XPeng Q1 2026 Earnings: GX Orders Run 30-Week Lead Times as Physical AI Pivot Takes Shape
XPeng Q1 2026 earnings call, May 28, 2026 — delivery guidance jumps 60%-plus into Q2 as flagship SUV demand surprises to the upside and robotaxi, humanoid, and overseas ambitions crystallize
GX Demand Is Running Ahead of Supply
The single most important datapoint to emerge from XPeng's Q1 2026 earnings call was the demand signal on the newly launched GX flagship SUV, which went on sale May 20. CEO He Xiaopeng told analysts bluntly that "the performance of GX is above our expectation," and the order data bear that out. The ultra flagship trim, priced above RMB 350,000, accounts for more than 80% of initial firm orders, a mix skew that is still growing. Lead times on that top variant have already stretched past 30 weeks, an unusually long backlog for a Chinese EV that has been on sale for barely a week. CFO James Wu confirmed that the GX carries the highest gross profit of anything in the portfolio, and that most SKUs are delivering better-than-expected unit margins, with only one configuration coming in below internal targets. Management guided Q2 total gross margin to hold roughly flat with Q1's 20.6%, which implies the mix benefit from GX is broadly offsetting ongoing headwinds from memory chip and battery raw-material cost inflation that began weighing on vehicle margins in Q1.
One nuance worth flagging: the MAX version of the GX is currently below 5% of the order mix, somewhat lower than XPeng had anticipated, and the extended-range variant started slowly before picking up momentum through targeted marketing in China's western and northern regions. Neither dynamic undermines the overall demand story, but they do suggest the ultra-premium BEV trim is carrying the commercial narrative almost entirely on its own at this stage. Management was explicit that commercial value, not volume maximization, is now the governing principle for pricing and configuration strategy across the entire lineup.
Q2 Inflection Is the Most Consequential Near-Term Event
XPeng guided Q2 2026 deliveries of 100,000 to 106,000 units, representing sequential growth of 59% to 69% over a Q1 that came in at 62,680 vehicles. Revenue guidance of RMB 19.6 billion to RMB 20.8 billion implies roughly 50% to 60% quarter-on-quarter growth. The Q1 numbers themselves were ugly in isolation — total revenues fell 17.6% year-on-year and 41.4% sequentially, vehicle revenue dropped 23.5% year-on-year, and the company swung from a Q4 net profit of RMB 380 million to a Q1 net loss of RMB 1.78 billion — but management positioned the quarter as a deliberate seasonal trough rather than a structural deterioration. He Xiaopeng was unambiguous: "I'm very confident that deliveries will grow substantially quarter-over-quarter in each of the remaining quarters this year."
The delivery ramp is being driven by four new SUV models launching within the next six months, starting with the GX and followed by three additional nameplates in the second half, all equipped with the Turing SoC and what XPeng calls VLA 2.0 intelligent driving. The company is also benefiting from the completion of its Turing SoC upgrade across the entire existing model lineup, a transition that appears to be resonating with buyers: more than 85% of Mona M3 customers are now choosing the MAX or Ultra SE variants, and the A-class sedan has held the top spot in its segment for 19 consecutive months.
Overseas: The Profitability Engine Most Investors Are Underweighting
Vice Chairman Brian Gu, speaking from XPeng's manufacturing partner in Austria, offered the clearest articulation yet of why international markets matter financially, not just strategically. International vehicle sales are "significantly better, even with obviously some of the tariff issues," generating "significant better gross profit as well as net profit contribution to our bottom line" relative to domestic sales. International deliveries already represent close to 20% of monthly volume, up from roughly 10% for full-year 2025. Management expects international revenue to exceed 20% of total revenue starting in Q2, and the target is to sustain that 30%-plus level through year-end as new global models launch.
The manufacturing footprint is expanding to support this: assembly operations in Indonesia and Malaysia serve local Southeast Asian demand, the Magna partnership in Austria covers Europe, and XPeng expects the majority of European sales to carry locally manufactured content. He Xiaopeng set a five-year ambition of 50% of revenue and profit from overseas markets, which, if achieved, would represent a fundamental transformation of the company's financial profile. The near-term milestone is monthly overseas deliveries exceeding 10,000 units in Q4 and more than doubling full-year international volumes.
