XPeng Inc. (HKG:9868)
Hong Kong flag Hong Kong · Delayed Price · Currency is HKD
38.28
+1.70 (4.65%)
Nov 17, 2026, 4:08 PM HKT
← View all transcripts

Earnings Call: Q3 2026

Nov 16, 2026

Summary

Q3 revenue grew 19.1% YoY, deliveries rose 15% QoQ, and vehicle margin improved sequentially. Q4 guidance points to record revenue and deliveries, while overseas expansion and Physical AI commercialization anchor the 2027 growth strategy.

Operator

Hello, ladies and gentlemen. Thank you for standing by for the Third Quarter 2026 Earnings Conference Call of XPeng Inc. At this time, all participants are in listen-only mode. After management 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 Xie, Head of Capital Markets of the company. Please go ahead, Alex.

Alex Xie
Head of Capital Markets, XPeng

Thank you. Hello, everyone, and welcome to XPENG's Third Quarter 2026 Earnings Conference Call. Our financial and operating results were issued via newswire services earlier today and available online. You can also view the earnings press release by visiting the IR section of our website at ir.xiaopeng.com. Participants on today's call from our management will include our co-founder, Chairman, and CEO, Mr. He Xiaopeng, Vice Chairman and President, Dr. Brian Gu, 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 conference call will be available on the IR section of our website. Before we continue, please note that today's discussion will contain forward-looking statements made under the safe harbor provisions of the U.S. Private Securities Litigation Reform Act of 1995. Forward-looking statements involve inherent risks and uncertainties.

As such, the company's results may be materially different from the views expressed today. Further information regarding these and other risks and uncertainties 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 forward-looking statements, except as required under applicable law. Please also note that XPENG ' s earnings press release and this conference call include the disclosure of unaudited GAAP financial measures as well as unaudited non-GAAP measures. XPENG'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 Xiaopeng. Please go ahead.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

In the second half of 2026, XPENG achieved growth despite challenges on both demand and cost fronts across the automotive industry, while continuing to improve our operational performance. In the third quarter, our vehicle deliveries reached 118,390 units, up 15% quarter-over-quarter. As sales in international markets and premium models grew rapidly, our vehicle margin improved sequentially to 13%, while our overall gross margin approached 19%, with operating cash flow exceeding CNY 2 billion. In October, our monthly deliveries surpassed 50,000 units for the first time, hitting a record high. Many believe that China's automotive industry is facing a challenging environment today, with both sales volumes and profitability under pressure. I believe, however, that this environment presents an opportunity for industry leaders to prevail.

Only companies with core competitive strengths, a discipline focused on operational excellence, a commitment to global expansion, and a forward-looking strategy can successfully navigate industry cycles. Looking ahead, we will not only pursue high-quality revenue growth but also deliver improved operating results through greater economies of scale and a more favorable business mix. This will lay a solid foundation to capture the long-term growth opportunities in Physical AI. Our technological leadership in Physical AI will also create more high-margin commercial opportunities with high barriers to entry across global markets.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

I am excited that our automotive industry business has entered a new phase of rapid growth, powered by a succession of global blockbuster models. In the third quarter, deliveries of the XPENG GX exceeded 20,000, ranking among the best-selling NEV models priced above CNY 300,000. MONA L03 has become one of the most popular new models in the industry this year, with deliveries exceeding 20,000 in October. Production capacity continues to ramp up. Following its launch in late October, MONA L05 quickly became a breakout hit in the large five-seat SUV segment, with 98% of customers choosing the Ultra SE and Max versions, equipped with advanced ADAS capabilities.

I expect MONA L05 to achieve the fastest production ramp-up in XPENG's history, with deliveries exceeding 10,000 units in the first month. We expect our monthly sales volume to surpass 60,000 for the first time in November. This will be another milestone for XPENG.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

Our international business is XPENG's strongest driver of growth and profitability. In October, as overseas deliveries of MONA L03 commenced, our total overseas sales reached 13,000 units, up 166% year-over-year, setting a new record. Our overseas sales target for the fourth quarter is to exceed 40,000 units, while full-year overseas sales are expected to surpass the 100,000-unit milestone for the first time in 2026. Our competitive advantages in technology and product quality are becoming increasingly evident in international markets, with per-vehicle profit contribution substantially higher than in the domestic market. We are rolling out multiple models, including the G6, G9L, and MONA L05 across global markets, and our order backlog is growing rapidly. In key markets such as Australia and the U.K., we are investing more resources in building locally operated subsidiaries.

