On February 10, Chinese automaker BYD launched 21 models equipped with advanced driver assistance systems (ADAS) in a single sweep, prompting swift reactions from competitors.
Though no names were explicitly called out, Great Wall Motor executives took to social media to declare that smart driving is “not a show.” Meanwhile, Richard Yu, chairman of Huawei’s Intelligent Automotive Solution business unit, made a pointed distinction between ADAS that merely functions and ADAS that is genuinely reliable and safe.
BYD’s strategy is straightforward: make smart driving mainstream. The company is offering high-level ADAS across its lineup at no additional cost, from the budget-friendly Seagull to the premium Song L electric vehicle. Its luxury brands—Denza, Fangchengbao, and Yangwang—are also integrating advanced driving features.
A preliminary breakdown suggests only three models—the Seagull, Qin, and Seal 05 DM-i—lack high-level ADAS in their base trims. In other words, nearly 90% of BYD’s newly launched vehicles now include smart driving features.
Given forecasts that BYD’s 2024 sales could reach 5.5 million units, the company might ship close to 5 million ADAS-equipped vehicles this year. “In the future, cars without smart driving features will become the minority,” BYD chairman Wang Chuanfu declared at the launch event.
But BYD isn’t cutting corners to achieve scale. Its entry-level DiPilot C system runs on capable hardware, such as Nvidia’s Orin N chip, allowing for future feature expansions.
At present, DiPilot C offers navigate-on-autopilot (NOA) for highways and expressways. However, its hardware is primed for future upgrades, including urban NOA and potentially nationwide ADAS—putting BYD in direct competition with the most advanced players in the industry.
Competitors scramble to keep up
BYD’s move has sent shockwaves through the industry.
Chang’an Automobile, bracing for the impact, plans to launch vehicles equipped with light detection and ranging (LiDAR) technology at a price point of around RMB 100,000 (USD 14,000) by August. Geely, meanwhile, is preparing an intelligence-focused product rollout aimed at making high-level ADAS accessible within the same price bracket.
Emerging EV makers aren’t sitting still either. Sources told 36Kr that Li Auto originally intended to release its new ADAS platform with only highway and commute NOA. But BYD’s aggressive push has forced the company to fast-track urban NOA integration. “Without this upgrade, we wouldn’t stand out against BYD,” one source said.
BYD’s year of smart driving
BYD’s ADAS transformation didn’t happen overnight.
Throughout 2023, the company’s smart driving division underwent multiple internal restructurings. At the start of 2024, BYD consolidated its ADAS and intelligent connectivity teams under its new technology institute. By midyear, the division split into two: Tianxuan, focused on in-house ADAS development, and Tianlang, handling lower-tier ADAS and supplier partnerships.
Then, in October 2024, the two teams were merged again. Leadership changed hands three times—passing from veteran Han Bing to rising star Xu Lingyun, before landing with current head Li Feng, who reports to Yang Dongsheng, the institute’s director.
Industry sources told 36Kr that Li stood out due to his extensive experience developing ADAS within BYD’s electronics division. His expertise in mass production and both highway and urban NOA made him a natural choice.
Once the team stabilized, BYD ramped up its in-house ADAS development. The project, launched in September 2024, covers highway NOA, urban NOA, and more.
Last December, BYD’s Shanghai-based ADAS employees were abruptly instructed to relocate to Shenzhen for an intensive development sprint. “BYD even increased travel budgets so employees wouldn’t have to take the cheapest red-eye flights,” one employee said.
Li himself pushed the team hard. Early in December 2024, he issued a directive signaling that development was in its final stretch, setting February 5 as the deadline for software readiness. Employees noted that Li was often still working with engineers at 5 or 6 a.m, driving bug fixes.
BYD’s ADAS push is top-down. Chairman Wang sees smart driving as an essential feature with massive potential. At the same time, growing consumer demand and feedback from BYD’s sales teams underscored the urgency of a faster rollout.
The company’s vision is now clear: smart driving for the masses.
A tiered approach to ADAS
BYD’s autonomous driving framework consists of three platforms, each tailored to different market segments.
- A platform: Powered by dual Nvidia Orin X chips (508 TOPS compute), featuring three LiDAR sensors for high-precision perception. This system is reserved for Yangwang-branded models.
- B platform: Equipped with a single Orin X chip (254 TOPS), with one or two LiDAR sensors. This system will be used in Denza and select premium BYD models.
- C platform: A cost-efficient system that omits LiDAR but relies on an optimized sensor suite for NOA and parking. While it currently lacks full urban autonomy, BYD plans to introduce commute NOA—enabling vehicles to remember and replicate driven routes—by the end of the year.
The C platform represents BYD’s largest ADAS deployment and is now the core focus of its in-house development efforts.
Even in end-to-end (E2E) AI-driven ADAS, BYD is covering its bases. The company reportedly has two separate teams working on this technology.
At the launch event, Wang revealed that BYD’s fleet now generates 72 million kilometers of new training data daily. Every seven days, the company can train a new over-the-air (OTA) update and roll it out to vehicles, continuously refining DiPilot C’s capabilities.
RELATED ARTICLE
Can BYD make smart driving mainstream?
In 2023, the industry’s ADAS focus was on E2E AI and large models, led by players like Huawei, Li Auto, Xpeng Motors, Momenta, and DeepRoute.ai.
BYD is shifting the narrative through sheer scale.
The company’s 21 newly launched ADAS-equipped models have largely maintained their price points. Even the relatively affordable Seagull now comes with highway NOA—arguably a bolder move than Xpeng’s Mona M03.
This is possible due to BYD’s manufacturing dominance. According to 36Kr, the entire hardware and software cost for DiPilot C is under RMB 5,000 (USD 700), significantly lower than third-party suppliers, which charge RMB 6,000–7,000 (USD 840–980).
With potential sales exceeding 5 million units, BYD wields unmatched bargaining power in AI chips, LiDAR, and other core components.
BYD’s smart driving push may appear sudden, but it has been years in the making—backed by clear strategy, relentless execution, and an industrial edge that few can match. The playbook is open, yet the industry still seems caught off guard.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This article was written by Li Anqi for 36Kr.