Update (August 29, 2026): XPeng reported second-quarter 2026 results as scheduled on Monday, August 24, 2026, before market open, with the call at 8:00am ET. This article originally published as a pre-earnings preview; it has been rewritten below with the actual reported numbers and the stock’s reaction. Revenue rose but missed analyst estimates, the net loss widened well beyond what Wall Street expected, and the stock fell roughly 7% on the day, even as the company’s robotics unit landed one of the largest funding rounds in Chinese AI history.
Key Takeaways
- XPeng (XPEV) reported Q2 2026 results Monday, August 24, 2026, before market open; shares fell roughly 7% on the day.
- Revenue reached RMB19.74 billion, up 8% year-over-year, inside the company’s own RMB19.60-20.80 billion guidance but below the roughly RMB20.57 billion analysts had modeled.
- Deliveries hit 103,295 vehicles, up 65% sequentially from Q1’s 62,682, but essentially flat year-over-year (+0.1%), a detail the market only learned on report day.
- Gross margin improved to 20.7%, up from 17.3% in the same quarter a year earlier, even as the per-share loss came in far wider than analysts expected.
- XPeng’s robotics unit (home to the IRON humanoid robot) raised more than $900 million at a valuation above $6.3 billion, reportedly the largest single-round financing yet in China’s embodied-AI industry, the report’s clearest bright spot.
What actually happened Monday
XPeng confirmed the results as scheduled, corroborated by the company’s own press release and independent coverage from Yahoo Finance, Investing.com, and Morningstar. Revenue for the quarter came in at RMB19.74 billion, up 8% from a year earlier and landing inside the RMB19.60-20.80 billion range XPeng itself had guided to before the print, but below the roughly RMB20.57 billion analysts had modeled. The market prices against consensus, not a company’s own guidance range, and the miss against that higher analyst bar is what mattered. XPeng also reported a per-share loss well wider than the loss analysts expected heading into the print. Shares fell approximately 7% on the day, per multiple market reports, as investors weighed the miss and a softer Q3 2026 outlook against the quarter’s one clear positive: a landmark funding round for the robotics business.
The delivery number: known ahead of time, but flat year-over-year
The 103,295-unit delivery figure was already public before the report. XPeng pre-announces deliveries ahead of formal earnings, standard practice across Chinese EV makers. What wasn’t public until Monday: that figure represented essentially no year-over-year growth (+0.1%), even though it was a 65% jump from Q1’s seasonally weak 62,682 units. The sequential rebound was real, but the flat year-over-year comparison is the more important context for anyone who read the delivery headline in isolation back in July.
Gross margin, the number this article flagged pre-earnings as the more important swing factor than the delivery count, came in at 20.7%, up from 17.3% in the same quarter last year. That improvement held even as the company posted a much wider net loss than expected: the core vehicle business is getting more efficient per unit sold, while spending elsewhere (R&D, robotics, international expansion) pushed the bottom line well below what analysts modeled.
The quieter positives: overseas deliveries and services revenue
Two figures got less attention than the headline miss but matter for anyone tracking XPeng’s business mix. Overseas deliveries surpassed 20,000 units for the first time in a single quarter, up 81% year-over-year, and now account for roughly a quarter of first-half revenue. Services and other revenue (technical R&D services, parts sales) grew 94% year-over-year to about RMB2.70 billion. It’s a smaller line item than vehicle sales, but one of the fastest-growing pieces of the business and a sign the company is building revenue streams beyond the vehicle-delivery number that dominates headlines.
The report’s bright spot: a landmark robotics funding round
XPeng’s robotics division, home to the IRON humanoid robot line previewed at the company’s November 2025 “Emergence” event, raised more than $900 million in its first funding round, valuing the unit at more than $6.3 billion post-money. Multiple market reports described it as the largest single-round private financing yet recorded in China’s embodied-AI industry. Alongside the funding news, XPeng’s L03 model reportedly secured a historic order volume at launch, with overseas deliveries of that model expected to begin in Q4.
None of this shows up in Q2 vehicle revenue, and it’s a separate business line from the EV operation investors are still valuing the stock on primarily. But it’s the detail that keeps XPeng’s robotaxi and humanoid-robot ambitions from reading as a roadmap slide: an outside investor round of this size, at this valuation, is a real, priced bet on the business.
What the options market got right, and what it didn’t
Ahead of Monday’s print, options traders were pricing an implied move of roughly 9.4% in either direction for XPEV, above the stock’s historical post-earnings average move. The actual reaction, a roughly 7% decline, landed inside that priced range: the options market sized the magnitude of uncertainty reasonably well, even though it couldn’t call the direction or the specific mix of a revenue miss, a wider-than-expected loss, and an offsetting robotics headline.
