XPeng achieves its first quarterly profit, plans to invest another 7 billion yuan in physical AI by 2026

“Scaling up allows us to survive in competition. Absolute leadership in physical AI technology and commercialization will become our core competitiveness,” said Xiaopeng Motors Chairman and CEO He Xiaopeng during the earnings call. He also revealed that by 2026, Xiaopeng’s investment in physical AI will reach 7 billion yuan.

Just one day earlier, Xiaopeng announced its first single-quarter profit since its founding. On March 20, Xiaopeng Motors released its unaudited financial results for Q4 2025 and the full year, with a net profit of 380 million yuan in Q4.

Among new carmakers with billions in revenue fluctuations, 380 million yuan is not a large absolute figure, but it is the first time Xiaopeng has achieved a single-quarter profit since its inception.

Sources close to Xiaopeng told Wall Street Insights that since 2023, Xiaopeng has been very focused on cost reduction, while also opening a second revenue stream by licensing intelligent driving technology to mass-market automakers. These two strategies are key to Xiaopeng’s quarterly profitability.

In China’s new energy vehicle market, which is deeply entrenched in stock competition and normalized price wars, this financial report offers a valuable business perspective. How a company once obsessed with pure electric and intelligent driving technology, suffering market setbacks, has managed to pull itself out of the mire through pragmatic strategic adjustments, strict cost control, and monetization targeting B-end customers.

Achieving profitability for the first time is just the surface; what’s more noteworthy is the change in Xiaopeng’s profit structure and business model. However, the money-burning investments are ongoing, and Xiaopeng is far from celebrating with champagne.

Profits Beyond Car Sales

In Q4 2025, Xiaopeng delivered 116,249 vehicles, a quarterly record; revenue reached 22.25 billion yuan, up 38.2% year-over-year; gross margin was 21.3%, up 6.9 percentage points; net profit was 380 million yuan, marking its first quarterly profit.

For the full year, Xiaopeng’s total revenue was 76.72 billion yuan, up 87.7%; total deliveries reached 429,445 units, up 125.9%; net loss narrowed sharply from 5.79 billion yuan in 2024 to 1.14 billion yuan.

However, in the fiercely competitive mass market priced between 150,000 and 250,000 yuan, maintaining high-level intelligent driving features while increasing gross margin by 6.9 percentage points and turning a profit cannot be achieved solely through scale of vehicle sales.

Xiaopeng Vice Chairman and Co-President Gu Hongdi stated in the earnings report, “We have achieved a profit path different from traditional automakers through a technology-led business model.” This points to Xiaopeng’s “services and other income” in the financial statements.

Data shows that in Q4 2025, Xiaopeng’s service and other income was 3.18 billion yuan, up 121.9% year-over-year and 36.7% quarter-over-quarter; for the full year, it totaled 8.34 billion yuan, up 65.6%.

This income mainly comes from three sources: revenue from providing R&D services to automakers (such as collaborations with Volkswagen), sales of parts and accessories, and carbon credit business income.

This indicates that an important part of Xiaopeng’s profit structure involves a significant technology licensing segment. While other new entrants still rely on car sales profits to cover R&D costs, Xiaopeng has already begun generating additional income through technology licensing.

At the same time, cost reduction in technology and scale effects have also played a role.

Data shows that in 2025, Xiaopeng’s gross margin was 12.8%, up from 8.3% in 2024. Considering the starting prices of Xiaopeng’s main models—MONA M03 at just 119,800 yuan and P7+ at 186,800 yuan—maintaining a 12.8% gross margin at this price level demonstrates the effectiveness of its cost reduction strategies.

However, compared to the overall gross margin of 18.9%, the automotive gross margin is 6.1 percentage points lower. This gap is primarily due to revenue from services and other income, such as technology licensing. In other words, without income from collaborations with Volkswagen and others, Xiaopeng’s overall gross margin would be significantly lower.

Despite achieving quarterly profit, Xiaopeng’s full-year net loss in 2025 still reached 1.14 billion yuan. Looking ahead, management provided a cautious outlook. Xiaopeng’s delivery guidance for Q1 2026 is between 61,000 and 66,000 units, with revenue expected between 12.2 billion and 13.28 billion yuan.

Given the impact of traditional seasonal factors like the Spring Festival, this guidance is quite conservative. It also indicates that quarterly profitability does not mean Xiaopeng has fully turned the corner; maintaining high R&D investment (28.7 billion yuan in Q4) remains a necessary expense to stay competitive in technological arms race.

7 Billion Yuan Gamble

Today’s automotive market features increasingly complex competition.

