On February 24th, Wall Street News learned that XPeng Chairman and CEO He Xiaopeng sent an internal work letter, themed &#34;Steady Progress and Breaking the Situation, Together Entering the New Decade of Physical AI in 2026,&#34; which set the strategic direction and goals for XPeng in 2026.
&#34;Steady progress&#34; and &#34;breaking the situation,&#34; two seemingly contradictory terms, precisely outline XPeng's current historical position: it must maintain the baseline of scale and quality in the mainstream automotive market while launching a comprehensive charge into the uncharted territory of physical AI, transitioning from capability to mass production.
In 2025, XPeng achieved a delivery volume of 429,400 units, with a year-on-year growth of 126% and a target completion rate of 113%, and was the first among new pure electric forces to achieve one million vehicles offline. However, as he stated in the letter, &#34;We have created several blockbuster products in the past, but that is not enough.&#34; In this fiercely competitive industry, XPeng needs more leverage to reach its ultimate goal.
<h2>01 The &#34;DeepSeek Moment&#34; of Autonomous Driving</h2>
In the letter, He Xiaopeng elaborated extensively on his views regarding autonomous driving. In his opinion, the turning point for autonomous driving has already arrived. For XPeng, it stands at a historic turning point and must seize the DeepSeek moment of autonomous driving in China.
The so-called DeepSeek moment refers not only to the emergence of technological capabilities but also to a dramatic decrease in costs and the popularization of experiences.
He Xiaopeng revealed that he spent the entire Spring Festival conducting internal testing of the &#34;second-generation VLA (Vision-Language-Action) model,&#34; describing its effects as &#34;stunning.&#34; This model, which he regards as &#34;the first version delivered for the L4 fully autonomous driving era,&#34; will be officially pushed out this quarter.
To this end, XPeng also made organizational adjustments. Wall Street News confirmed from XPeng insiders that in early February, XPeng merged its Autonomous Driving Center and Intelligent Cockpit Center into a General Intelligence Center. This means that driving decisions and human-machine interactions will no longer be governed by two separate logical systems but will share the same AI foundational model.
The scenario depicted by He Xiaopeng in the letter is highly vivid: in the future, users will only need to say, &#34;Take the road ahead, avoiding the red lights,&#34; and the vehicle will execute the command automatically. This integrated cabin and driving experience is evolving from a geek novelty to an essential feature for families.
The most significant piece of information in this letter is that Volkswagen will become the first customer of XPeng's second-generation VLA model.
Since Volkswagen's investment in 2023, the cooperation between the two has deepened from platform procurement to electronic and electrical architecture, chips, and even intelligent driving systems. Having a global sales giant pay for China's AI technology is not only a signing of a commercial contract but also a milestone in XPeng's evolution from a technology importer to a technology exporter.
For Volkswagen, the joint venture in China urgently needs leading intelligent solutions to regain market share; for XPeng, Volkswagen's large-scale procurement can not only dilute enormous R&amp;D costs but also signifies that its autonomous driving technology possesses global competitiveness that can be output and reused.
<h2>02 The Watershed of Mass Production in Physical AI</h2>
If the automotive business represents today, then physical AI is the tomorrow in XPeng's eyes
He Xiaopeng clearly stated in the letter: 2026 is a watershed year for XPeng's embodied intelligence to transition from &#34;capable&#34; to &#34;mass production.&#34; XPeng aims to become the world's first technology company to achieve full-scale production of three cutting-edge AI businesses: robots, flying cars, and Robotaxi in the same year.
In terms of Robotaxi, XPeng plans to launch a pilot operation for ride-hailing models this year, with the core goal of running through the initial cycle of technology, customers, and business.
It is worth noting that among the four new SUVs that XPeng plans to launch, except for the small-sized SUV of the MONA series, the other three are all planned to be equipped with the Robotaxi autonomous driving system.
This design concept of a private car and an unmanned vehicle fleet sharing the same architecture is also conducive to diluting hardware costs through scaling, laying the groundwork for Robotaxi to enter the &#34;initial rapid growth phase&#34; in 2027-2028.
Currently, the split-type flying car &#34;land aircraft carrier&#34; has entered trial production and will achieve large-scale production and delivery this year. From taking to the skies to being delivered, this is a key leap for the low-altitude economy from concept to reality.
At the same time, He Xiaopeng also revealed that the highly anticipated new generation IRON robot will start mass production by the end of the year, aiming to become the world's first large-scale production advanced humanoid robot. It will first be deployed in scenarios such as guiding and shopping assistance, and an SDK will be opened to global developers.
In November last year, the Iron robot faced skepticism due to its human-like walking, and He Xiaopeng's video of &#34;self-proof&#34; went viral online, indirectly confirming that its technological accumulation has gained market recognition.
These three business lines may seem dispersed, but their core is highly unified—they are different carriers of physical AI, sharing the same underlying technological foundation. He Xiaopeng's ambition is to transfer the perception, decision-making, and planning capabilities accumulated in the automotive field to bipedal robots and low-altitude aircraft, building a multi-terminal embodied intelligence ecosystem.
Indeed, there is a huge gap between being capable and being able to mass-produce, involving engineering, supply chain, cost control, and regulatory access. In 2026, the market will test whether XPeng has the ability to cross these three chasms simultaneously.
<h2>03 A Year of Product and Globalization</h2>
