Somewhere in Guangzhou right now, Xpeng is trying to pull off three separate miracles inside the same twelve months: mass produce a flying car, mass produce a humanoid robot, and put a fleet of self driving taxis on public roads. For anyone not familiar with the Chinese auto market, this sounds like a pitch deck written by someone who has never actually had to build anything.
It is, in fact, the roadmap of a man who has already built two companies from nothing, sold one for billions, and nearly ran the other into the ground before turning it into one of China’s most technically ambitious automakers. That man is He Xiaopeng, and the road from mobile browser executive to flying car mogul is one of the stranger origin stories in Chinese tech, even by the industry’s own high bar for strangeness. Here’s all you need to know about Xiaopeng and his Xpeng brand.
Winning the Browser Wars
Long before “smart EV” meant anything, He Xiaopeng was fighting a much smaller war over a much smaller screen. Born in 1977 in Huangshi, Hubei province, he studied computer science at South China University of Technology and spent several years at a telecom software firm before, in 2004, co-founding UCWeb with Liang Jie, a fellow graduate of the same university.
UCWeb built mobile browsers just as China’s internet was moving off the desktop, and its flagship product, UC Browser, became the default way huge numbers of people in China, Indonesia, and India got online through low-spec phones and patchy networks. Much of that success came down to engineering rather than marketing: UCWeb built its own compression and rendering engine, the U3 kernel, to make pages load fast on hardware that had no business handling them, and it spent a decade fending off domestic rivals like Tencent’s QQ Browser and international ones like Opera Mini.
By 2014, UCWeb was valuable enough that Alibaba, which already held a stake, bought out the rest of the company in a deal reported at the time as the largest merger in Chinese internet history. He stayed on inside Alibaba for three years, running its mobile business group, chairing Alibaba Games, and overseeing the video platform Tudou. It was, by any normal measure, a soft landing: wealthy, senior, comfortable. He left anyway, walking away from Alibaba entirely in August 2017 to go run a small, struggling electric car startup that, as it happened, already carried his name.
A Startup That Already Had His Name on It
The actual founding story is messier than the tidy version usually told. Xpeng was registered in August 2014 by Xia Heng and He Tao, two engineers who had cut their teeth at state automaker GAC Group, with early money coming from He Xiaopeng, still an Alibaba executive at the time, and from Xiaomi founder Lei Jun.
He Xiaopeng effectively bankrolled the venture and lent it his name well before he ran it day to day, taking the chairman and chief executive titles only once he cut ties with Alibaba in 2017. It’s a founding myth that flatters the self-made narrative a little less than Xpeng’s marketing usually implies, but it also explains a lot about how the company was built: automotive engineering discipline from Xia and He Tao, welded onto the internet-platform instincts and capital He brought from his UCWeb years.
The early years were rough. Xpeng built just fifteen units of a concept car called Identity X before its first real production vehicle, the compact G3 crossover, reached Chinese roads in 2019, built under contract at Haima Automobile’s factory because Xpeng didn’t yet have its own plant. The P7 sedan followed, delivered from 2020, and gave the company its first real design and range credibility against domestic rivals.
Xpeng went public on the New York Stock Exchange in August 2020, pricing at fifteen dollars a share and popping more than 40 percent on debut, then followed with a secondary listing in Hong Kong in 2021, becoming the first Chinese smart EV maker with dual primary listings on both exchanges.
Crash, Rebuild, Repeat
Going public did not make the business easy. Between 2022 and 2023, Xpeng’s sales stalled badly enough that its stock cratered and He’s personal fortune, tied up almost entirely in company shares, reportedly fell by around 80 percent from its 2021 peak. The company responded with a real overhaul.
In 2023 it bought Didi’s smart-car development unit for roughly 744 million dollars in stock, a deal widely credited as the technical seed for the budget-focused Mona sub-brand, and it sold Volkswagen a stake just under five percent for 700 million dollars, a vote of confidence in Xpeng’s driver-assistance software from a company that actually knows how to build cars at scale.
The bet paid off in stages. The Mona M03 and its SUV sibling, the L03, became genuine volume hits, and by the fourth quarter of 2025, Xpeng posted its first ever profitable quarter on record deliveries.
That recovery hasn’t been a straight line since, which is where the real story diverges from the press release. In the second quarter of 2026, Xpeng delivered a record 103,295 vehicles and grew revenue 8 percent year over year to nearly 20 billion yuan, yet its net loss nearly tripled to 1.34 billion yuan, because everyone in China’s EV market is still slashing prices to move metal.
Vehicle margin slipped from 14.3 to 12.1 percent even as overall gross margin held above 20 percent, a level Xpeng says puts it ahead of both Tesla and BYD on that specific metric. Only three Chinese EV brands turned a profit at all in the first half of 2026. It’s a useful reminder that record deliveries and financial health aren’t the same sentence in China’s EV market right now, and Xpeng is spending heavily on autonomous driving, robotics, and flying cars precisely while its core car business still fights for margin.
