XPeng's Billion-Dollar Robot Arm Can't Distract From a Bleeding Auto Core

XPeng's EV losses triple in Q2 and guidance misses, while robotics unit raises $900M at $6.3B valuation—market remains skeptical.

Investors are being asked to weigh two very different versions of XPeng this week. One is a robotics division that just pulled in more than $900 million from blue-chip backers at a valuation north of $6.3 billion. The other is an electric-vehicle business whose losses nearly tripled in the second quarter and whose guidance for the months ahead fell well short of Wall Street's hopes.

The market's verdict on Monday was unambiguous: the shares dropped around 7 percent, with another 9 percent shaved off in Hong Kong trading the following day. For all the fanfare around the humanoid-robot story, the market is pricing the struggling core business more heavily than the robotics optionality — at least for now.

The Numbers Beneath the Surface

XPeng's headline figures for the April-to-June period looked respectable at first glance. Revenue climbed 8 percent year over year to 19.74 billion yuan, a sharp 51.5 percent jump from the first quarter. Deliveries reached 103,295 vehicles, while gross margin held above the psychologically important 20 percent threshold at 20.7 percent.

But the profit-and-loss statement tells a less flattering story. The net loss ballooned to 1.34 billion yuan, up from 480 million yuan in the same period a year earlier. The adjusted loss per ADS of 1.29 yuan came in well above the FactSet consensus estimate of 0.91 yuan — a miss that underscores how expensive growth has become.

The vehicle margin, a metric analysts watch obsessively, slipped to 12.1 percent from 14.3 percent in the prior quarter. Citi points to delivery bottlenecks around the MONA L03 model as a key drag. That margin pressure matters because XPeng's own third-quarter revenue guidance of 21.7 billion to 23.4 billion yuan sits noticeably below the roughly 25.9 billion to 26.7 billion yuan the market had penciled in. The delivery forecast of 115,000 to 121,000 vehicles implies sequential improvement, but it still undershoots expectations — and puts the company's stated goal of 60,000 monthly deliveries by year-end in serious doubt.

The Robot That's Worth More Than the Car Business

Against that sobering operational picture, the robotics funding round stands out as the brightest spot. Led by IDG Capital, with Tencent, Alibaba and Gaorong Ventures participating, the round values the Dogotix subsidiary at over $6.3 billion — the largest single raise in China's embodied-AI sector to date. The IRON humanoid robot, equipped with 76 degrees of freedom and three Turing AI chips, is slated for series production by the end of 2026, with external commercial deliveries beginning in 2027.

Goldman Sachs calculates that the robotics stake alone represents roughly 53 percent of XPeng's entire $11.8 billion market capitalization. Citi values the remaining core business at around $6.5 billion — a split that suggests the market isn't fully adding up the two segments. XPeng retains about 82 percent of Dogotix after the round, or 68.41 percent if all option rights are exercised, meaning the parent remains the primary beneficiary if the IRON robot hits its production targets.

Morgan Stanley cautions that the valuation sits at the upper end of comparable private humanoid-robotics peers, though it doesn't see that as inconsistent with the market's overall size. Seeking Alpha has issued a buy rating with a $22 price target, citing the robotics optionality alongside upcoming model launches like the G9L and Mona L05 in the fourth quarter.

A Cautionary Tale From Unitree

The bear case draws directly from a recent precedent. Unitree, a fellow robotics player, has seen its shares collapse 45 percent from their peak since its August 19 stock market debut, wiping out $30 billion in market capitalization. The lesson, as some strategists frame it: embodied-AI valuations can correct violently once operational reality collides with hype. XPeng has no commercial robot sales yet — just a production plan — which leaves the robotics narrative vulnerable to exactly that kind of repricing.

There are also broader concerns about stretched AI valuations, with market strategists like Abby Joseph Cohen warning that the sector as a whole may be running ahead of fundamentals.

The Bull Case Has a Pulse

Optimists counter that the current price already discounts a lot of bad news. The stock sits roughly 45 percent below its level at the start of the year, and about 32 percent under its 200-day moving average. At 11.40 US dollars following Monday's slide, it trades 59 percent below its 52-week high of 24.40 euros.

There are genuine bright spots beneath the gloom. International deliveries jumped 81 percent in the quarter to over 20,000 units, with overseas revenue accounting for more than a quarter of total sales in the first half. The services segment grew 94 percent to 2.70 billion yuan, driven by technical development work for another automaker — evidence that XPeng's revenue streams are diversifying beyond vehicle sales.

The stock did stage a partial recovery on Tuesday and Wednesday, climbing 4.2 percent from the prior close of 9.95 euros. Whether that marks a genuine reassessment of the robotics option or merely a pause in the selloff remains an open question.

What to Watch Next

The immediate test comes with third-quarter delivery numbers, which will show whether XPeng can hold the line at the lower end of its 115,000-to-121,000-vehicle range. The G9L and Mona L05 launches in the fourth quarter will also be closely scrutinized for signs that the product cycle can regain momentum.

If vehicle margins stabilize and deliveries track toward the guidance range, the bull case of an undervalued dual structure — core auto business plus robotics — remains intact. If margins deteriorate further or XPeng misses even its reduced delivery forecast, the market may start questioning the Dogotix valuation with the same skepticism it applied to Unitree. The robotics story is real and backed by serious capital, but its commercial payoff is still two years away. In the meantime, the core business has to stand on its own — and right now, that's the part that's wobbling.