This approach has been well received by dealers and customers, driving faster expansion of our sales network and market share. We are also ramping up our production capacity in Europe and other overseas markets. By deepening our partnerships with global automotive industry and supply chain partners, we can better meet customer expectations and regulatory requirements.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

Looking ahead to 2027, I believe XPENG will sustain robust growth and significantly outperform the broader industry. We plan to launch multiple new models and facelift versions to expand our product portfolio and grow our market share in China and overseas. In the domestic market, we will expand our sales network through more diverse, high-quality formats, bringing XPENG series vehicles, MONA series, and our robots to cover more regional markets. We will also offer more targeted products and services to better address diverse customer needs. In international markets, we expect our sales network to exceed 1,000 stores, supporting more than a doubling of overseas sales compared with 2026. Our international business will account for a substantially higher share of total vehicle sales and revenue and make meaningful profit.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

Our goal is that XPENG continue to lead the deployment and commercialization of autonomous driving technologies in China and around the world. In the third quarter, we rolled out a new version of the second generation VLA to users in China, including VLA Lite for Max owners. User experience and satisfaction have improved markedly. During the Paris Motor Show in October, XPENG was the only Chinese automaker to partner with the event organizer to offer autonomous driving testing experiences, competing head-to-head with another international automaker. Despite complex road conditions in central Paris and traffic rules different from those in China, XPENG's NGP successfully navigated the Arc de Triomphe roundabout. We expect to be among the first batch to obtain regulatory approval in the E.U. in 2027, enabling us to take the lead in deploying urban ADAS in Europe.

We have been particularly encouraged by overseas customers' keen interest in our AI-powered ADAS and smart cockpit technologies. Even before relevant local regulations are in place, a sizable portion of customers have already chosen the premium Ultra version. I believe the rollout of VLA and VLM will give XPENG a powerful competitive differentiator globally, driving higher sales volumes and average selling prices. It will also play a pivotal role in building XPENG into a global technology brand. More importantly, our ultimate goal for XPENG NGP is to achieve level 4 autonomous driving capabilities in mass-produced vehicles for consumers. I believe the level 4 experience will create highly sticky demand for paid autonomous driving capabilities in the consumer market.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

XPENG's Robotaxi business has taken a major step toward commercial operation. We have recently been granted approval to conduct intelligent connected vehicle demonstration operations to carry out L4 Robotaxi demonstrations across Guangzhou. We have officially launched XPENG YOYO as our Robotaxi service brand. Starting in the fourth quarter of this year, we plan to deploy a replicable products and service model in Guangzhou and begin offering end-to-end solutions to global partners next year. As our technical and operational capabilities continue to advance, I believe XPENG YOYO will begin scaling the operations in 2028, starting with overseas markets. L4 autonomous driving will make mobility safer and more convenient while generating higher revenue from hardware and software sales, services, and revenue sharing from operations.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

Our next-generation humanoid robot is accelerating toward the hardware mass production by year end. In September, we put our first automated production line for advanced general-purpose humanoid robots into operation. The production line has an automation rate of over 80% across core manufacturing processes and is designed to support mass production at automotive-grade quality standards. We have also finished signing supplier nomination agreement with our suppliers. Leading component manufacturers have provided active support to ramp up production. Our full-stack in-house R&D capabilities, featuring deep integration of hardware and software, are complemented by years of experience in manufacturing and supply chain management. Together, these strengths give me confidence that XPENG's humanoid robot will be industry leading in design, quality, production efficiency, and cost competitiveness.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

Advanced intelligence is XPENG IRON's greatest competitive advantage. We would like to do away with remote. We want to operate without the rules dictated by preset software. IRON can recognize multiple people and remember their identities. When interacting with people, it can observe and think in real time, respond promptly to individual needs, and provide personalized customer engagement and in-depth product explanations, just like our top salespeople. In the first quarter of 2027, the first batch of mass-produced IRON robots will begin trial deployment in some XPENG stores. I expect IRON to become one of our star employees, delivering a delightful experience to customers visiting our stores. In the second quarter of 2027, we will officially launch XPENG IRON for external commercial customers.

