XPENG-W (09868) Q1 2026 Earnings Call Transcript

Hello, ladies and gentlemen. Thank you for standing by for the first quarter 2026 earnings conference call for X1 Inc. At this time, all participants are in listen-only mode. After management's remarks, there will be a question and answer session. Today's conference call is being recorded. I will now turn the call over to your host, Mr. Alex E., Head of Investor Relations and Capital Markets of the company. Please go ahead, Alex.

Hello, everyone, and welcome to EXPON's First Quarter 2026 Earnings Conference Call. Our financial and operating results were issued by Newswire Services early today and available online. You can also view the earnings press release by visiting the IR section of our website at ir.xiaong.com. Participants on today's call from our management will include a co-founder, chairman, and CEO, Mr. He Chang-Hung, Vice Chairman and President, Dr. Brian Gu, Vice President, Mr. Charles Chang, Vice President of Finance and Accounting, Mr. James Wu, and myself. Management will begin with prepared remarks, and the call will conclude with a Q&A session. A webcast replay of this call will be available in the IR section of our website. Before we continue, please know that today's discussion will contain four looking statements made under the safe harbor provisions of the U.S. Private Securities Litigation Reform Act of 1995. Four-looking statements involve inherent risks and certainties. As such, the company's results may be materially different from the views expressed today. Third information regarding these and other risks and certainties is included in the relevant public filings of the company as filed with the U.S. Securities and Exchange Commission. The company does not assume any obligation to update any full-looking statement except as required under applicable law. Please also note that EXPO's earnings press release and Z's conference call include the disclosure of un-audited GAAP financials as well as un-audited non-GAAP financial measures. EXPO's earnings press release contains a reconciliation of the unaudited non-GAAP measures to the unaudited GAAP measures. I will now turn the call over to our co-founder, chairman, and CEO, Mr. He Xiaophang.

Please go ahead. 智能电动车的业务将会高速增长 持续贡献可观的盈利和强劲的现金 支撑我们面向物理AI的研发投论 那么在今天我相信物理AI应用正在处于从量产落地 到规模爆发式增长的前沿 我们已经验证了Skilling Law 同样在物理AI包括自动架持跟基权领域是适用的 我们现在通过加大工程和研发投入突破 上线的路径已经变得了无比清晰 所以在这个时间点 我们会选择坚定的加大在AI相关的研发下注物理AI 我相信物理AI应用会在未来十年是全球性的战略机会 我会带领团队在今年完成Robotaxi和能行基层的量产 Hello everyone.

In the third quarter of 2026, we formally changed our official Chinese name from X-Pen Motor to X-Pen Group, reflecting X-Pen's transformation from a smart EV company to a physical AI company. Within the X-Bank Group ecosystem, our smart EV business will drive rapid growth, consistently contributing substantial profitability and robust cash flow. Today, physical AI applications stand on the cusp of transforming from mass production deployment to explosive growth at scale. We have proven that scaling law holds true in both autonomous driving and robotics, making the path to breaking technological ceilings through accelerated R&D investment undeniably That is why, at this pivotal moment, we choose to bet firmly on physical AI with increasing R&D on AI. I believe physical AI applications represent one of the most significant global strategic opportunities of the next decade. This year, I'll lead the team to bring robo-taxes and humanoid robots into mass production while also building the commercial ecosystem around them. Our goal is to turn our leadership in physical AI technologies, including our mixed-gen intelligent-assisted driving system, into a powerful new engine for revenue and profit growth, and ultimately create substantial commercial value. 仅仅关注规模 我们更关注是销量跟经营质量的平衡 并且长期如何发展 我非常有信心在今年接下来的每个季度的销量都会环比大幅度增长

那么其中我预计二季度交付量为10万到10万6千台 环比增长约60% 从GX车型开始 我们有四款全新的SUV车型 将会在最近六个月上市并且交付 那么这几款车型从产品定义的初始 就是全球车型 我相信小鹏将迎来历年来最强劲的销量增长曲线

In the first quarter, amid broad market volatility in China's domestic new energy vehicle market, we delivered a total of 62,682 vehicles. Even in a market downturn, our focus extends beyond scale. We also place greater emphasis on balancing delivery volume with operating quality, maintaining a long-term perspective. I'm very confident that deliveries will grow substantially quarter over quarter in each of the remaining quarters of the year. I expect Q2 deliveries to reach between 100,000 and 106,000 units, reflecting quarter-over-quarter growth of over 60%. Starting with the GX, we plan to launch and begin deliveries of four new SUV models within the next six months. These models have been defined and designed from day one as global vehicles.

I believe XPeng is entering the strongest delivery growth trajectory in our history. 今年4月份,2026款小鹏的MORA M03上市,我们推出来搭载图灵芯片的全新的MAX版本,以及支持第二代VLA的OCHR SE版本,这标志着我们完成的所有车型向图灵芯片的升级。 So we will see from the chip to AI to the entire field of the software and the technology of the M03 We also have over 85% of the users

to choose Max and O2 SE The M03 is the M03 of the M03 of 20 years is a solid-rated win In April 2026, we launched the 2026 X-PEN MONA M03, including a new MAX version powered by our Turing AI SOC and an Ultra SE version supporting the LA 2.0. This also marks the completion of the Turing AI SOC upgrade across our entire model lineup. Through our four-step in-house R&D capabilities from SOC to ALI, we have made advanced computing power and technology more accessible. As a result, more than 85% of Mona MO3's customers chose the Max or Ultra SE version. The Mona M03 has remained China's top-selling A-class pure electric sedan for 19 consecutive months. With its leading technology and stylish design rarely seen in the A-class segment, Mona has become the brand of choice for young users. We're confident that the Mona lineup will continue to achieve sustained success, not only in China, but also in overseas markets.