VLA 2.0 Scaling Law Thesis and the August Inflection
XPeng continues to make the case that its Vision-Language-Action model is compounding in a manner consistent with AI scaling laws, and the April data point — ADAS mileage penetration on VLA 2.0-equipped vehicles surpassing 50% for the first time — is the empirical foundation for that argument. He Xiaopeng described VLA 2.0 as "a key reason customers choose XPeng, creating a strong and lasting user mindshare channel in the market."
The next scheduled capability upgrade arrives in Q3, specifically targeting August, where management expects "a significant performance improvement" in what the CEO characterized as the subjective feel of the system — smoother, smarter, better generalization — rather than the safety and engineering basics that VLA 1.0 prioritized. A second OTA release is planned for year-end. He Xiaopeng made a forward-looking claim that deserves attention: "I think that with gradual R&D development, we will be able to achieve L4 capability with 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 in all shape or form of the whole business." He explicitly declined to elaborate further, flagging this as work in progress.
XPeng is also testing VLA 2.0 in Europe and hopes to receive regulatory approval in multiple countries in 2027, which would allow overseas delivery of the technology and serve as the foundation for international robotaxi deployment.
Robotaxi: A Post-2027 Commercial Story, Not a 2026 Catalyst
On robotaxi, He Xiaopeng calibrated investor expectations carefully. Current operations are confined to Guangzhou under existing licensing, and despite a widely noted regulatory tightening in China's autonomous vehicle rules, management says it has not materially disrupted the development rhythm. The honest framing was that "after 2028, we're going to expect a huge commercial opportunity for robotaxi." In 2027, XPeng plans to launch a dedicated economy-car platform to validate the robotaxi business model domestically.
The business model XPeng intends to pursue is an asset-light one: it will sell vehicles and technology to operating partners, take a commission on deployments, and avoid directly running taxi fleets. He Xiaopeng noted that since announcing the GX as a pre-installed, mass-produced robotaxi-capable vehicle — the first of its kind in China with full hardware redundancy — the company has received significant inbound interest from prospective operating partners both domestically and internationally. The regulatory and commercial scale, however, remains a second half of the decade event in management's own telling.
Humanoid Robots: Automotive-Grade Ambition, Year-End Mass Production Target
XPeng's humanoid robot program, named Ren, was the most substantively detailed it has been on an earnings call. Management set a year-end mass production target for Ren, with initial trial commercial deployments in XPeng's own showrooms, followed by commercial customer deliveries in China and overseas in 2027. A next-generation version featuring multilingual communication, human-like full-body motion, and gradually autonomous professional task execution is slated for a Q3 showcase.
He Xiaopeng made the competitive differentiation argument around manufacturing quality rather than raw capability: most competing humanoid robots are built to consumer-product grade tolerances, whereas XPeng is designing Ren to automotive-grade safety and reliability standards, incorporating existing automotive supply chain partners as component suppliers. The company has completed development of a proprietary next-generation dexterous hand that is described as significantly more agile and substantially cheaper than predecessors.
On the economics, He Xiaopeng was candid that the cost structure of a humanoid robot today resembles that of a car, but argued that hardware margins will be superior to automotive because of the compute density embedded in each unit, with additional software licensing and cloud model revenue on top. On the question of overseas versus domestic economics, the CEO was direct: "In the overseas market, the payback period can be much shorter than that of what you would achieve in China, so it would make more sense economically for overseas business owners to buy humanoid robots." He framed overseas markets as likely generating greater commercial returns than domestic, mirroring the dynamic already emerging in the EV business.
R&D Trajectory and Balance Sheet
R&D spending rose 46.8% year-on-year to RMB 2.91 billion in Q1 2026, and management signaled this is a deliberate acceleration tied to the physical AI investment thesis rather than a temporary spike. The company ended Q1 with a cash position of RMB 42.09 billion, which provides meaningful runway for the investment cycle ahead. Starting in Q2, XPeng will begin delivering its Turing SoC to Volkswagen at scale, and VP Charles Zhang confirmed that technology and services revenue guidance for full-year 2026 is comparable to 2025 levels, with the Volkswagen supply relationship described as "a very attractive business." Management characterized the company as "quite open minded" to expanding similar technology commercialization arrangements beyond the current Volkswagen partnership.