As an advanced general-purpose humanoid robot, IRON can quickly learn and adapt to different tasks across a wide range of commercial scenarios, taking on different roles at different times based on needs, 24/7. As IRON deliveries scale up, large volumes of high-precision data collected from real-world scenarios will substantially accelerate the iteration of our robotics AI models and further strengthen our technological leadership. Over the past several months, as I have become more deeply involved in the technical deployment, development, and organizational management of our robotics business, my confidence in successfully deploying and developing advanced humanoid robots has grown steadily. On the business model, I'm optimistic that XPENG IRON has the potential to far surpass the most popular premium vehicle models on the market in terms of both sales potential and gross profit.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

At the upcoming XPENG AI Tech Day this month, I plan to share our technical perspectives on automobiles, robotics, and the Physical AI paradigm, along with some of the results achieved through real-world applications. As large language model companies enter the robotics space, Physical AI companies, in my view, will retain irreplaceable competitive advantages. Leveraging capabilities in hardware, mass production, data, edge computing power, and commercial deployment, these companies can enable the safe, reliable, and large-scale deployment of cars as robots and humanoid robots across countless real-world scenarios, unlocking tremendous productivity gains. At XPENG AI Tech Day and the Guangzhou Auto Show, the latest version of XPENG IRON will be showcased and will feature a range of distinctive capabilities. I look forward to sharing these exciting developments with you.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

As autonomous driving regulations ease globally and the competitive landscape evolves, international automakers and mobility companies face an increasingly urgent need for a technological transformation. In response, we officially launched our technology solutions business, XPENG AI Engineering, in October. We're opening up opportunities for collaboration with external partners in smart electric vehicle technologies while offering Physical AI solutions, spanning Turing AI chips, foundation models, autonomous driving, and robotics. We have established a dedicated business development team that is already in discussions with multiple companies. Our goal is to bring XPENG's technologies into the global automotive, mobility, and robotics ecosystem, fostering greater openness and shared success.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

For the fourth quarter of 2026, I expect deliveries to be approximately 169,000- 174,000 units, representing year-over-year growth of 45.4%-49.7%. Total revenues are expected to be between CNY 30.4 billion and CNY 31.8 billion, up 36.6%-42.9% year-over-year. I am very confident that Q4 will set new records in both deliveries and revenue while continuing to improve the quality of our operations. Moreover, we will enter the first quarter of 2027, a traditionally low season for the industry, with a substantial order backlog. Industry headwinds will further differentiate companies, and we aim to widen our lead over competitors through sustained growth momentum, abundant cash flow, and healthy profit margins.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

Physical AI development is not a sprint but a marathon over the next decade. Looking ahead at 2027, I plan to devote more attention to enhancing our overall operational performance and strengthening our ability to generate cash. At the same time, we will accelerate the mass production and delivery of advanced humanoid robots and substantially strengthen XPENG's local market presence and brand momentum worldwide. Over the coming quarters, you will see the benefits of economies of scale increasingly translate into our better operating results, with technology-driven revenue growth and cost optimization becoming more evident. Our innovations in AI Infra and the deployment of recursive self-improvement, or RSI, in autonomous driving have significantly improved our R&D productivity and operational efficiency. This enables us to achieve far greater progress in R&D than before with comparable levels of human and computing resources.

In the short to medium term, we aim to achieve positive cash flow and profitability for our automotive business, providing a solid financial foundation for our strategic investments in Physical AI. XPENG's emerging Physical AI businesses benefit from strong synergies with our automotive business, enabling highly efficient R&D investment while developing diversified revenue streams across both B2C and B2B markets. Ultimately, our strategic investments will build lasting competitive moats in technology, brand recognition, and commercial ecosystems. Step by step, we will move towards sustainable profitability as we pursue our vision of becoming a global leader in Physical AI.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

Thank you, everyone. With that, I will now turn the call over to our VP of Finance, Mr. James Wu, who will walk you through our financial performance for the third quarter of 2026.

James Wu
VP of Finance and Accounting, XPeng

Thank you, Xiaopeng. Now let me provide a brief overview of our financial results for the third quarter of 2026. I will reference CNY only in my discussion today unless otherwise stated. Our total revenues were CNY 24.28 billion for the third quarter of 2026, an increase of 19.1% year-over-year and an increase of 23% quarter-over-quarter. Revenues from vehicle sales were CNY 21.63 billion for the third quarter of 2026, an increase of 19.8% year-over-year and an increase of 26.9% quarter-over-quarter. The year-over-year and quarter-over-quarter increases were primarily attributable to higher vehicle deliveries and higher average selling price as a result of the improvement in product mix of models. Revenues from services and others were CNY 2.65 billion for the third quarter of 2026, representing an increase of 13.7% year-over-year and a decrease of 1.9% quarter-over-quarter. The year-over-year increase was primarily attributable to increased revenues from carbon credit trading.