It is a large old-time L4 R2 testở by the other is the second-time给 its brands. It is the first to be built for electronic and use of the Q3 of a robo-tiko-tiko. It will become the U-V market system for the new worldwide. 我们已经提升GX的产能 让用户更早提升 同时在今年下半年我们上市的三款车型 都将搭载基于图林芯片的第二代VIA 并且具备一车双冷 同时面向全球发售 随着GS跟后续三款新车的规模交付 我相信小鹏集团的季度交付量 环比会大幅度增长 那么特别开心的是 GS成功于小鹏集团品牌向上的关键一步

On May 20th, we launched a GX, a flexion model built for the L4 era. It is also China's first pre-installed mass-produced RoboTaxi model with full hardware redundancy, representing a new starting point for us to continuously expand our share in the SUV market. Among the initial firm orders for the GX, the ultra-flagship train price above 350,000 RMB accounted for over 80%, making the GX one of the most sought-after products in the premium luxury vehicle market. The GX's luxury is defined not only by its design, but more importantly, by its leading technology and ultimate safety standards, setting a new benchmark for tech-defined luxury. I'll spare no effort to work with our supply chain partners to ramp up GX production capacity and deliver vehicles to customers sooner. The three new models launching in the second half of this year will all be equipped with touring SOC Power VLA 2.0, featuring one-vehicle dual-energy capabilities, and will be launched for global markets. With deliveries of the GX and the upcoming three new models ramping up, I'm confident that X-Lang Group's quarterly delivery volume will grow significantly quarter over quarter. The success of the GX must be a key step in elevating the X-Lang Group brand.

In the second half of this year, we'll build on this momentum by launching a series of technology products at high price points, including humanoid robots and flying cars, further strengthening our brand equity and profitability. 在今年下半年随着我们四款全球化新车型的交付 并且小峰后面所有的新车型都是全球车 我们的目标四季度实现海外月销持续过万 在全年海外销量是实现翻倍以上的增长 那么因为我们从今年起发布的后面的所有车型都已经考虑了国际市场

所以我们在将来海外和国内上市的时间差将会大幅度缩短 实现海外销量和交付的加速度 我们希望在海外市场被视为长期增根本地化经营的企业 并且与当地的用户跟合作伙伴经历更强的连接 从去年开始小鹏在海外已经建立了三个本地化生产基地 In 2026, we are accelerating international expansion on the back of a growing lineup of high-quality technology-led products and deepening localized operations. The P7 Plus's overseas delivery launch in April pushed our monthly international deliveries above 6,000 units for the first time. Starting in Q2, international revenue is expected to exceed 20% of total revenue. In the second half of this year, we plan to introduce four models for global markets. Starting with the GX, every upcoming X-Men model will be built as a global vehicle. Our target is to achieve sustained monthly overseas deliveries of over 10,000 units in the fourth quarter and to more than double full-year overseas deliveries. Starting this year, by embedding international market needs and certification requirements early in the vehicle deployment process, we will significantly shorten the time gap between domestic and overseas launches and accelerate our overseas sales momentum. Our goal is to be recognized as a company committed to long-term localized operations in every market we operate in, building stronger ties with customers and partners across each To that end, we have established three localized production bases overseas since last year, and our Munich R&D Center has become our fastest growing research hub. 用户选择小工最重要的理由之一

也在市场上建立起新的用户薪资的壁垒 我认为第二代VR让我们看到了两个巨大的变化 第一个就是从原来的端到端的范式 进入到全新的跟物理事件模型相关的范式的认证的成功 第二个点就是用户从可以用自能辅助驾驶 到将来可能可以变成要用好用的一个全新的变化 所以在第二代VIA的第一个版本的成功 就验证了整个的通过放大数据 通过扩大模型的参数量所带来的能力飞跃 我们也看到了物理世界AI的Skylnore 在后面我们会继续加码投入Skylnore 在今年三季度我们计划推出了新版本 会非常大幅度地提高模型能力的上限 加速扩大我们在自动驾驶的领先优势 我相信第二代VIA能力上限高 不使用高精力图 泛滑能力极强 我们一开始就考虑了如何部署到海外食料 并且支持前装量产的Robotizing 我们跟自动驾驶团队的目标是就到 市场上绝对的第一 并且走向真正的L4 In April, ADAS mileage penetration on VLA 2.0 equipped Xpeng vehicles surpassed 50% for the first time,

signaling that advanced intelligent driving is becoming a cool must-have feature for users. VLA 2.0, with its generational lead in intelligent driving, has become a key reason customers choose Xpeng, creating a strong and lasting user mindshare mode in the market. The success of VLA 2.0's first version also reinforces our belief that scaling data and model parameters can drive meaningful breakthroughs in real-world AI capability, strengthening our conviction in the scaling law for physical AI. We're set to accelerate our investment in scaling up. With the upcoming release scheduled for Q3 this year, will substantially elevate our upper limit of our model's performance, further widening our lead in industry. The LA 2.0 features a high-capability ceiling, operates with our HD maps, offers exceptional ability to generalize, enabling rapid deployment across overseas markets, while supporting pre-installed, mass-produced Robotexies at scale. Our goal is to become the undisputed number one in the domestic market and take a critical step towards true global leadership in L4 autonomous driving, including RoboTaxi. 开始在多个国家获得监管许可,并且逐步推送给我们的客户。

经过我们近期在全球客户的调用,我们发现一个非常大的变化就是全球用户对高等级的智能主持驾驶系统,包括例如VoA和全场景的多语言本地对话系统,例如VoA的期盼远高过我们最开始的判断。 In the next half of the year of 2016,小空 will be able to enable China's cars to be able to move forward and move forward. In this process, I see that no matter whether it's 2B or 2C, there are huge companies.

I'm pleased to see the accelerated rollout of the UN-led DCAS regulatory framework. With Europe, our most important international market, now beginning to open certification pathways, for high-level ADAS. Our VLA 2.0 is currently being tested in Europe, and we hope to receive regulatory approval in multiple countries next year, allowing us to deliver the technology to overseas customers and begin the global generalization of our VLA 2.0. Our recent research shows that global users' demand for assisted smart driving system, VLA, and full-scenario multilingual conversational system, VLM, are far stronger than we had anticipated. In the second half of 2026, XPeng will lead and accelerate the intelligent transformation of China's automotive exports. In the process, I see substantial commercial opportunities emerging on both the B2B and B2C front.

Thank you.

