--- title: "XPeng's Bold Bet on \"Physical AI\": Will It Secure a Ticket to the Future?" type: "News" locale: "en" url: "https://longbridge.com/en/news/288093644.md" description: "Shedding the label of a traditional automaker" datetime: "2026-05-29T15:52:34.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/288093644.md) - [en](https://longbridge.com/en/news/288093644.md) - [zh-HK](https://longbridge.com/zh-HK/news/288093644.md) --- # XPeng's Bold Bet on "Physical AI": Will It Secure a Ticket to the Future? Recently, XPENG-W officially released its financial data for the first quarter of 2026. In the first three months of 2026, XPENG-W recorded revenue of RMB 13.03 billion, a year-on-year decrease of 17.6%. Although the overall gross margin increased by 5.0 percentage points to 20.6% during the same period, the company still reported a net loss of RMB 1.78 billion, a significant expansion compared to the loss of RMB 660 million in the same period of 2025. The direct drivers behind the widening loss include a decline in delivery volume compared to the same period last year, but rising R&D expenses were also a major factor—R&D investment alone reached RMB 2.91 billion in the quarter, showing significant growth both year-on-year and quarter-on-quarter. During XPeng’s Q1 earnings conference call on May 28 and in a podcast released the same day, He Xiaopeng, Chairman of XPeng, provided a directional explanation for the increase in R&D spending in the first three months of 2026: The company is fully transitioning from a "patchwork" architecture of multiple stitched-together AI models to a new system driven by a single general-purpose large model, termed "Physical AI." Around this pivot, he made a series of time-bound commitments, such as a substantial quarter-on-quarter increase in delivery guidance for the second quarter; mass production of Robotaxis has been completed, with demonstration operations starting in Guangzhou in the second half of the year; mass production of the humanoid robot IRON in the fourth quarter of this year, entering XPeng stores as sales guides in the first quarter of 2027, and exporting overseas in the second quarter. With upcoming new models like the L03 and the facelifted G9, XPeng is using a financially pressured short-term report to exchange for entry into the commercialization cycle of Physical AI. Whether this represents strategic foresight or overly aggressive early-stage investment depends on the extent to which these promises are fulfilled over the next four quarters. ## 01 Behind the Widening Loss From the financial report, the core contradiction in XPeng’s Q1 results lies in why the overall loss expanded despite improved per-vehicle gross profit. The answer is reflected in the expense structure. In terms of revenue structure, total revenue in the first quarter was RMB 13.03 billion, of which automotive sales revenue was approximately RMB 11 billion, a 23.5% decrease from the same period in 2025. This decline was mainly due to the drop in delivery volume in the first quarter. Specific delivery data shows that XPeng delivered 62,682 vehicles in the first quarter of 2026, a 33.3% decrease from the 94,008 vehicles delivered in the same period of 2025. However, the overall gross margin remained at 20.6%, an increase of 5 percentage points compared to the same period in 2025. Despite the year-on-year decline in deliveries, the gross margin of the automotive business rose from 10.5% in the same period last year to 12.1%. This indicates an improvement in XPeng’s sales mix, with a higher proportion of high-priced models, leading to better profitability on a per-vehicle basis. Cost reduction through technology and platform sharing are taking effect, resulting in higher profit margins per vehicle sold compared to a year ago. Looking at the expense side, selling and administrative expenses were RMB 1.88 billion, a 3.2% year-on-year decrease and a 32.5% quarter-on-quarter decrease. Against the backdrop of plummeting deliveries and the need for marketing investment for new car launches, the contraction of these expenses suggests that management is still controlling costs. Another point eroding profits is the rise in R&D expenses. R&D expenses in the first quarter amounted to RMB 2.91 billion, an increase of over 46.8% from RMB 1.98 billion in the same period of 2025, and also an increase from the fourth quarter of 2025. This incremental funding was primarily directed towards three areas: intelligent driving R&D, the development of the humanoid robot IRON, and software-hardware adaptation for the Robotaxi fleet. He Xiaopeng also stated in the interview on May 28 that the company is in a critical investment phase of "migrating from an AI patchwork to a unified Physical AI architecture," making these structural costs inevitable. So, where lies the short-term financial turning point? This depends on the pace of new car releases. XPeng’s earnings call provided guidance for Q2 deliveries pointing to a significant quarter-on-quarter increase, specifically around 100,000 to 106,000 units. Supporting this expectation is the intensive launch of several new models. The GX was launched in May. According to research by the Zhongtai Securities auto team, its orders far exceeded expectations, with nearly 25,000 firm orders within 12 hours, and monthly sales are expected to stabilize at around 7,000 units, whereas the market previously expected only 2,000 to 3,000 units. The more crucial volume-driver model, the L03, positioned as a pure electric SUV under RMB 150,000, is scheduled for launch in the third quarter. This model shares a significant amount of underlying architecture with the M03, which has long topped the sales charts for pure electric sedans in the RMB 100,000–150,000 range. Market expectations suggest its monthly sales could