Gross margin was 18.9% for the third quarter of 2026, compared with 20.1% for the same period of 2025 and 20.7% for the second quarter of 2026. Vehicle margin was 13.1% for the third quarter of 2026, compared with 13.1% for the same period of 2025 and 12.1% for the second quarter of 2026. The quarter-over-quarter increase was due to the improvement in product mix of models. R&D expenses were CNY 2.85 billion for the third quarter of 2026, representing an increase of 17.5% year-over-year and a decrease of 2.1% quarter-over-quarter. The year-over-year increase was mainly due to higher expenses related to development of new vehicle models and AI-related technologies as the company extended its product portfolio to support future growth. SG&A expenses were CNY 2.78 billion for the third quarter of 2026, representing an increase of 11.4% year-over-year and an increase of 11.2% quarter-over-quarter.

The year-over-year increase was primarily due to higher marketing and advertising expenses and higher personnel costs related to the sales functions. The quarter-over-quarter increase was primarily due to higher marketing and advertising expenses and higher commission to the franchise stores. As a result of the foregoing, loss from operations was CNY 0.75 billion for the third quarter of 2026, compared with CNY 0.75 billion year-over-year and CNY 1.14 billion quarter-over-quarter. Net loss was CNY 0.87 billion for the third quarter of 2026, compared with net loss of CNY 0.38 billion year-over-year and net loss of CNY 1.34 billion quarter-over-quarter. As of September 30, 2026, our cash position was CNY 45.41 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 our third quarter 2026 financial results. This concludes our prepared remarks.

We will now open the call to questions. Operator, please go ahead.

Operator

Thank you. If you wish to ask a question, please press star one on your telephone and wait for your name to be announced. If you wish to cancel your request, please press star two. If you are on a speakerphone, please pick up your handset to ask your question. For the benefit of all participants on today's call, if you wish to ask your question to the 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, and then feel free to follow up with your next question. Your first question comes from the line of Tim Hsiao with Morgan Stanley. Please go ahead.

Tim Hsiao
Analyst, Morgan Stanley

[Non-English content] So my first question is about the 2027 outlook and growth.

How do you see Chinese autos domestic and overseas sales growth in 2027? Which products and pricing segment will drive XPENG growth? What could help the company to buck the cyclical industry downturn and gain market share? That is my first question. Thank you.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

First of all, thank you for the question. I think that in 2027, the domestic automotive industry will still face headwinds, but it is going to be slightly better than 2026. In terms of the overseas market opportunities, I think it will represent a very significant growth for XPENG, definitely a major and rapidly growing opportunity for us because this year, if you look at the total vehicles exported from China, they are mainly composed of a considerable proportion of ICE vehicles and hybrid vehicles. However, I have firm belief that BEVs will actually excel in the future in 2027 and 2028. Looking ahead to 2027, I think that XPENG will significantly outpace the industry's sustained growth because one of the most important factor is, first of all, to sell well and do well in the current models that we have in our current price segment.

This year, we actually place a lot of emphasis on particular models starting from, for example, GX, L05, and L03. We can actually see that three months or six months after they were launched, we already see very good results and market feedback in terms of the sales volume. Next year, in 2027, we are also going to launch new models and facelift versions, thereby broadening our entire product portfolio. Knowledge international-wise, we also will expand and diversify our channels as well, not only internationally but also domestically. For example, previously, we mainly prioritized top-tier cities in China. However, we believe that we can actually deepen our distribution to lower-tier cities, for example, in the northern, northwest, and also western part of the domestic market, which will allow us to basically build our differentiated factor to improve conversion efficiency among our customers.

At the same time, as we just mentioned, the overseas market will be a huge driving force for us. The growth rate will also increase at least by double next year. I am hoping for actually a lot of growth in our domestic and international market next year. I think it is going to represent basically the first year of our rapid growth for XPENG, no matter if you look at the product capability and also our profitability and autonomous driving capability. Thank you.

