custom AI cockpits with global luxury automakers, delivering a world-class luxury product and expanding the range of solutions we can deploy with other partners. Owning the Flyme layer is central to our strategy. Completing the acquisition changed 3 critical things for us operationally. The first is a fully integrated system. Our Cloud Peak software stack handles the underlying cross-domain middleware, while Flyme Auto acts as the application and interaction layer, delivering the user interface, visual design and cross-device experience. Flyme Auto also supports over-the-air updates, which means the vehicle keeps improving long after it leaves the factory. That is what a software-defined vehicle looks like in practice. The second is cross-selling. A fully integrated hardware and software offering strengthens our competitive position with every automaker we work with. It creates more opportunities to sell across our existing customer base. Third is reoccurring revenue. Flyme can be monetized through software licensing, app ecosystems and cockpit deployments. These revenue streams are not tied to hardware volume and can carry higher margins than hardware sales. As Ziyu mentioned, Flyme also reaches beyond the vehicle. Building on Flyme's existing Super Aicy AI assistant, Tencent Cloud's WorkBuddy brings the Agentic AI to smartphones, AI devices and Flyme Auto into the vehicle. Users simply describe a task in natural language and WorkBuddy completes it with consistent experience across every device. If software is how we connect with users, silicon is what powers the experience. Full integration of hardware and software from the user interface right down to the silicon is what drives the performance and efficiency our customers demand. We can already see how we are reinforcing the foundation of this long-term strategy with Arm, with whom we've recently signed an MOU to develop next-generation AI computing platforms for automotive applications. This collaboration is an important validation of the technology behind our software-defined vehicles, which we have built over many years. As intelligent mobility continues to evolve beyond software-defined vehicles toward increasingly AI-driven experiences, the ability to tightly integrate silicon, software and user-centric applications will differentiate us in the market. Working with a global technology leader such as Arm reflects the capabilities we have developed across the full technology stack and our ability to translate advanced technologies into commercial solutions for the end user. Strategically, this collaboration further advances ECARX's position at the intersection of silicon software and AI. It expands our opportunity to set beyond our core automotive markets while reinforcing the value of our vertically integrated approach, which spans hardware, operating systems, applications and user experiences. Just as importantly, it demonstrates the growing recognition of ECARX as a trusted technology partner capable of delivering complex full stack solutions at scale. Our work with the ARM ecosystem dates back to the incubation of SiEngine. SiEngine continues to prove our capabilities in silicon. The 7-nanometer SE1000 chip already in production in our Antora solution and the 5-nanometer SE2000 chip now in active development were co-developed with SiEngine. The SE2000 integrates native large language model capabilities and higher AI compute and is custom built for our next-generation Antora computing platforms and for Flyme Auto. This is critical because the silicon and software that power our vehicles can also power other intelligent devices. Robotics and other physical AI applications need many of the same components as a modern vehicle, which are high-performance computing at the edge, low latency and high efficiency. With a stack that is already proven at scale in vehicles, we are well positioned to extend into those markets over time. Looking ahead, our pipeline remains strong. We are highly engaged with both existing and prospective customers, and we are encouraged by the level of commercial activity. While we are not in a position to disclose specific opportunities today, multiple partnership discussions and potential programs are advancing through various stages of development, which we believe will further accelerate our expansion and diversify our revenue streams. Our focus is on converting this engagement into sustainable long-term growth. I want to leave you with one final thought. Last quarter, I said that very few companies can offer a closely integrated stack from silicon to software to sensors. We built that stack for the vehicle, and it is now proven in more than 12 million of them. What this quarter shows is that, that same stack can bring AI from the cloud to the edge across a much wider range of intelligent devices. That is the opportunity ahead of us, and we are building towards it from a position of strength. With that, I will turn the call over to Dylan.