Returning to specific automotive products, He Xiaopeng candidly stated, &#34;In the past, we created several blockbuster products, but that is not enough.&#34; This marks a profound shift in XPeng's product strategy: from relying on a single blockbuster to competing for every core niche market with a &#34;product matrix.&#34;
2026 is defined as an unprecedented year for products, with an astonishing density of attacks. In the first quarter, three super extended-range products will be launched first, fully implementing a dual-energy strategy. This means XPeng officially bids farewell to the history of relying solely on pure electric vehicles. The first XPeng X9 equipped with super extended-range technology delivered 5,424 units in December 2025 and 4,219 units in January 2026, initially validating the feasibility of the extended-range route in the high-end MPV market.
According to public information, XPeng plans to release more than ten new products within the year, including the 2026 P7+, G7 super extended-range, G6, G9 released in January, and four new SUVs. Behind this product density is XPeng's desire for scale effects. Only with a sufficiently large sales base can the high costs of intelligent research and development be diluted, and a global channel and service network be supported
This year, XPeng not only has many products but also aims to sell these cars globally.
In 2025, XPeng's overseas delivery volume exceeded 45,000 units, a year-on-year increase of 96%, ranking first in the sales of pure electric vehicles exported by Chinese new car manufacturers. However, He Xiaopeng is clearly not satisfied with this; in his letter, he proposed a more ambitious goal: to double overseas sales by 2026 and reach annual overseas sales of 1 million units by 2030, contributing over 70% of profits.
The strategy supporting this goal is summarized by He Xiaopeng in eight words: &#34;sharp knife&#34; to break the deadlock, &#34;red carpet&#34; to retain talent.
&#34;Sharp knife&#34; refers to precise selection of products and markets. On the product side, XPeng will launch 6 global models, with price ranges covering 100,000 to 200,000; on the market side, it will focus on five benchmark markets: Israel, Germany, Norway, Thailand, and France, using these as pivot points to radiate across the Eurasian continent. In terms of channel construction, the number of overseas outlets is planned to double to 680 this year.
&#34;Red carpet&#34; refers to delivery and service. He Xiaopeng is particularly wary of the globalization trap—too many car companies have failed in their globalization efforts due to poorly built teams and inadequate capabilities, ultimately leading to a collapse in quality and reputation. Therefore, he repeatedly emphasizes the need to solidify the scale of global supply and production capabilities, from supply chain, manufacturing, logistics to spare parts service, moving towards global collaboration.
It is noteworthy that XPeng's overseas expansion is no longer just about selling cars. With the P7+ completing trial production in Austria and set to be delivered in 25 European countries, XPeng is replicating China's intelligent experience overseas. Whether it can establish a differentiated advantage in intelligent driving in the European market, the stronghold of BBA, will be a litmus test for XPeng's technological export quality.
<h2>04 Challenges Continue</h2>
The core of implementing the strategy lies in people. He Xiaopeng disclosed a set of data in his letter: by 2026, XPeng's global employee count will increase by another 8,000, with plans to recruit 5,000 campus talents.
In terms of talent philosophy, He Xiaopeng proposed two points: first, to train more people who are &#34;willing to fight and can shoulder responsibilities&#34; on the global battlefield, expecting more students to go overseas to compete; second, to make AI Agents the &#34;capability extension&#34; for everyone, using AI to enhance efficiency from code writing to training generation.
This reflects XPeng's deepening thinking about organizational structure. As the company scales, simply parachuting in senior executives cannot solve cultural integration issues; therefore, He Xiaopeng emphasizes that &#34;future XPeng leaders will more often come from internal growth and promotion.&#34; This strategy of cultivating &#34;homegrown talent&#34; will be key to maintaining organizational stability and ensuring consistent execution globally.
While outlining a grand blueprint, XPeng also faces real financial challenges. Despite achieving record sales in 2025, the price war in the automotive industry continues, and XPeng's R&amp;D investment in cutting-edge fields such as AI and robotics is at a historical high.
Additionally, sources indicate that XPeng has set a sales target of 550,000 to 600,000 units for 2026 in an internal strategic meeting, with brokerages also providing guidance for annual sales of 600,000 units. This target means achieving approximately 40% growth from the 429,400 units sold in 2025. In the increasingly competitive Chinese market, this is no easy task
From &#34;steady progress towards a long-term goal&#34; in 2023 to 2025, and &#34;steady advancement to break the situation&#34; in 2026, XPeng is undergoing a transition from the survival phase to the development phase.
For He Xiaopeng, the track in 2026 is filled with both infinite scenery and thorns. Whether it can seize the opportunity during the &#34;DeepSeek moment&#34; of autonomous driving, and whether it can turn the three major cutting-edge AI businesses from blueprints into products, will determine whether XPeng can become a key player in the &#34;new decade of physical AI.&#34;

XPEV

09868

515030

DRIV

516110

CARZ

159565

MOTO

516380

The full matrix is open for battle

- On February 24, Xpeng Motors' CEO He Xiaopeng outlined the company's strategy for 2026, emphasizing a balance between maintaining market share and advancing in physical AI.  
- The company aims to be the first to mass-produce robots, flying cars, and Robotaxis this year, with significant partnerships and organizational changes to support this agenda.  
- Facing a competitive landscape, Xpeng targets 550,000 to 600,000 vehicle sales in 2026 and aims to double overseas sales, positioning itself as a leader in the evolving AI-driven automotive sector.

XPeng's New Year Sword, pointing towards Physical AI