Representative image of a human in a self-navigating Tesla vehicle.
Ni Tao is IE’s bi-weekly columnist, giving exclusive insight into China’s technology and engineering ecosystem. His Inside China column explores the issues that shape discussions and understanding about Chinese innovation, providing fresh perspectives not found elsewhere.
The recent entry of Tesla’s Full Self-Driving (FSD) technology into China has got the country’s automobile industry buzzing with speculation about what’s in store for it.
Tesla has long envisioned expanding its FSD operation into China. Until recently, it was only available in the US, Australia, Canada, and a handful of European countries.
As an early adopter of end-to-end (E2E) autonomous driving solutions, Tesla has made significant strides since the 2021 release of its FSD Beta, with the latest offering being FSD V12.
Tesla’s database edge is enormous, thanks to its acquisition of over a billion miles of driving data globally and more than 10 million real human driving videos. The company has leveraged these inputs to train its algorithms using neural networks, creating a data pool that Chinese carmakers and autotech firms can only envy.
In this sense, Tesla’s FSD does pose a significant threat. But does that signal a looming demise of these domestic players? The reality is more complex than meets the eye.
China’s autonomous driving landscape consists of two types of players. One group, which includes EV makers like Xpeng and Li Auto and DeepRoute.ai, a tech startup, adopts a technical approach similar to Tesla’s E2E.
They use AI models to process data and directly generate control commands, enabling vehicles to recognize road conditions and pilot themselves. However, unlike Tesla’s camera-based visual solution, Xpeng and Li Auto also equip their cars with lidars, millimeter-wave radars, and cameras to collect data.
The other approach relies on multi-sensor fusion perception technology, with Huawei as a major proponent. Huawei’s autonomous driving system ADS2.0, available in car models like AITO M7 and M9, integrates lidars, millimeter-wave radars, and cameras to enhance perception.
It allows L4 autonomy with low highway takeover rates and includes features like autonomous valet parking.
Huawei bets big on multi-sensor fusion perception because it believes that camera, millimeter-wave radar, and lidar have their shortcomings, making integration necessary.
The arrival of Tesla’s FSD has elicited mixed reactions. Some see it as a threat, while others welcome it enthusiastically. Companies like Xpeng and DeepRoute.ai, which embrace the same E2E approach, stand to benefit. It’s no wonder Xpeng’s founder and CEO, He Xiaopeng, openly welcomed Tesla’s arrival.
Even Huawei, Tesla’s arch-rival, positively views the coming clash with FSD, suggesting it could help push the industry forward. That said, Richard Yu, chairman of Huawei’s smart vehicle affiliate, has boasted that Huawei’s enhanced lidar-driven perception capabilities give it an edge over Tesla.
Against this backdrop, the competition between Tesla and the pack of domestic players is set to be fierce. Who will finally emerge on top is anyone’s guess, and it’s premature to say the pack will gang up on Tesla or Tesla will take all their market share.
My prediction is different: Both vision-based and multi-sensor fusion technologies will coexist in China. Despite their current advantages, a “winner-takes-all” scenario is unlikely for the E2E adopters.
In fact, these technologies may converge as lidar and other sensors become more mature and cost-effective, potentially leading Elon Musk to reconsider his previous dismissal of lidar as a “fool’s errand.”
Several facts support my argument. Although Tesla is a strong E2E player, domestic manufacturers like Xpeng and DeepRoute.ai didn’t just jump on the E2E bandwagon. With years of experience behind them in this field, their performance may not necessarily be inferior to Tesla’s FSD.
From a broader perspective, Tesla’s FSD rollout signifies a shift from electrification to intelligence. In the next five to ten years, intelligent driving will be the main driver of automotive upgrades in China and beyond.
Today, Chinese car companies and suppliers are on a more solid footing than when Tesla first entered China ten years ago. Some have become market incumbents, plunging huge sums into driverless vehicle technology. Therefore, it’s hard to say if they will lag even if Tesla’s FSD doesn’t enter China.
