In February 2026, Volkswagen did something no global automaker had done before: it agreed to license Level 4 autonomous driving software from a Chinese company. From XPeng, a Guangzhou-based EV maker that did not exist until 2014. The deal built on a $700 million equity stake VW had already taken in XPeng back in 2023, and on a jointly developed electronic architecture that will now underpin not just Volkswagen’s China lineup, but six or more global models. XPeng’s own CEO called it a moment with meaning “far beyond the cooperation itself.” He was not exaggerating.
For an industry that spent four decades assuming technology flowed from West to East, that is a genuinely strange sentence to write. Volkswagen entered China in 1984 to build Santanas in Shanghai and, in the process, teach a struggling domestic industry how modern cars get made. Four decades later, it is paying one of its own students to teach it how to build a modern car.
That reversal did not happen by accident, or at least not the kind of accident anyone in Wolfsburg, Detroit, or Tokyo intended. It is the long, slow payoff of a bargain Western automakers struck with Beijing back when China’s auto industry barely existed.
The Bargain
China’s modern car industry effectively begins with Beijing Jeep, a 1983 joint venture between Beijing Automotive and American Motors Corporation, majority owned by the Chinese side from day one. Shanghai Volkswagen followed in October 1984, with the first Santana rolling off the line the following September.
A wave of others arrived through the late 1980s and 1990s under a policy framework the government called “three large, three small, two mini,” which paired approved domestic manufacturers with approved foreign partners: Toyota with FAW and GAC, Honda with Guangzhou and Dongfeng, Nissan with Dongfeng, Hyundai with Beijing Auto, BMW with Brilliance.
In 1994, Beijing formalized the arrangement in its Automobile Industry Policy: any foreign automaker that wanted to build passenger cars in China had to do it through a joint venture, capped at 50 percent foreign ownership, with no more than two such partnerships per vehicle segment.
The logic on both sides was straightforward. China had almost no modern manufacturing base and wanted one, along with the management practices, supplier networks, and quality-control systems that came with it. Foreign automakers had exactly that expertise, and in exchange for sharing it, they got a front-row seat to what was obviously going to become the largest car market on earth.
Volkswagen, General Motors, Toyota, Honda, Nissan, Hyundai, and BMW all signed up. SAIC, FAW, Dongfeng, GAC, Changan, and Beijing Auto became the Chinese half of nearly every major nameplate sold in the country for the next thirty years.
What the Joint Ventures Actually Taught
The direct technology transfer the JV policy was designed to force turned out to matter less than the ecosystem it accidentally built around it. Stanford researchers who studied the policy found real quality gains at JV-affiliated models, roughly 4 to 20 percent better than unaffiliated models between 2007 and 2014. But they also concluded the overall effect was modest.
Automakers with no foreign partner at all gained almost as much over the same stretch, through industry-wide spillover that had nothing to do with any formal joint venture agreement. The knowledge did not stay locked inside the partnerships. It spread through the supplier base, the contract manufacturers, the engineers who moved between companies, and the precision manufacturing standards that foreign automakers had to import just to hit their own quality bar.
Meanwhile, the Chinese partners were doing something the original policy never quite anticipated: building their own in-house brands alongside the joint venture, funded in part by JV profits and staffed in part by JV-trained engineers. SAIC built Roewe, then took over the MG name outright. FAW revived Hongqi as a serious brand rather than a government limousine line. Dongfeng launched Voyah. GAC built Aion and Trumpchi. Changan spun up Avatr and Deepal.
All of this was built on institutional knowledge that thirty years of joint venture manufacturing had put inside Chinese companies, whether or not that was ever the plan.
The Reset Nobody Saw Coming
For a long time, none of this looked especially threatening to the foreign partners, because the game was still combustion engines, and that is where Western and Japanese automakers had a century of head start that no amount of shop-floor osmosis could fully close.
Electrification changed the terms. Beijing poured subsidies and mandates into new energy vehicles for over a decade, and battery makers like CATL and BYD’s own in-house battery arm built a manufacturing and cost advantage that owed nothing to any joint venture. Once the industry’s center of gravity shifted from engines to batteries, software, and electronic architecture, the JV brands were still running decision-making processes designed in Wolfsburg or Detroit, on platforms built for a different era.
Their Chinese counterparts, both the JV partners’ own house brands and fully independent players like BYD, Geely, Nio, and XPeng, iterated at a pace the old structure was never built for. Volkswagen’s own recent numbers make the point better than any outside analyst could. Its newest China-market EV, developed on architecture shared with XPeng, went from drawing board to production line in 24 months, a fraction of the four or five years such projects traditionally took.
The Reckoning, By the Numbers
The scale of the reversal became undeniable in 2026. In June, the combined market share of joint venture and foreign brands in China fell below 25 percent for the first time on record, down from more than 60 percent as recently as 2020. Volkswagen’s China deliveries fell 26 percent in the first half of the year, to 971,000 units. Honda dropped 35 percent. Toyota fell 17 percent. BMW slid 20 percent.
Changan Ford, once a real presence in the market, sold just 28,800 cars in six months. Chinese brands, meanwhile, crossed 70 percent market share for the first time in the same period.
Tellingly, the one foreign nameplate that kept growing was Tesla, up more than 28 percent in the first half, and that is worth sitting with for a second. Tesla is not a joint venture. It owns its Shanghai plant outright, a structure that only became legal for passenger cars in 2022, when China finally lifted the ownership cap it had imposed on itself back in 1994.
The lesson is not that foreign brands cannot win in China. It is that the joint venture model itself, built for a market China no longer resembles, has become the liability.
Full Circle
The sharpest irony is what is happening outside China. SAIC, the joint venture partner Volkswagen and General Motors have both relied on for four decades, posted 2.045 million global sales in the first half of 2026, the first time any Chinese automaker has cleared two million units in six months.
A large chunk of that came from MG, the once-British marque SAIC now owns outright, which sold more than 190,000 cars in Europe in that same half year alone, on pace to beat its 307,000-unit full-year total from 2025. BYD is chasing hard, targeting more than 50,000 sales in Germany this year through steep discounting, aiming to knock MG off its perch as Europe’s top Chinese brand.
Chery’s Omoda, Jaecoo, and Jetour brands grew 267 percent across the EU in the first four months of 2026. Leapmotor grew 559 percent in the same window, distributed through a joint venture with Stellantis, the company that today owns the Jeep brand: the same Jeep that partnered with Beijing Automotive back in 1983 to start this whole story.
Brussels has noticed. The EU imposed countervailing duties of up to 35.3 percent on China-made EVs in October 2024, and in January 2026 pivoted toward a minimum price floor system instead. It echoes the exact market-access leverage China itself used in 1994, deployed in reverse. SAIC is already building a plant in Spain, partly to sidestep the tariff that hits it hardest. It has learned the precise lesson the original joint venture policy was built to teach: if you want to sell inside a protected market, build inside the wall.
Our Take
None of this means Western automakers are finished in China, or that Chinese brands will simply repeat the takeover in Europe and North America unchallenged. Tariffs, trade politics, and real gaps in areas like driving feel and long-term reliability data still favor the incumbents in plenty of markets.
But the joint venture bargain that once looked like a straightforward trade of market access for manufacturing know-how has resolved into something closer to a masterclass, taught patiently over thirty years, that the students have now finished.
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