Beyond China
Xpeng’s answer to that domestic bloodbath has been to get bigger somewhere else. It entered Norway in 2021, its first European market, and has since expanded to roughly 28 European countries, entering the UK in February 2025 with a right-hand-drive version of the G6 sold through International Motors.
By mid-2026 the company operated around 380 stores across its overseas markets, more than 290 of them across those 28 European countries alone, and it’s increasingly building locally rather than just shipping finished cars: an Indonesian plant turning out the X9 from July 2025, an Austrian plant near Graz run with Magna, marking the first time that contract manufacturer had ever assembled cars for a Chinese automaker, and a right-hand-drive plant in Malacca, Malaysia, that came online in June 2026.
Overseas deliveries in the second quarter of 2026 topped 20,000 units, up 81 percent year over year, and now account for more than a quarter of total revenue at an average export price above 40,000 euros, notably richer than Xpeng’s typical price at home. The company has also rolled out extended-range EV versions of the G6, G7, and P7+, partly to sidestep the fully-electric-only tariffs some Western governments have slapped on Chinese EVs.
The Flying Car Is Actually Happening
Which brings us to the part of this story that sounds like science fiction and is, improbably, the closest to real. Xpeng’s flying car unit traces back to 2013 and was formalized in 2020 as a majority-owned subsidiary then known as AeroHT.
Its current flagship, the Land Aircraft Carrier, pairs a 5.5-meter, six-wheeled “mothership” van, built on an 800-volt silicon carbide range-extender platform good for more than 1,000 kilometers of combined range on China’s CLTC cycle, with a detachable two-seat eVTOL aircraft that folds into the van’s rear bay and deploys on command.
It received airworthiness type certification from Chinese aviation regulators in 2024, rolled the first units off a purpose-built Guangzhou assembly line in late 2025 that the company calls the world’s first mass-production line for flying cars, and carries a price cap of roughly 2 million yuan, a little under 300,000 dollars.
In October 2025, the unit rebranded entirely as Aridge, distancing itself from the Xpeng name as it courts outside investors ahead of a planned Hong Kong IPO, and flew a manned demonstration flight in Dubai, its first public manned flight anywhere outside China.
That trip came with an order for 600 units from a consortium of Gulf distributors, reportedly the largest single overseas order the flying car industry has seen, plus a new long-range sibling called the A868, a tiltrotor hybrid eVTOL aimed at business travel with a claimed range over 500 kilometers and a six-seat cabin.
Customer deliveries of the Land Aircraft Carrier are still targeted for late 2026, which, if Xpeng actually hits it, would make this one of the only flying car programs anywhere to move from concept renderings to a real production line rather than staying a trade-show curiosity. Airspace regulation everywhere outside China remains the real bottleneck, and a healthy dose of skepticism is warranted until paying customers are actually flying these things to work.
Betting the House on Physical AI
The flying car is, in some ways, just one leg of a broader thesis. Xpeng now describes itself as a physical AI company as much as a carmaker, built around its self-developed Turing chip, a 40-core processor rated at 750 trillion operations per second that shows up in its cars, its robotaxi hardware, and its humanoid robot program.
That robot, called Iron, appeared in a more lifelike second generation in late 2025, with a biomimetic spine, twenty-two degrees of freedom across its hands, and a curved display standing in for a face, running on three Turing chips for a combined 2,250 TOPS.
Xpeng is targeting large-scale production of Iron by the end of 2026 and raised more than 900 million dollars for the robotics unit in mid-2026 at a 6.2 billion dollar valuation, backed by IDG Capital alongside Tencent and, in a nice bit of symmetry, Alibaba, the same company that once bought He’s browser business. A robotaxi push run with Alibaba’s Amap mapping service is also slated to begin trial operations this year.
Our Take
He Xiaopeng has been unusually candid, for a Chinese auto executive, about how brutal all of this is. Over the past two years he has repeatedly warned that China’s EV market is nowhere near finished shaking out, predicting in one interview that only seven major Chinese carmakers would survive the next decade out of the hundreds that once tried, then sharpening that forecast last year to just five survivors within five years.
That the timeline keeps getting shorter and more urgent, not less, says something about how he actually views the state of play, regardless of how the marketing copy reads. The Elon Musk comparisons follow him everywhere, mostly because both men funded an EV company with proceeds from an earlier tech exit and both now chase robots and autonomy with equal enthusiasm. The comparison sells him a little short. Musk rarely dwells publicly on how much of his own industry might not survive; He does it routinely, about a market his own company still has to win.
It’s the same restless, pattern-recognizing instinct that once bet on mobile browsers before most of China owned a smartphone worth browsing on. Two decades later, it’s betting on flying cars and humanoid robots before most of the world takes either seriously. Whether Xpeng ends up on the right side of the shakeout he keeps predicting, while simultaneously trying to ship flying cars, humanoid robots, and robotaxis in the same calendar year, is the actual story worth watching from here.
Hillary started his automotive writing journey at HotCars, and has written for CarNewsChina, GlobalSUV, and many other top auto blogs.
He loves to read, play chess, and supports Liverpool during the weekends