Our fully redundant GX feed is already undergoing L4 testing on topic roads in Guangzhou. We aim to launch pilot passenger operations for our RoboTaxi service in the city in Q3. The GX's L4 full redundancy hardware and software are decoupled from the vehicle platform, making them deployable across our entire lineup, including the Mono series. I believe our RoboTaxi offering has a clear edge over incumbent RoboTaxi companies in terms of ability to generalize, cost efficiency, and scalability. These advantages position us to build a multi-stateholder ecosystem where operating partners and X-Peng work together to create and share commercial values. Following the overseas rollout of VLA 2.0, we'll also actively pursue bringing X-Peng's cost-effective robotech supervision to overseas and domestic markets. 同时小空能型继承的Ilon面向量产版本的软硬件研发目前进的都顺利 即将进入ETR的软硬件核位阶段 那么量产版的Ilon我们是按照车规矩的安全可靠性来进行研发和设计的 我们有一大批优秀的汽车零部件合作伙伴 也成为Ilon零部件的合作伙伴

我们在最近完成了自研零巧手的开发 新一代零巧手不仅非常灵活 而且成本大幅度下降 我们为训练机前的大老小老等 搭建了多个维度的数据体系跟模型 当前的数据量正在快速增加 模型的训练效果显出提高 小鹏可能是国内唯一一个全战乱硬件 都自研的机前公司 从芯片到物理大模型 从数据的生产到模型的预训链和后训链 从关节到零巧手 从全新一代的运控到VLA,VLM,VLT 深度软硬件资源跨域进行融合创新 这才使艾朗具有更高颜值和高组织以及全面可量产的能力 我期待在三季度会让大家看到新一代 埃朗连机器人的相关的演示 他不仅能够多元沟通 具有理人的全运动姿态 还可以逐步开始执行专业的工作 我们目标是在今年年底前实现埃朗的量产

并且首先我们在小鹏的门店进行私商用 在明年我们会对中国和海外的商业客户进行交付 The software and hardware development for our mass production humanoid robot IRON is progressing smoothly and is about to enter the ET2 software hardware integration stage. The mass production version of IRON will be built to automotive-grade safety and reliability standards. And many of our existing automotive supply chain partners have also become component partners or suppliers for IRON. We have also recently completed development of our proprietary next-generation Dexterous PAN, which is significantly more agile and substantially lower in cost. We have built-in multidimensional data systems to train our own brain and cerebellum models, with our training data and scaling rapidly and outcomes improving significantly. Xpeng is the only robotics company in China with full-stack in-house R&D capabilities of both hard-end software, spanning SOCs to physical AI foundation models, data generation to pre-training and post-training, joints to destitute hands, and next-generation motion control to VLA and VLM. Through deep in-house software and hardware R&D and cross-domain innovation, IRON will deliver a more refined design, higher quality, and more comprehensive capabilities. I look forward to showcasing the next-generation IRON in the third quarter, featuring multilingual communication, human-like full-body motion, and gradually autonomous execution of professional tasks. We are targeting to achieve mass production of RM by year-end, with its initial trial commercial deployment in X-lang showrooms, followed by commercial customer deliveries in China and overseas next year. I believe that once human or robots reach mass production, the data flywheel will drive technology iteration and sell growth at a pace that outstrips what we saw in EVs. Starting next year, revenues from humanoid robot hardware and AI models are expected to emerge as a key driver for revenue and growth, rather grow for XM Group. Xpeng is now fully committed to advancing the mass production and global expansion of three fiscal AI applications, DLA 2.0, RoboTaxi, and Humanoid Robots. We firmly believe these three areas present enormous potential in terms of both commercial scale and investment returns. Looking at our roadmap, the B2B market will be the first to take off, while international markets will generate greater commercial returns than domestic markets. Backed by the deep experience we've gained through our partnership with Volkswagen, along with a business model already proven through mass production, Xpeng is well positioned to execute this next phase of growth. We are making the large-scale commercialization of physical AI a company-wide strategic priority and will move decisively towards large-scale deployment.

Thank you very much. 小红将会不断冲击更高的效率目标 与此同时 我们的经营质量也正在显著改善 我们的付款涨期在新造设室里面处于领先水平 我们的毛利率 面对成本的压力保持嫩轻 汽车业务的规模效应会持续增强 我期待小红集团通过物理AI技术的决定领先 构建规模化 网络化

以及以群效应的全新的商业模型 在未来 汽车 罗伯塔斯 机权 都会是差异化巨大的物理AI质量 无论是硬件的销售规模 还是AI模型的经常业收入 都将高速成长 那我们在AI研发投资产生 极其丰厚的回报 For the second quarter of 2026 We expect deliveries of 100,000 to 106,000 units up 59.5% to 69.1% quarter over quarter, with revenue of 19.6 billion RMB to 20.8 billion RMB, up 50.4% to 59.6% quarter over quarter. Having passed the seasonal trial, we are entering a period of strong growth driven by four new models, increasing production capacity and expanding international business in the third and fourth quarters will continuously strive for higher sales target concurrently our operational quality will improve significantly our supplier payment terms remain at an industry-leading level growth margin demonstrates strong resilience against cost pressures and economies of scale in our EV business continues to strengthen i expect that the expand group to build a an entirely new business model anchored by our absolute leadership in fiscal ai technology with scale and network effects both expand vehicles robo taxis and humanoids will become highly differentiated fiscal ai agents moving forward both the hardware sales scale and recurring AI model revenue are poised for high-speed expansion, allotting immense returns on our AI R&B capital.

Thank you, everyone.

With that, I'll now turn the call over to our VP of Finance, James, who will walk you through our financial performance for the first quarter of 2026.