Consumer preferences and market demand provide another catalyst, as car buyers in China show a strong appetite for autonomous driving features.
On the other hand, Tesla itself faces significant hurdles. Musk has lobbied painstakingly to get FSD licensed in China. For FSD to thrive, however, two core issues must be resolved: obtaining geographic data on public roads and ensuring data security compliance.
The first one is easier to fix. Chinese laws require overseas carmakers to work with certified Chinese companies to acquire a mapping license for developing autonomous driving technologies.
Tesla recently solved this by partnering with Baidu to obtain a license and gain access to Baidu’s lane-level navigation system.
The data compliance challenge proves more thorny. In a positive development, an official document indicated early this year that all models produced by Tesla’s Shanghai Gigafactory comply with regulatory requirements in human face and cockpit data collection, storage, and processing.
This finding addressed data security concerns, allowing Tesla cars to enter previously restricted areas like government compounds.
Subsequent reports by international media, such as Reuters, suggest Tesla plans to establish a data center in China to process and train autonomous driving algorithms locally.
This represents a departure from Musk’s earlier stance. Previously, he insisted data collected in China be sent to the US to be processed by its Dojo supercomputer. This is normally how Tesla improves upon its E2E capabilities.
However, this practice is incompatible with China’s national data security law, which has strict requirements for cross-border data flow.
Bloomberg reported that Musk had implored Beijing on a whirlwind trip to China in late April for permission to transfer Tesla’s China-specific data to the stateside, to no avail. Hence, the alternative is to build a data center.
The question is how to obtain the crucial AI chips. Industry insiders told me that Nvidia’s GPU Orin X is a chip of choice for many domestic carmakers and suppliers working on intelligent driving.
One thing is for sure, though. Musk cannot expect the Biden administration to make an exception for Tesla China to import advanced chips from the US.
Which brings me back to what I said earlier. In the short term, Tesla’s FSD is unlikely to crush domestic opponents due to insufficient computing power.
From the regulator’s perspective, Beijing has welcomed Tesla’s FSD entry. Chinese policymakers are known for long-term thinking and should be credited with this.
Six years ago, despite the risk of jeopardizing the entire domestic automobile sector, they gave the green light to the Tesla Gigafactory in Shanghai. In hindsight, what was thought of then as a gamble proved to be a resounding victory.
The introduction of Tesla as a disruptor has forced China’s EV brands to work harder to play catch up, albeit on a different racetrack. Some of them have come to beat formidable rivals at their own game.
By allowing FSD into China this time, Beijing appears to be playing its card of “calculated cannibalism” again, using FSD to stimulate competition and drive excellence at home.
Manufacturing prowess and supply strengths still matter, but they alone will not ensure China’s lead in the global EV race. In the new era of autonomous driving, success hinges upon a closer integration of industries like integrated circuits, big data, and AI.
And who else is in a better place than Tesla to help Chinese car companies grow faster and elevate their game?
If things go as planned, Beijing’s latest maneuvers could play out the same way as is often touted in its “win-win” strategy: With Tesla reaping big gains from its FSD business in China, domestic automakers and suppliers are also poised to consolidate their advantages, further outpacing European and Japanese rivals in the autonomous driving race.
Stay up-to-date on engineering, tech, space, and science news with The Blueprint.
By clicking sign up, you confirm that you accept this site's Terms of Use and Privacy Policy
Ni Tao Ni Tao worked with state-owned Chinese media for over a decade before he decided to quit and venture down the rabbit hole of mass communication and part-time teaching. Toward the end of his stint as a journalist, he developed a keen interest in China's booming tech ecosystem. Since then, he has been an avid follower of news from sectors like robotics, AI, autonomous driving, intelligent hardware, and eVTOL. When he's not writing, you can expect him to be on his beloved Yanagisawa saxophones, trying to play some jazz riffs, often in vain and occasionally against the protests of an angry neighbor.
0