Huawei’s HarmonyOS Smart Driving is building a strong brand presence in the high-end intelligent driving market through a lightweight, asset-light technology empowerment model; Xiaomi Auto leverages its unique traffic and ecosystem advantages, demonstrating a powerful siphoning effect. Traditional automakers are thriving in the new energy sector, capturing large market shares.

Faced with these tech giants with vast ecosystems, Xiaopeng relying solely on incremental hardware innovations and short-term cost-performance improvements is unlikely to establish a lasting competitive barrier.

This forces Xiaopeng to venture into deeper waters—moving beyond vehicle manufacturing alone, toward foundational shifts in computing power, software ecosystems, and even general physical AI.

From He Xiaopeng’s comments during the earnings call, it’s clear that Xiaopeng is systematically shifting its business narrative toward “physical AI.”

He Xiaopeng revealed that in 2025, the company’s R&D expenditure was 9.5 billion yuan, with 4.5 billion yuan dedicated to AI-related investments. This year, Xiaopeng will continue increasing AI R&D, with physical AI investments expected to rise to 7 billion yuan.

Beyond automotive plans, two strategic moves were highlighted during the call: first, the large-scale deployment and external supply of its self-developed “Turing” AI chips; second, testing commercialization of Robotaxi and high-end software.

Historically, automaker chip R&D was mainly to ensure supply and reduce costs by avoiding high premiums from general-purpose hardware providers. Xiaopeng’s application of the Turing chip in its own models confirms this cost-saving logic.

He Xiaopeng explicitly welcomed other automakers and robotics companies to use the Turing chip. This move essentially aims for Xiaopeng to enter the ecosystem of foundational computing providers.

In this winner-takes-all semiconductor and computing power market, if Xiaopeng can provide foundational computing as a Tier 1 or even lower-tier supplier, its role will no longer be just a participant but a rule-maker.

Alongside this shift, the automotive software revenue model is also being reconstructed. Xiaopeng plans to advance Robotaxi operations in the second half of 2026, which is not a heavy-capital ride-hailing service.

Unlike traditional ride-hailing platforms, Xiaopeng prefers to act as a technology and hardware provider. With large models supporting end-to-end solutions, both consumer vehicles and B2B Robotaxi operations will essentially become massive data collectors and computing terminals.

Industry insiders told Wall Street Insights that this transformation means Xiaopeng’s revenue will gradually move beyond one-time hardware sales. Advanced autonomous driving capabilities will no longer be just a vehicle feature but a service that can be charged via high-end software subscriptions and per-kilometer Robotaxi operations.

This long-tail revenue model with extremely low marginal costs and high gross margins exemplifies a SaaS-like business scenario. From “selling hardware” to “selling software” to “selling services,” collaborative projects with the public are just the beginning of this model’s rollout.

During the earnings call, He Xiaopeng also shared the latest progress on humanoid robots: IRON is expected to begin mass production by the end of 2026, with a monthly capacity of over a thousand units; equipped with three Turing AI chips, the second-generation VLA technology stack has been tested on the robot.

IRON will initially be used in Xiaopeng stores and campuses for guiding and shopping assistance, with future expansion into industrial and home scenarios. He Xiaopeng stated that in the next 5–10 years, the market space for physical AI applications will be even broader than the automotive industry—“global Robotaxi and humanoid robot markets will each be worth trillions.”

Looking back at Q4 2025, the 380 million yuan quarterly profit proves Xiaopeng has moved past its most dangerous phase.

But the other side of this achievement is also clear: a full-year net loss of 1.4 billion yuan still indicates a long way to go for sustained profitability; the decline in Q1 delivery guidance hints at the pain of model transitions; and the 7 billion yuan investment in physical AI will test cash flow management.

He Xiaopeng stated during the call, “Scaling up allows us to survive in competition. Absolute leadership in physical AI technology and commercialization will become our core advantage.”

Industry insiders told Wall Street Insights that Xiaopeng’s current strategy is clear: use short-term profits to buy time for technological advancement, then leverage technology to unlock new business models. But how long this window lasts remains uncertain.

By 2026, Xiaopeng must prove that its quarterly profitability is not a flash in the pan, and that the 7 billion yuan gamble on physical AI can truly pay off.

Risk Warning and Disclaimer

Market risks exist; investments should be cautious. This article does not constitute personal investment advice and does not consider individual users’ specific investment goals, financial situations, or needs. Users should evaluate whether any opinions, views, or conclusions herein are suitable for their particular circumstances. Invest at your own risk.