Thank you, Xiaobeng. Now, let me provide a brief overview of our financial results for the first quarter of 2026. I'll reference R&B only in my discussion today, unless otherwise stated. Our total revenues were $13.03 billion for the first quarter of 2026, a decrease of 17.6% year-over-year, and a decrease of 41.4% quarter-over-quarter. Revenues from vehicle sales were $11 billion for the first quarter of 2026, a decrease of 23.5% year-over-year, and a decrease of 42.3% quarter-over-quarter. The year-over-year and quarter-over-quarter decreases were mainly attributable to lower vehicle deliveries. Revenues from services and others were $2.03 billion for the first quarter of 2026, representing an increase of 41.2% year-over-year and a decrease of 36.1% quarter-over-quarter. The year-over-year increase was primarily attributable to the increased revenues from technical R&D services and parts and accessory sales. The quarter-over-quarter decrease was primarily due to the reduction of technical R&D services revenues following a significant milestone catch-up in a prior quarter, as well as no revenue contribution from carbon credit trading in a current quarter. Gross margin was 20.6% for the first quarter of 2026, compared with 15.6% for the same period of 2025 and 21.3% for the first quarter of 2025. Vehicle margin was 12.1% for the first quarter of 2026, compared with 10.5% for the same period of 2025 and 13% for the fourth quarter of 2025. The year-over-year increase was primarily attributable to the cost reduction and improvement in product mix of models. The quarter-over-quarter decrease was due to higher unit vehicle costs, resulting from increased memory chip and battery-related costs. R&D expenses were $2.91 billion for the first quarter of 2026, representing an increase of 46.8% year-over-year and an increase of 1.1% quarter-over-quarter. The year-over-year increase was mainly due to higher expenses related to the development of new vehicle models and AI-related technologies as the company expanded its product portfolio to support the future growth. SG&A expenses were $1.88 billion for the first quarter of 2026, representing a decrease of 3.2% year-over-year and a decrease of 32.5% quarter-over-quarter. The year-over-year and quarter-over-quarter decreases were primarily due to the lower commission to the franchise stores. As a result of the foregoing, loss from operations was $1.87 billion for the first quarter of 2026, compared with $1.04 billion year-over-year and $0.04 billion quarter-over-quarter. Net loss was $1.78 billion for the first quarter of 2026, compared with net loss of $0.66 billion year-over-year and net profit of $0.38 billion quarter-over-quarter. As of March 31, 2026, our cash position was $42.09 billion. To be mindful of the length of our earnings call, I would encourage listeners to refer to our earnings press release for more details on first quarter 2026 financial results. This concludes our prepared remarks. We will now open the call to questions. Operator, please go ahead.

Thank you. If you wish to cancel your request, please press star 2. If you are on a specific phone, please pick up the handset to ask your questions. For the benefit of all participants on today's call, if you wish to ask your question to management in Chinese, please immediately repeat your question in English. For the sake of clarity and order, please ask one question at a time. Management will respond, then feel free to follow up with your next question. Your first question comes from Tim Hiziao with Morgan Stanley. Please go ahead.

Hi, I'm Morgan Stanley of Tim. Thank you for your question and congratulations to the company's new on-screen GX. I have two questions. The first question is about the new on-screen GX. My first question is about GX, because we noticed the new model has been selling very well since its launch on May 20th. I could imagine you can share the current order book, Your Steady State Sales Ball and Target, and how we should think about the vehicle growth margin. Thank you.

The situation is still in the middle of the day. So to move on to the point of view, at the beginning of our旗舰 version, the 8成, the 8成, the 8成, the 8成, the 8成. But we've seen some interesting data, including, like, like, like, Thank you very much. 三四万以上市场的头部的报款 那么说实话 我们第一正在全力推动供应链的支持与产能 但是我们也非常期望能够推出高品质的整个的服务 跟质量 能够稳定GS在长期的交付的数据 所以换个角度 我对GS的持续性是最关注的 那么GS因为作为小工的一个旗舰 实际上毛利是非常不错的 When I was talking to my friends with the media, I said that only one SKU is better than one SKU is better than one SKU is better than one SKU.

Thank you for the question.

Honestly, the performance of GX sales is above our expectations. It's been performing very low. We also observed some interesting data regarding the sales numbers of GX as well. So currently, if you look at the BEV flagship version, right now the lead time has surpassed 30 weeks. Even under the expectation of converting more to the ultra version, it's still growing very, very fast. On the other hand, our flagship model with initial orders over 80% of the total orders continue to grow in the mix as well. We also observed something very interesting. For example, the max version is right now less than 5% of the mix, which is a little bit lower than our expectation. An extended range version initially actually was performing less popular than our BEV version, but right now it's approaching the level of BEV popularity, especially with our increased promotion and marketing campaigns in the western and northern part of China, which is very, very promising. We definitely have high hopes for GX, for it to be one of the top-selling vehicle models, is the above $300,000 on the price range market for NDVs, high-end luxury, big six-seater. And to be honest, we are working very closely with our supply chain partners to support the capacity ramp-up, and we definitely want our top-quality products to come with top-quality service and after-sales services as well to support this sustainable GX delivery. Our priority right now not only is about the deliveries of the product, but also the sustainability for GX being a flexion model of X Bank Group. Now, being an flexion model, obviously the GP margin of GX is quite good. As I mentioned earlier with the media, this GX, only one SKU under the GX series is performing with a lower GP margin than our expectation. But the majority of SKUs, of GX, is actually having a higher than expectation kind of GP margin performance. Thank you. I also would like to add that starting this new year, all of our new vehicles, when it comes to their configuration and also their pricing it will be under the consideration of their commercial value so commercial consideration will be one of our key priorities starting this year and the second thing is that we are looking for long-term sustainable sales performance rather than having a big beginning and subsequent diminishing sales so we're looking for a better quality in terms of supply chain management and the modular management quality of the whole supply chain system to secure our supply and also production capacity ramp up to support our vehicle delivery. So starting with GX, we are expecting to achieve a better balance between commercial value and also scale as well as long-term stable sales.

I have a very good answer.