Tim Hsiao
Analyst, Morgan Stanley

[Non-English content] My second question is about the humanoid robots. As more LLM heavyweights enter Physical AI, where would XPENG's robotics moat lie? Is that going to be models, data, motion control, hardware, or faster deployment at scale? Does the heavy investment in proprietary model still make economic sense for a company like XPENG? That's my second question. Thank you.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

Thank you. This is a very interesting question. Actually, it has been the subject of extensive discussion recently, including our internal discussions within our own company. Regarding this topic, I actually would like to draw everyone's attention back to, for example, back in 2015- 2020, around that time, that is when the R&D of autonomous driving first started. We actually already saw many people achieving autonomous driving through computer simulation. However, you can actually see them failing to be deployed in the real world due to lack of precision and deployment capabilities. Also, there is issues regarding the downstream security as well. We also saw that back then, I remember it was around 2018, when there were a lot of L4 companies made a lot of great demonstrations, but today there isn't really a lot of their roll out on the road as expected.

There was a lot of actual deployment issue as well. Sometimes it has to do with their cost structure. There were also at a time when the market was prevalent with a lot of combination of cloud and edge architecture, but actually today you don't see them deploying on the road. The most critical issue is latency. What I am trying to convey here is, first of all, based on our assessment, it is very difficult to try to use a general model, a System 2 model to achieve the combined capability of this System 0, 1 and 2. It is actually very difficult to deploy it using traditional routes. Different companies now are working on different robots. I think they will actually present different competitive moats. For XPENG, as our company, we need to focus on four capabilities or four directions of development.

The first one is actually our hardware capability. A lot of people, when they talk about Physical AI, they thought that the emphasis is on AI. No, the emphasis should be on physics. The physical capability is very important. The second factor is large-scale commercial production or mass production capability. When we, XPENG, enter the space, we come with our manufacturing capability and our sales capability, which will be very useful in the future of Physical AI day and age. The third area is basically the data capabilities with scale. Right now, I actually see that many companies are spending a lot of money to purchase high-quality diverse data, especially on visual data. For example, when it comes to Physical AI, it is just like when it comes to the functions of a human being.

It is very important that we have, for example, the touch feedback, the pressure, the temperature, haptic data, et cetera, tactile sensation. These are very critical to having Physical AI capability. The fact that we actually have a lot of visual data is very competitive in the market. With hardware and real-world data feedback, it would make Physical AI more possible. Obviously, the true generalization and deployment of Physical AI ultimately depend on the computing power, not only in the cloud, but also at the edge as well. We are going to see that inference at the edge and also cloud-based training, cloud-based simulation, all requiring massive computing power support on different fronts. In this regard, I believe that Chinese EV companies with in-house development capability have actually strong moats compared to international players. That is basically my take on your question. Thank you.

Tim Hsiao
Analyst, Morgan Stanley

[Non-English content]

Operator

Thank you. The next question comes from the line of Tina Hou with Goldman Sachs. Please go ahead.

Tina Hou
Analyst, Goldman Sachs

[Non-English content] So thanks for taking my question.

My first question is regarding the 2027 overseas sales target, which is at least doubling from 2026. Among this, could you give us some color on different geographic distribution and also on the new model pipeline in overseas market, as well as how do you plan to do a combination of pricing, spec, and smart capabilities to raise your market share in overseas market? Thank you.

Brian Gu
Vice Chairman and President, XPeng

Hey, Tina, it's Brian. Let me answer your question here. First of all, yes, we're very excited about the overseas opportunity that Xiaopeng mentioned. Next year, we aim to more than double what we achieved in terms of volume this year. I think it comes from a number of factors. From geographic split, I would say Europe's still going to be probably around 50% or more of our volume. Then I think followed by Southeast Asia and Asia Pacific, probably around 30%. Then the rest will be the Middle East and Latin America, about 20%. For us, we look at these markets, we think the growth will come from some of the recently launched models. For example, we actually launched the G9L, the L05, the G6, all going to be global models next year.

Some of those models, for example, even G9L, L03, when it's launched, I think only have about less than a quarter of delivery this year. You can see that most of the volume growth will be driven by these newly launched, I would say, global products. Also, our product covers, I would say, pretty complete market segments. From the more family mass market focus, we'll have the L03, the L05, and G6 covering the mainstream family-used SUV market. For the high-end family-used SUV, the G9 and G9L, as well as the GX for select Middle Eastern markets, will be also launched to cover the high-end uses. Also, our X9 is still one of the most popular MPVs in Southeast Asia, and increasingly, we're also launching in Europe as well.