That’s the general lesson worth carrying into the next print, for XPEV or any other stock: the at-the-money straddle (the closest-to-the-money call and put expiring right after the event, combined premium times roughly 0.85) estimates the size of an expected move, not its direction or its cause. A trader who only reads a pre-announced delivery number and assumes the print is fully priced in can still be surprised by a story like this one, where a funding round for a barely-revenue-generating robotics unit ends up mattering to the stock almost as much as the quarterly numbers themselves.
Whoever is checking a future XPEV print, or any other name’s, close to the report date should pull the current front-week ATM straddle directly from a live options chain rather than relying on a prior quarter’s number. A tool like Webull shows the full options chain and real-time implied volatility for free, the fastest way to check where the market is pricing an upcoming move before deciding whether an options structure is worth the premium.
Context matters here too: XPEV has traded in a wide 52-week range, and China-ADR names as a group have carried an elevated volatility premium through 2026 on top of company-specific news, the same dynamic covered in TRDC’s Alibaba earnings options setup and PDD Holdings earnings options setup. A US-listed China ADR earnings event isn’t just pricing the print, it’s pricing regulatory and delisting-related headline risk that a purely domestic company doesn’t carry, part of why the implied move on these names tends to run hotter than a similarly sized US company’s, and part of why this quarter’s 9.4% priced move was itself elevated relative to XPeng’s own trailing average.
Reviewing two hypothetical structures against the actual move
These are illustrative examples only, not trade recommendations, reviewed after the fact purely to show how each structure interacts with a real outcome rather than an assumed one.
| Structure | View expressed | Risk profile | Best fit |
|---|---|---|---|
| Long straddle/strangle (hypothetical) | The actual move will exceed what’s priced in | Defined risk (premium paid), unlimited theoretical reward | A trader who thinks a specific catalyst creates two-sided surprise potential beyond the historically priced range |
| Iron condor around the expected move (hypothetical) | The stock stays inside the market’s priced range | Defined risk and defined reward, both capped | A trader who thinks the known information is already priced, and new details won’t move the stock as much as the options market expects |
Take the earlier hypothetical from this article’s original pre-earnings version: XPEV trading near $14 with an implied 8% move priced a range of roughly $12.90 to $15.10, illustrative numbers built from the stock’s price level heading into the print, not this specific report’s actual inputs. The real result, a roughly 7% move, would have landed inside that illustrative range: a long strangle bought outside it wouldn’t have cleared its breakeven, while an iron condor sold around it would have kept its premium. That’s a useful illustration of how these structures behave when a move comes in near the priced expectation rather than beyond it, not a claim about what any specific position actually returned, since real strikes, premiums, and fills vary by when and how a trade was placed.
Bottom line
XPeng’s Q2 2026 report was a genuine mixed bag: revenue grew but missed consensus, the net loss came in far wider than expected, and the stock fell roughly 7%, all while the delivery number the market already knew turned out to be flat year-over-year rather than the growth story the sequential jump implied. The one clear win, a landmark funding round for the robotics unit, didn’t show up in the quarter’s financials at all. The options market priced a 9.4% move and got the magnitude roughly right without knowing which of these threads would end up driving the reaction.
FAQ
Q: What were XPeng’s actual Q2 2026 results?
A: Revenue of RMB19.74 billion, up 8% year-over-year but below the roughly RMB20.57 billion analysts expected. The company reported a per-share loss well wider than what analysts had modeled. Gross margin improved to 20.7%, up from 17.3% a year earlier.
Q: How did XPEV stock react?
A: Shares fell roughly 7% on report day, per multiple market reports, as investors weighed the revenue miss and wider loss against a softer Q3 2026 outlook. Exact intraday figures varied slightly by source and by whether pre-market or full-day trading was measured.
Q: Didn’t XPeng already report its delivery numbers before earnings?
A: Yes, XPeng pre-announced 103,295 vehicles delivered, standard practice for Chinese EV makers. What wasn’t known until the formal report: that figure was essentially flat year-over-year (+0.1%), even though it was a 65% sequential jump from a seasonally weak Q1.
Q: What was the one clearly positive part of the report?
A: XPeng’s robotics division raised more than $900 million at a valuation above $6.3 billion, reportedly the largest single-round private financing yet in China’s embodied-AI industry. It’s a separate business line from the EV operation and doesn’t show up in Q2 revenue.
Q: How is the options market’s “expected move” actually calculated, and did it work here?
A: Take the price of the at-the-money call plus the at-the-money put expiring closest to the event, multiply by roughly 0.85; that approximates the one-standard-deviation range the market is pricing for the stock’s move. Ahead of this report, that priced roughly a 9.4% move; the actual roughly 7% move landed inside that range, meaning the market’s estimate of magnitude was reasonable even though it couldn’t predict the specific mix of a revenue miss and a robotics funding headline.
For more on how earnings-day option pricing works and where it tends to go wrong, see TRDC’s IV crush explained with real examples guide.