exceed 15,000 units. Additionally, the facelifted G9 and the L05 are expected to be launched and delivered sequentially in the second half of the year, with a total of seven "dual-energy" models planned for the full year. If the sales expectations for the L03 and GX are met as anticipated, XPeng will enter a steep ramp-up period for deliveries in the third quarter. The dilution of manufacturing costs driven by sales scale, combined with the positive trend in gross margins for existing models, is expected to push the automotive business gross margin higher within the year. Service and other income, mainly comprising technology licensing and intelligent driving subscriptions, reached RMB 2.03 billion in the first quarter. As the VLA distilled version is gradually installed in vehicles this year and technical solutions are exported to more automakers, this high-margin revenue stream will become an important buffer for adjusting overall profitability. The new car cycle starting in the second quarter will be the first window to test whether XPeng’s investments can be converted into scale and profit. ## 02 The Trilogy of "Physical AI" Beyond the financial report, He Xiaopeng is betting XPeng’s future on three things: intelligent driving technology export, Robotaxi operations, and mass production of humanoid robots. At the May 28 earnings conference call, he clarified three core business lines: advanced intelligent driving, Robotaxi, and the humanoid robot IRON. All point to a qualitative shift from "Digital AI" to "Physical AI" capable of interacting with the real world. Intelligent driving remains the foundation. He Xiaopeng believes that its VLA 2.0 capabilities have reached the absolute forefront domestically, and the company has begun gradually installing the distilled model in vehicles this year, planning to achieve global deployment of VLA 2.0 by 2027. However, Huawei’s Qiankun ADS is expanding rapidly on Aito and Luxeed models, and Tesla’s FSD is accelerating local testing in China. Whether XPeng can maintain its intelligent driving advantage depends on the speed of technological iteration and cost. Using a unified large model to understand, plan, and control vehicles marks a complete departure from the previous rule-driven "patchwork" mode. Prior to this, the fact that XPeng’s VLA 2.0 accounted for over 50% of intelligent driving mileage in its first month of rollout serves as key proof of the correctness of this technical route. This indicates that user dependence on and trust in the new intelligent driving system are crossing the chasm, laying the commercial foundation for XPeng to provide technology licensing to other automakers and expand service revenue. The "countdown to mass production" for Robotaxi and the humanoid robot IRON pushes XPeng’s valuation ceiling from that of an automaker to that of a broader Physical AI service provider. Regarding Robotaxi, based on its closed-loop intelligent driving technology, vehicles have already rolled off the production line, with plans to start demonstration operations in Guangzhou in the second half of the year. Furthermore, He Xiaopeng revealed that future Robotaxis will feature lower-priced models and expand overseas. The advanced humanoid robot IRON represents an even bolder step. Its commercialization path is exceptionally clear and pragmatic: mass production in the fourth quarter of this year, entering its own stores as sales guides in the first quarter of 2027. This "use it first" strategy allows for data accumulation and technological iteration in real-world scenarios while exploring commercial value at a relatively controllable cost. Meanwhile, the humanoid robot IRON also has overseas plans, with exports expected in the second quarter of 2027. XPeng is telling a story of "multiple uses for one core asset," where core AI capabilities serve as the brain, simultaneously driving intelligent vehicles, Robotaxi fleets, and humanoid robots. If the market accepts this narrative, XPeng’s valuation anchor could shift from the price-to-sales ratio typical of automakers to the logic of a technology company. However, currently, revenue from intelligent driving technology licensing is volatile, and both Robotaxi and robotics are still in the investment phase, with no single line independently contributing to profit. What the market needs to see is tangible large-scale implementation and the realization of financial data, not just a perfect technical architecture and business plan. Can IRON truly replace human labor in stores? What will the unit economics look like after the scaled operation of Robotaxis? Until these questions are answered, "Physical AI" remains merely a narrative. For now, XPeng has taken the lead in securing a ticket to the next race track, but the competition has only just begun. ### Related Stocks - [XPEV.US](https://longbridge.com/en/quote/XPEV.US.md) - [09868.HK](https://longbridge.com/en/quote/09868.HK.md) - [516110.CN](https://longbridge.com/en/quote/516110.CN.md) - [516380.CN](https://longbridge.com/en/quote/516380.CN.md) - [515030.CN](https://longbridge.com/en/quote/515030.CN.md) - [159565.CN](https://longbridge.com/en/quote/159565.CN.md) - [159323.CN](https://longbridge.com/en/quote/159323.CN.md) - [TSLA.US](https://longbridge.com/en/quote/TSLA.US.md) ## Related News & Research - [XPENG Reports First Quarter 2026 Unaudited Financial Results | XPEV Stock News](https://longbridge.com/en/news/287887137.md) - [BYD expands ‘God’s Eye’ rollout amid EV price war](https://longbridge.com/en/news/287928967.md) - [Don't use AI when emailing Paul Graham](https://longbridge.com/en/news/287659485.md) - [How to fix AI's branding problem, according to top marketers](https://longbridge.com/en/news/287767399.md) - [COLUMN- Lawyers, meet your AI 'twin'](https://longbridge.com/en/news/288047755.md)