I think given these new product launches, I think we have a lot of confidence of significant volume and share gains. Also in terms of channels and market development, we also continue to push some of the high potential markets. For example, I recently came back from Australia and also U.K. I think these markets, we are now doubling down our commitment and investment in these markets, and we think the channel development will ensure strong market share growth in these key overseas markets. Internationally, we are looking to build our channel in terms of sales and service points to be more than 1,000. Also, I think in addition to the markets that I mentioned, the U.K. and Australia, we also continue to develop core markets in Europe, core markets in Southeast Asia, as well as in Middle East.

Another thing I think we're very, I would say, happy to see is that XPENG brand positioning in global markets actually has been really differentiated from a lot of other Chinese EV brands. We are really viewed as tech-centric with a premium positioning. Especially till next year, we would like to introduce our high-end ADAS skills, the NGP, in European markets, as well as other technology features. I think the continued leadership in tech differentiation will actually help us differentiate our brand positioning, also allow us to achieve premium pricing and premium product perception as well. One last point I want to make is that internationally, there's a lot of talks about potential trade barriers and tariffs. I think as a company, we have been dealing with that already for a few years.

We'll continue to be very nimble in terms of our product positioning, our pricing strategy, as well as our channel business models. In addition, we also have been looking at earlier deployment of manufacturing investments, for example, in Europe, in Southeast Asia, and other markets where we think through localization, through build-out of local capabilities and scale-based cost reductions, can actually allow us to be more competitive in these international markets. Overall, we're very excited about the growth, and also we think the volume growth will bring us significant profitability as well.

Tina Hou
Analyst, Goldman Sachs

[Non-English content] So thanks, Brian. My second question is regarding the gross margin as well as unit net profit for the overseas vehicle sales. How do you plan to deal with the next year's export tax rebate decline? Which measures are you planning to take in order to mitigate the impact on that? Thank you.

James Wu
VP of Finance and Accounting, XPeng

Hi, Tina. This is James. I will address your question here. First of all, in the third quarter, our export volume has already exceeded 20% of our total volume. From a revenue perspective, obviously, it's even higher. We would expect the Q4 export portion to be at a similar level and obviously continue to expand over time. From a gross margin of our export business perspective, it is, I would say, more than double versus our domestic business from a gross margin perspective. But if you look into contribution margin on net profit level, the export business is even much more healthier into the several multiples versus domestic business, partially due to its sales structure difference. For example, in certain markets, we operate under a distributorship model where our gross margin effectively is our net profit.

Overall, we believe our export business margin is among the top tier in the industry. I think XPENG brand is very well positioned overseas with price premium. Our goal in 2027 is obviously to continue to grow overseas market with more than double the volume versus 2026 and supporting a better operating performance for the group in the coming years. Now, with regard to the tax policy evolvement, these are just normal cost of business. There are several actions that we are considering to take to counter that impact. One of that is to accelerate our overseas localization effort to not only achieve a better scale but also counter the tax rebate scheme changes. That scheme is obviously related to the CBU exports versus parts and KD exports. Accelerating overseas localization is something that should help us overcome that impact.

The other thing we can do is on the pricing side for pretty much all of the new products that we are pricing at the moment, already considering tax rebate changes into 2027. We are also planning on product cadence adjustments to cope with potential further tax rebate phasing out in 2028. These are the actions we are considering. Lastly, I would say from a scale perspective, next year we expect to achieve a higher scale with new products selling not only for a couple of months but for the full year, including the new products that were launched in the second half of this year, including L03, L05, G9L and GX, which will subsequently supporting us to achieve scale as well as better profitability for our overseas market. Thanks.

Tina Hou
Analyst, Goldman Sachs

Thank you, James.

Operator

Thank you. Your next question comes from the line of Ming Hsun Lee with Bank of America. Please go ahead.

Ming Hsun Lee
Analyst, Bank of America

[Non-English content] So looking for 2027, could you discuss your expectation for the gross margin trend and also how will you reduce your loss and also the major cadence driver? Thank you.

James Wu
VP of Finance and Accounting, XPeng

Yeah. Thanks. This is James. I will address your question. First of all, I think it is still a little too early to provide 2027 earnings guidance. I think we will provide more information in our next earnings call. If you look at our gross margin in the third quarter, there are a couple of things that are changing. One is our vehicle margin has improved quarter-over-quarter driven by our better product mix, primarily due to the launch of GX, which accounts for almost 20% of our volume in the third quarter. Again, export volume increased from below 20% to more than 20%. From a trend perspective, in Q4, given L05 delivery will start to commence and account for a bigger portion of our volume. From an ASP perspective, we do expect a slightly lower ASP in the fourth quarter.

