China's EV Market Has Evolved Into an AI Arms Race, Not Just About Low Prices
Chinese electric vehicle manufacturers are increasingly incorporating the same AI features to survive the ongoing price war in the world's largest automotive market, China. The competition has shifted from battery range and driver-assistance systems to a focus on a range of in-car AI functions.

ByteDance's cloud platform, Volcano Engine, announced at the Beijing Auto Show last Friday that over 50 automotive brands are now using its Doubao AI model. The tech division had a booth at the show next to autonomous driving taxi company Pony.ai.
Volcano Engine stated that this means Doubao is applied to 145 models, covering over 7 million vehicles. Beyond domestic brands, Doubao AI has also been integrated into new models from foreign brands, such as the all-electric Mercedes-Benz GLC, SAIC Audi E7X, and SAIC Volkswagen ID. ERA 9X.
“We will continue to accelerate the speed of new feature integration,” Fermin Soneira, CEO of the Audi-SAIC joint venture, told reporters earlier this month ahead of the auto show. He pointed out that automakers can remotely and quickly deploy technology updates via OTA.
Despite the rapid rollout of new features, automakers still face ongoing sales pressure.
He said, “The situation will remain difficult because there is capacity. This price war won’t really end next month.”
The shift towards AI reflects consumer demand for connected features, including interfaces compatible with Huawei smartphones or voice assistants like Doubao.
According to consulting firm Chozan, ByteDance’s Doubao is currently the most widely used AI chatbot in China, with over 155 million weekly active users as of the beginning of this year. Volcano Engine showcased Chinese and English AI systems for automotive use at its auto show booth.
Stephen Dyer, Partner and Managing Director, Automotive & Industrial Practice, AlixPartners Asia, said the price war has evolved into a battle over functionality within the cockpit.
However, the challenge is that many of these technologies quickly become similar, making it harder for companies to stand out.
According to AlixPartners data, among the 20 best-selling electric vehicle models in China, those priced at 100,000 yuan (US$14,645) and above offer similar driver-assistance and in-car entertainment features.
“With technology, they have to race to develop it, and they have to constantly compete because technology spreads so quickly, you can never maintain a long-term technological advantage,” Dyer said.
Instead, he expects Chinese companies to begin competing more on “outside-the-car experiences,” similar to luxury brands offering exclusive lifestyle experiences.
Chinese automaker Nio, for example, offers customers exclusive products and clubhouses in addition to vehicles with premium interior materials.
The Chinese EV company has faced the dual challenges of cost pressures from offering such perks and slowing market growth. However, Nio announced last week that its ES8 is the first model priced above 400,000 yuan to deliver 100,000 units in just 215 days.
Alibaba also announced on Friday that its Qwen AI model will be integrated into vehicles from several automakers, including BYD, as well as a local joint venture with Volkswagen. The system allows drivers to order meals, book hotels, purchase attraction tickets, and track packages using voice commands.
The model will use Nvidia’s automotive chip system and will function even with limited network connectivity.
Tu Le, founder and managing director of consulting firm Sino Auto Insights, told CNBC’s Eunice Yoon that, ultimately, AI should run in the background to support the user experience, and doesn’t necessarily need to be a feature of the vehicle.
Even if automakers struggle to stand out in the Chinese market, they may be able to compete more effectively with foreign counterparts.
“What we see as a simple feature and standard configuration in Volkswagen in the Chinese market will soon become standard in Western markets,” Le said. (Bianews, Electric Vehicle World, AI Pris compiled)