My second question is about Robotaxi. The second question is about RoboTaxi, could you please update us on your domestic operations? and the overseas expansion plans uh has the restaurant regulatory tightening in china had any adverse impact to the progress and how do you view the impact of that to be

global taxi business uh to the groups to see passenger vehicle sales uh that's my second question thank you thank you we看到了中国国内近期对自动家的监管的收紧 但它对我们的整体的节奏并没有任何影响 因为从我们的角度来看 从我角度来看 诺博特斯是在2028年和之后会是一个巨大的全新的商业机会 那么在现在小鹏正在国内和全球来尝试把它跑通 所以我们整个的节奏是把VA2.0在中国和海外落地并且做好 我们在中国会通过当前车进行快速的研发跟实验 并且在2027年小红会推出经济型的车 来做好ロボタシー的示范和商业验证 小红的ロボタシー我相信有两点 我觉得第一点我们只做产品并且抽拥 我们不愿意直接做运营 所以我们会有非常多的合作伙伴 第二点小红的ロボタシー主力做全球

Thank you for the question, and thank you for mentioning the recent tightening of the autonomous vehicle regulation in China. However, it hasn't placed any adversity to our rhythm of development yet, because from our perspective, after 2028, we're going to expect a huge commercial opportunity for RoboTexi. And right now, we are doing step-by-step to prepare for that opportunity, both in and out of China. And so our current plan is that we are going to do a lot of deployment and trial within China, as well as the international market with VLA 2.0. And we are confident that we can achieve success. In China, we are going to conduct, and we are actually in the midst of conducting rapid research and development experimentation using our current vehicle models. And in 2027, we're going to launch an economy car model to demonstrate and validate how to start the RoboTaxi business model in this area in China. And we are sure that we have the capability to achieve high-level success. And we want to clarify that with Xpeng, first of all, we want to focus on the products, and we can take a commission from working with our operational partners. We are not going to directly involve ourselves in the operation. Hence, we expect to have many, many domestic and international partners in the operation for RoboTaxi. And the second thing is that we are going to work not only in China, but also globally.

We believe that global RoboTaxi has lots and lots of commercial potential and value to tap into. 我们的想法吧 我觉得第一个 我觉得当前做Lobotesty的测试跟研发试验 实际上对C端是有正面的影响 因为在之后 小朋友VIA会提供多个不同的智能辅助加持的策略 例如急速模式 例如类似Lobotesty的这种无扫接管的模式 我认为Lobotesty之后会进入到 逐步进入到一个政策法 can be open and open and open and open and open and open and open. I think that Lobotard will be completely separate from now and C-doll. It may be a different way of driving. Thank you.

Now regarding the second question that you asked about the impact of our B2B business to the B2C business, let me just add a few of our thoughts. First of all, we believe that the current testing and R&D and experimentation actually on RoboTaxi has a positive impact on our B2C side of the business because our VLA model is going to offer all kinds of intelligent driving assistance strategies, for example, the speed mode, for example, the less human intervention mode, et cetera, that we can borrow from RoboTaxi. And I believe that for the future, the market is going to digress and actually transition into a period where policies and regulations and market situation is going to be more open with the B2B side versus B2C side being two different market segments. So that is our expectation for the future. Thank you.

Thank you.

Your next question comes from Nick Lai with JP Morgan.

Please go ahead.

Thank you. 谢谢小冯总跟华尔森 我是David Morgan的 Nick 有两个比较重要 中长期战略的问题 刚在电话会议中 小冯总花了很多时间 在讲这个VLA 2.0跟OTA的 这个中长期的这个战略 那我想就请教 今年下半年 未来的一年左右 这个VLA 2.0的几个OTA 会有什么比较重要的趋势跟变化 还有刚小冯总有提到 以后物理AI的应用跟商业化 这几个部分可以帮我们 再展开讲一讲 My first question is to talk repeatedly about the VOA 2.0 and medium and lower term strategy, including production roadmap and the OTA. And I wonder if the chairman can also share with us the major change and advantage that we can anticipate from here.

Thanks.

Thank you. NPI的接管率 那么在今年的年底 那么我们还期望 第一次我们把VLA VLM通过组织上的合并 能够能体现出来 除了自动驾驶的司机能力 更有语言沟通的能力 甚至为未来的汽车的管家能力 打下基础 我认为就是L5的初见能力 我们期望在今年年底初步

Thank you. I believe that within this year of 2026, S-Pen's VLA or VLA plus VLM capability is going to have two key models or experience two key development phases. The first phase will actually happen in Q3 August this year. During this period of time, VLA is going to have our second version, which will appear as smarter, more communicative, and also have better generalization capability with less NPI or human takeover or intervention penetration. Whereas in the past with our VLA 1.0 version, it mainly focused on the basic capabilities of safety, engineering, and also basic experience. We didn't or we were not able to increase the ceiling of the capability significantly. However, by August and Q3 this year, We're expecting a lot of significant performance improvement on that point. On the other hand, with the combination of VLA and VLM, we're expecting not just ADAS capability, but also effective language communication capability and something that approach the capability of the so-called Butler-like kind of experience in her. And so we are going to achieve gradual implementation through the OTA or three OTA releases in August this year, and also at the end of this year. Thank you. I think that with gradual R&D in development, we will be able to achieve L4 software capability on an L2 hardware in the future. And by that time, we expect to see tremendous changes in terms of the business model and all shape or form of the whole business. This is something that we are in the work of. We are not going to discuss the detail about it today.

My second question is related to humanoid robot in terms of cost advantages compared with competitors. and in 2027, Japan plans to export from a robot overseas market.

Can we also share more about that long-term strategy? 安全的系列的事情 又比如说绝大部分当前的机械不考虑可靠性 稳定性 可维修性 都是销规矩 它不是车规矩 然后很多的能力机械公司 它没有考虑如何把它speal up的规模和供应链怎么去解决良品率品质的系列的问题 以及它没有考虑到比如说这些硬件 实际上软件的底层 比如说我们的芯片 我们的EA电子电架构 它是在本地的云端 如果在云端如何解决数据隐私 流量 算力 延时等系列东西 所以换个角度 这是为什么我们从去年初开始就在可以说在能力进行的硬件 软件 除了电芯没有自研 我们基本上唯一的全占自研的机器公司 所以这才使我们能够超级集成 我认为机器人最开始的硬件的成本结构会类似一台汽车 但是会逐步跟一台在长期来说会更有 可有更规模更精细的解决方案 但是在今天,机成的成本结构跟一台车上是非常接近。 非常接近。 Thank you.