Gross margin is slightly sliding quarter-over-quarter versus Q2. That is partially because in Q2, if you remember, we have our technical R&D service revenue from our collaboration with Volkswagen. There was a major milestone achieved in the second quarter, and lacking that, driving a slightly lower other revenue. That is partially offset by higher carbon credit sales revenue that we start to recognize in Q3, and expect a similar level of carbon credit sales in Q4, because of our projected volume in Europe. Overall, we still expect our group margin in 2026 to be in the high teens. Okay? I would say in 2027, from an outlook perspective, the overall scale is expected to grow. Again, overseas volume will be more than double, better mix from overseas, and better scale.

From a challenge perspective, we have seen obviously continued challenge from a cost side due to AI-related memory chips and et cetera, export VAT we found, as mentioned earlier. There is also battery consumption tax that is being implemented, and we are in the process of discussing with our suppliers. On the expense side, we would expect relatively moderate growth in R&D expense versus 2026. Again, better efficiencies in the sales expense. From a ratio perspective, as a percentage of revenue, both R&D and SG&A, we believe will show a better trend in 2027. Lastly, we expect commercial revenues to continue to commence and continue from our Physical AI business. It is going to be at scale and grow as our expectation.

The profitability of our robotics business, as Xiaopeng mentioned earlier, is expected to be meaningfully better versus, let us say, a premium EV in the market. That is all we can share at this point. Thanks.

Ming Hsun Lee
Analyst, Bank of America

Yeah. Thank you. That is all my questions.

Operator

Thank you. Your next question comes from the line of Xiaoyi Lei with Jefferies. Please go ahead.

Xiaoyi Lei
Analyst, Jefferies

[Non-English content] Thanks for taking my question, and congrats on the strong quarter. I just have one on humanoid robots. Your next generation robot is heading into mass production by the year end. How do you see demand over the next one to two years? Which kinds of companies and use cases do we expect to be the first buyers? Thank you.

He Xiaopeng
Co-Founder, Chairman, and CEO, XPeng

[Non-English content]

Speaker 4

Thank you for the question. I think this is something that everyone in the space would like to go and explore. The issue is that today in China, or even around the world, robotics as an industry is mainly towards the B2B side, even for larger businesses, where they offer a lot of highly customized products. However, for us, we actually don't want our robots to require customization. Instead, we envision it as a highly standardized product for small to medium B2B customers, and eventually for even individual household customers that doesn't need to be customized at all. When it comes to the entire business possibility or the business logic in the future, first—

—I think that if I were in Europe, I would say the best business sector to deploy robots right now is in industries. But I think the best scenarios for China is really in business and family. For industrial robots, the requirement regarding safety may be slightly less stringent. However, the requirements for the hand dexterity is slightly more demanding in terms of the capabilities and skills that you can have through this robot. The criteria for cost is also more stringent as well. However, in the context of commercial use and also for family household use, the emphasis placed more on the safety and also the overall aesthetic appeal of the design of the robot.

So the requirements are pretty much different depending on how and where you want to deploy your robot. It is going to help us to build a great differentiation when we target different market segments. Thank you.

Xiaoyi Lei
Analyst, Jefferies

Thank you, Mr. Chair.

[Non-English content]

Speaker 4

We believe that XPENG's robot may be different from many other products in the market. First, we aim for a highly human-like and versatile design. Since our production line is deeply integrated into our in-house R&D and manufacturing production lines, I am very confident about our manufacturing capability and our supply chain capability, whereas other competitors in the market may face a lot of headwind in this regard. However, from my perspective, whoever in the market is the first to develop high-quality robots, also with high safety features, will be the one that actually wins in the market first. Because when it comes to the robot application and its generalized capability for use in commercial sense and in household sense, it is very important to have high quality together with high safety and reliability.

That is why right now, currently, when you look at the robots that have been sold, a lot of them are left sitting in the warehouse and doing nothing because they are unable to meet high safety and quality standards. That is also a challenge that we have to face, and we need to work to overcome it. Nevertheless, I am very confident that by the end of this year, specifically starting mass production on January 1 next year, with deliveries beginning in the second quarter of next year, we will actually achieve high quality and greater safety than any other robots available in the industry. I believe that once we have overcome that challenge, we will be able to actually grow at a very rapid pace, far surpassing the current production capacity and sales scale of the not-so-high-end robot models. Thank you.

Xiaoyi Lei
Analyst, Jefferies

[Non-English content]

Operator

Thank you. Your next question comes from the line of Jingyue Wu with CITIC Securities. Please go ahead.