We actually encounter all kinds of differences during the mass production of humanoid robots between the structure of a robot versus an EV, especially when it comes to different aspects of consideration. Clearly, there's a lot of areas where we need to address when it comes to the mass production of humanoid robots. For example, the safety, the hardware of humanoid robots are quite different from cars, and currently the available robots in the market do not concurrently consider safety and security of the simulation and also of the humanoid robots and how humanoid robots will interact with humans across all kinds of scenarios. At the same time, the vast majority of the existing products out there do not consider reliability, stability, and maintainability of the robots as well. If you look at the level of mass-scale structural production of the humanoid robots, a lot of it is actually consumer product grade versus our expectation of being car grade safety and production quality. At the same time, a lot of our competitors do not consider the supply chain, capability development when humanoid robots achieve a certain level of mass production, how do you secure supply chain capability and safety? At the same time, if you deploy your SOC during a chip in the humanoid robot, is the data stored locally or are they on the cloud? Are they on the cloud? If we consider privacy, safety, and security, et cetera, et cetera. That's why since the beginning of last year, 2025, we had already started a series of overall of our design and also a preparation of human or robot having the full-stack hardware and software capability of human or robot allow us to actually have a more comprehensive skill set and capability preparation for human or robot to release its potential and commercial value in the future so i would say that for now except for batteries all the other parts of human or robots that we produce are in-house full-stack self-developed um so in the future for sure we believe that we can come up with a more scalable and economical solution to humanoid robots. However, today, the cost structure of humanoid robots is actually very similar to that of a car. Thank you.

Okay, thank you.

So, regarding the issue of the foreign industry, I'll add a little bit about this. Just like our cars, our cars are currently thinking about the future. Our first product is thinking about the future. 实际上机器人在全球的销售的逻辑比中国可能会更佳 因为从辅助甚至取代岗位的角度来看 在海外的能工对机器人更适合 第二 在硬件如何符合海外的政策和软件以及数据如何符合海外的政策 我们做了很多事情 比如说我举个例子吧 小鹏当前VA2每个小时的token使用量大概在两个亿 那么如果把机器人的模型放到云端 我认为大概每个小时的token可能有五到十个亿

每个小时的云端的流量费可能在100GB以上 所以他不可能是把他的算力放在通过云端的方式来完全解决 And the second point that I would like to make is about the overseas market. I think just like EV, all of our models developed nowadays are considered as global vehicles and our robots are considered as global robots as well. Potentially, if you look at the overseas market, there's actually a bigger commercial value to replace human workers with humanoid robots. So there's lots of commercial potential to tap into globally speaking. At the same time, when you look at the regulatory side of things in terms of the hardware and software and also data privacy, we also have prepared sufficiently in our humanoid robot design as well. When you look at our current VLA 2.0 usage, we're looking at 200 million users per hour that are for the VLA models outside of the cloud. And if you were to consume everything or do the computing on the cloud, you're looking at the consumption of data of 100 gigabytes per hour, which is enormous. And so we have to start from day one to consider how we incorporate that kind of data consumption or model usage in our humanoid robots that are maybe etched, defined or locally deployed as well. And so the entire system for our humanoid robots has been designed for the global market since day one. And we expect to actually see more potential progresses internationally as our domestic R&D system continues to evolve going forward.

I'm very期待 for the development of Taiwan.

Thank you. Your next question comes from Tina Howell with Goldman Sachs. Please go ahead.

Thank you, Goyal. This is my question. I'm Goyal of Tina. My first question is regarding our robot taxi business. So wondering what is our plan in terms of expanding to more cities outside Guangzhou and the timing of that. Also, accordingly, what has been the progress of our licensing approval? Thank you.

We have more partners in the country and abroad, and they have more partners in the country. In the past, we have announced that we have received a lot of international and foreign partners in the country. They will be very interested to watch our next year's整体推進. So, as soon as we have released the 4-year-old RoboTaxi in 2017, we will see that we will continue to increase in RoboTaxi.

Thank you, Sina, for your question. Our RoboTaxi deployment for exploration, I mean experimentation right now, is limited only in Guangzhou, where we already got the license off. Our plan for RoboTaxi is that we are going to test the water here in Guangzhou. And after we successfully develop the technology, the product and the business model here in Guangzhou, we then can expand our business partnership in the whole of China and also outside of China. And we expect to work with multiple partners for their own localized operation. And since we made the announcement to right now, we have received lots of inquiries and lots of interest, both in and out of China from our potential business partners. And so they are keeping a close eye on our future progress in the upcoming year. And I believe that after 2027, when we were able to launch our new economy car model for RoboTexi, after we are able to announce maybe a better total solution for RoboTexi in the future, we believe that we can actually increase our commercialization capability will see a global taxi in and out of China in the future. my second question is regarding our second quarter growth margin so on the one hand we have higher volume as well as a better model mix with GX on the other hand we have some headwind from higher raw material and some component cost so just wondering what would be the guidance for vehicle growth margin as well as the company blended gross margin. Thank you.

Hi, Tina. This is James. Yeah, so as you can see, first of all, in the first quarter of our total gross profit is pretty close to the prior quarter Q4 of last year. So, you know, in that we did see some level of cost increase as mentioned in the earlier script around the memory chip cost increase as well as the battery raw material cost. That is partially included in Q1, and we continue to expect that to be included in the following quarters in the year. You did mention that we have launched the GX, the full-size SUV, be in the second quarter, and we will start to deliver that in the following quarters as well. The GX gross profit is among the highest in our portfolio, so from a product mix perspective, we'll start to see better mix in Q2 as well as in the second half of the year. So all of that considered, we expect the Q2 total gross margin to be around the same level as Q1. So, hopefully that answers your question.

好的,非常感謝. Thank you. Your next question will come to me with Bank of America. Please go ahead.

can bring us what kind of money or a cost of money. Then, for example, we can talk about some cars in the export. If it's been a magnitude of the magnitude, we can have a lower cost, but we can have some of the company's production. Then, what kind of money will be better? This is the question of the management. So, management, could you elaborate more about your long-term overseas market growth driver, especially for your overseas road production profitability versus your export business model?