Jingyue Wu
Analyst, CITIC Securities

[Non-English content] Okay, this is Jingyue Wu from CITIC Securities International. My first question is, what is the progress of our overseas localization in these strategic regions, and what percentage of vehicles will be locally produced next year?

James Wu
VP of Finance and Accounting, XPeng

Hi, this is James. I think this question was touched already a little bit earlier, but I will briefly address again. If you look at our overseas, from a localization perspective, we have localization footprints in Europe, as well as in Southeast Asia, particularly. There are several manufacturing footprints, including contract manufacturing efforts that we have already established. Different areas would require different localization requirements, and the calculations are different. We have already started exploring another step of deeper localization, including local source parts with suppliers in all of these markets. I think these discussions are ongoing. Eventually, we will meet whatever the localization requirement from the local market in order to achieve the savings from a duties perspective as well as from the local subsidy benefit perspective. I think overall, you can expect our overseas localization to be deeper going forward.

In the meantime, that would also help us to counter, as earlier mentioned, the CBU-related tax, VAT tax refund rebate that is phasing out in 2027 and 2028. With a deeper localization footprint overseas, that will help us to not only become more competitive from a cost perspective overseas, but also saving duty shortcomings from the tax benefits perspective.

Brian Gu
Vice Chairman and President, XPeng

Yeah, let me just add. I think you know that we already have local manufacturing in, for example, Europe in a partnership with Magna in Austria. In Southeast Asia, we actually have also manufacturing in Malaysia and Indonesia. We are looking at definitely expanding and deepening localization, also manufacturing deployment next year. The exact, obviously, markets are yet to be announced, so just stay tuned. I would say that we are very focused on making sure that we can develop a scaled and cost-competitive overseas supply chain that give us the ability to not just produce, but also meet the requirement for local sort of regulatory as well as to satisfy our local customer needs. I think all of that, I think are the sort of the focus for us in terms of development and investments.

Jingyue Wu
Analyst, CITIC Securities

Thank you, Brian. Thank you, James.

[Non-English content] My second question is that in October, the company announced plans to open up its smart vehicle technology to global partners, offering full ecosystem collaboration that spans from vehicle technologies to AI chips and foundation models.

Could you share any updates on the eco-openness and the rollout of the solutions also regarding our partnership with Volkswagen? How should we understand the announcement of expansion to overseas market? Thank you.

Brian Gu
Vice Chairman and President, XPeng

Yeah, this is Brian again. First of all, I think the commercialization of our technology is a key priority for us. Formation of this new XPENG AI Engineering services practice, I think is really taken in the experience we received from the Volkswagen collaboration. I think throughout the collaboration, we have developed the ability to work with third party, service third party, and also generating revenue from our technology stack. I think just a few points I want to make. First of all is that we're actually getting a lot of inbound requests to really work with us on various level of tech stacks. Obviously, in addition to the EV and powertrain-related, vehicle platform-related, we also have seen significant interest in our AI, Physical AI core that we have developed, leveraging our chips, leveraging our model training, AI Infra, et cetera.

I think there's a full stack of interest that we have received. In order to address that, I think we have formed this new business unit, which actually I personally am leading that effort. I think the goal is that we want to really focus on significant partnership opportunities in the platform areas, as well as developing new categories that can leverage our AI and the Physical AI capabilities that we developed to more, I would say, more industry and more segments. That's our goal. I would say this effort is just, I think we just started making that effort as a new business. We obviously have currently already been working with Volkswagen on existing projects. There's other additional collaborations that we're discussing with Volkswagen and also other partners.

I think as we deepen the development of that opportunity, I think we see significant opportunity generate starts next year. That's something I think will leverage both the, I would say, OEM partnership relationship model, which I would say demonstrated by the Volkswagen partnerships. But also, I would say building upon our new Physical AI technologies into products that actually can be partnered not just with OEM, but also other segments of Physical AI, embodied intelligence, robotics, Robotaxis, robo-vans, and other areas as well. Needless to say, we're very open to these discussions, and we're currently focusing on multiple high-potential opportunities.

Jingyue Wu
Analyst, CITIC Securities

Thank you, Brian. I have no further questions.

Operator

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

Alex Xie
Head of Capital Markets, XPeng

Thank you once again for joining us today. If you have further questions, please feel free to contact XPENG's investor relations through the contact information provided on our website or the Piacente Financial Communications.

Operator

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