I hope in this five years, the sales of the half is half. The sales and the yield will be over half. I think that following the last half year, the four cars come out after the next year, the next year, the 2017 and 2018, the total rate will be greatly increased. This is a very good thing for me. 因为我们以前的车型主要是粘管,所以整个海外,无论是从硬件,还是从我们AI的软件,还是从我们的渠道,还是从我们的服务,包括本地化的充电能力,超充能力,还是我们的利润,我们会无论在汽车还是汽车里面都会在海外做巨大的投入。 So, in fact, this year, the development of our cars are mainly in the United States. So, I have a very trust in the United States in the future of the United States.

First of all, since you asked two questions, I'm going to address the first one first. Regarding our overseas market expansion or development, first of all, it's one of our most important four strategies, and in the coming five years, we expect to have maybe 50% of our revenue and profit coming from the overseas market. And as we approach into the second half of the year, with the launch of our four new models upcoming, we believe that we are going to be able to tap into the overseas market even more following 2027 and 2028, because in the past we mainly had two models serving the international market. And when it comes to our EV future and also robotics future, we have always considered overseas market as one important component or important market for us. So when it comes to our capability development, it doesn't matter if it's hardware, software, our distribution channels, our servicing network, including localized charging capabilities, supercharging capabilities, as well as our profitability consideration. You know, overseas has always been in the roadmap of development. And with our dedicated and committed R&D spending, starting, you know, many, many years ago, we believe that we are, you know, well-pointed to tapping to the overseas market development future, especially in the coming three years.

Hey, I mean, this is Brian. Let me just add here.

Hi. So, yeah, I mean, let me just add here. First of all, if you look at the international sales volume contribution, in the latest month, I think we already see it represent close to 20% of our volume. As you saw in last year, international sales is roughly 10% of our global volume. So you can see a significant increase in terms of the proportion that international sales is now represented in our global sales. Second point is on the probability of our international vehicle sales is significantly better, even with, obviously, some of the tariff issues we faced, some of the cost increases we saw this year. Still, I think that the international business generates significant better most profit as well as net profit contributions to our bottom line. And also, you probably saw that the new models that Xiaopo mentioned for the global markets has yet to be launched. So we anticipate that momentum will carry on throughout the year. So I think the contribution at the 20% levels will be consistent throughout the year, because obviously China, we saw significant growth expectations as well. And then in terms of really dealing with sort of tariff and other challenges, we are increasing our investment locally. I'm actually currently in our partner, Austria, right now to make sure that we have capacity to tackle the expected growth for our Europe as well as global markets. So I think it is going to be, I would say, a very exciting contributor to our overall momentum as well as profitability.

The next question comes from Paynewu, Lucidic Securities.

Please go ahead.

Thank you for taking my question. I'm Pingyue Wu, representing Pacific Security and also CFI. And my first question is regarding the overseas production. And we've seen significant progress on XPoint's overseas production localization this year. And could you share more color on what percentage of vehicles for overseas market will be produced at local plants in this year and next year? And does the localization rate vary materially by region?

Yeah, so this is Brian. Let me address that question again.

On the localization production, right now we have two plants in Southeast Asia, in Indonesia and Malaysia, mostly addressing local demand. And also we have the partnership with Magna in Austria, where we manufacture vehicles for the European market. All three of these manufacturing layouts will see increased capacity this year, as well as the new models being produced in those local markets. I would say that for the Southeastern Asian countries, it's mostly for the local markets, and obviously Austria is for Europe. So I would say the Austrian operation will be a very important capacity be contributing to our European sales. I expect the majority of our European sales will have local manufacturing at this production. And elsewhere, where I think we don't have manufacturing facilities yet, so I think it will continue to maintain the current business model. But I think as we increase our volume as well as market share in some of these large markets, We are actively looking at ways to increase our production capacity as well as localization efforts to make sure we satisfy the localization content rules as we move into a more deepened local production model.

感谢

然后我第二个问题是关于机器人 因为管理团刚才也提到了时间点 年底的时候量产 27年QE的时候进入现在门店 然后想请问能否可以举几个小例子 来帮助我们理解机器人在我们的门店 以及客户的门店的一些 客户的商业场景的一些功能 然后是否可以在 特别是在外部客户那边 介绍一下小鹏的机器人 And my second question is about humanoid robot. And the management team just mentioned that X-Con is aiming for mass production of humanoid robots by year-end and store development in first quarter 27. And could you share a few specific examples of what functions the robot will perform in F.3 retail store and also in the client application.

好的好的

And additionally, could you elaborate on what specific application scenario can we see in the external corporate customers and what's our pricing strategy? 对 我们的基础实际上跟大部分的基础不同 我们是想在一个跟能交互的环境进行使用 因此我们考虑的很多的环境就算是商业环境 我们最开始思考的是在导览导购 不管是中国还是海外 它可以介绍 比如说在我们的门店介绍我们的车 但是这个车的试驾 签单可能还是有一个人来负责 但是对公司的介绍对产品的介绍对性能的介绍等我觉得可以由机器人来帮助人来实现 那么当导览导购这个在全球已经非常多的岗位场景做得不错的形容下 我们实际上更希望是能够往这个销售的零售角度去思考 比如说一个收银员您可以看到在全球的收银员或者销售零售的岗位 在全球的岗位数量是非常之多的 所以换一个角度 小鹏在基权的体系里面 从商业化我们会聚焦在 从导购的角度的深入 但是我们会开放我们很多的合作 跟我们的合作伙伴 他们可能有不同的岗位的需求 他们可能会跟我们一起 来生成数据 来进行预训练和后训练 来生成不同的模型 那么这些模型我们会放到 我们的基权对外合作的 For example, Apple Store, you can download and download it on Apple Store.

So, let's get back to the first one. tour guides or assisted shoppers. For example, in our offline stores or dealer shops, we can have these human robots to introduce the cars and introduce our products for you. And then there will be our human salesperson that come along and do the test drive with you and sign the order or sales order for the car. But I think robots can really help with the basic information introduction and do some performance of the product introduction, et cetera, to achieve a higher efficiency. In addition to that, once that capability is more sophisticated, we then can open up more opportunities or collaboration with our partners in the ecosystem who may have different kind of various job requirements that we can work with, you know, to place on the shoulders of our humanoid robots. And we also can work with the ecosystem partners to generate data for pre-training and post-training to create different models for different humanoid robots, for different scenarios. For example, in the retail space, there is opportunity for them to be a cashier person, for example, or a sister shopper with a more sophisticated level of capabilities. That's our plan. Thank you.

So we talked about the third thing about the future.

We are also in the discussion of this. But we saw that the future is like a car. 嗯 所以基层的销售价格肯定是比一台一般的车要贵的啊 所以基层是会有比较好的毛利在硬件上 同时我相信虽然我们的基层身上有多颗图灵芯片有小几千体的算力 但是我们也认为基层在云端使用的token量也是要有的 所以我们也期望基础是在软件收费里面更快地跨出那一步 所以换一个角度 我觉得对一家商家或者一个小币来说 他在买一个基础的时候 他会考虑我多少个月我能够把这个钱给挣回来 所以我们相信在海外我们能够甚至几个月 就把这个事情给能实现商家的目标 对中国可能有时间稍微长一些 所以我们一直在考虑客户如何去使用 why it's bought and why it's worth the price. To explain its 3D logic.

Now, when it comes to the commercialization of our humanoid robots, we're still in the discussion phase of it. Even though the bomb structure of a humanoid robot is very similar to that of a car, the RASP, the retail price of a humanoid robot is more expensive than a car, naturally. And so the GP margin, in terms of the hardware of a humanoid robot, but it will be more superior than that of a car. We have incorporated multiple chewing chips in our humanoid robot with various types of computing power. Even though we have that in the hardware, we believe that there's still a lot of commercial value for the software usage in terms of the cloud usage of their humanoid robot computing power. So for the software commercialization step or the software licensing fee revenue, I think humanoid robots have a better potential than cars. Now, on the other hand, if you look at things from our potential customers, for example, business owners' perspective, if they are a small business owner, when they make the human decision of buying a humanoid robot, what they prioritize is the payback period, right? And so I think in the overseas market, the payback period can be much shorter than that of what you would achieve in China. And so it would make more sense economically for overseas business owners to buy humanoid robots. And so we have always been thinking, you know, the rationale of how customers use the humanoid robots, why they would purchase it, and how our humanoid robots can add value to them. Thank you.

Thank you.

Your next question comes from Yu Yan Ding with HSBC.

Please go ahead.

各位管理层好 我是汇丰的雨倩 我有两个问题 第一个是关于中长期的这个战略 还有一个是关于短期的这个财务 那第一个问题 第一个战略方向的问题想请教 看到公司最近有改名成这个小鹏集团 看起来像是走向一个 Physical AI Platform Company的这个趋势 那管理层能不能分享一下 对于长期的商业模式的眼镜的思考 然后我们如何去看待各个战略板块 产生收入的时间点 和就整个中长期在收入中的这个结构的变化 I've got two questions One on the mid to long term strategy and second is financial in the near term So the first question is we noticed the company changed the name to Expon Inc and it's shaping like a physical AI platform company So can management share a bit more discussion and strategic thinking about the business model evolving into mid to long term? And how do we compartment the strategic division's commercialization timeline? And how does into the mid to long term the revenue structure change?

Is it a good question?

Thank you. 那么这个问题我只能非常简要的跟你分享一点点我们的一些简单思考 我觉得在当前实际上最主要的收入肯定来自于规模效应 主要来自硬件 以及我们如果从生态角度是体系的效应 但是我觉得全球很少有汽车厂商在硬件的规模效应上 能够搭建一个软件的planphone,形成下一个的网络效应,有软件的收费,有整个的多边网络效应,那么再往后我觉得整个智能的agent的起来会产生蚁群效应,也就是像蚂蚁的蚁群之间在中心化跟非中心化里面产生新的效应,我觉得这三个效应都会提高壁垒,都有不同的商业逻辑,

Thank you for the question. This is a broad question, and I'm trying to share with you my brief answer. First of all, right now, our main revenue stream definitely comes from the economy, the scale economy, basically. It comes from our hardware and comes from our ecosystem. and I think globally speaking very few car manufacturers can actually build a software platform and form the next network effect based on the economies of scale with their hardware so network effects include both for example the software fee software revenue and the entire multilateral network effect which we actually have the potential to achieve going forward we'll believe that the whole entire you know intelligent agent emergence will produce the so-called and cluster effects. So basically, it's like the interaction between and colonies where you have network effects in the decentralized and centralized clusters. What we're trying to say is that these three effects are going to have their impact on our business model, and with these three effects playing together, interacting together within X and H, we are going to actually enhance our competitive mode. And so with this system, we actually have a very optimistic expectation for our future business strategy and our value creation. Thank you. Let me just add that in the short to medium term, we will prioritize more on the scale of effects of our brand equity and also gross profit. With the scale of globalization and enhanced GP margin, we will be able to have a positive effect on our overall profitability. And also with sufficient profitability, we will be able to support adequate RV investment, which will also bring about new technological competitiveness and competitive modes to the company.

Thank you, Mr. Peng.

My second question is about the service revenue. Can management share the trend about this year, especially in terms of the scope? Is that possible? we can expand our partnership from China to outside of China.

You change this, Charles. So I think that we continue to maintain the guidance that the total revenue generated from the technology and the services and IP licensing revenue in 2026 is comparable to that of 2025. As you may be also aware of, starting from Q2, we will start a delivery of the Turing SOC to our partner Volkswagen at scale. And I think that we continue to believe that the monetization of technology commercialization through such type of collaboration, it is a very attractive business to us. So I think that given all the proprietary technology we have in-house, I think that we are quite open-minded on expanding such commercialization opportunities of our technologies.

Got it. Thank you, Dean. Thank you. As there are no further questions, now I'd like to turn the call back over to the company for closing remarks.

So thank you once again for joining us today. If you have further questions, please feel free to contact Excellence Investor Relations through the contact information provided on our website or the PSN Financial Communications.

Thank you. This concludes today's conference call. You may now disconnect your lines.