"BYD Intelligent Driving Strategy: Integrating In-house Research and Development for Mass Production, Aiming to Bring NOA to Vehicles Priced Under 100,000 RMB

10/22 2024 353

Author | Dexin

Editor | Dexin

According to multiple sources, BYD intends to integrate its in-house intelligent driving team under its New Technology Institute after a series of adjustments this year, with the goal of achieving mass production of in-house algorithms as early as November.

This could mark a significant milestone in BYD's intelligent driving research and development.

During a meeting in September, it was internally proposed to introduce advanced intelligent driving (high-speed NOA-level) capabilities to models priced between 100,000 and 200,000 RMB by next year.

As the top priority for BYD's senior management, the competition within BYD's intelligent driving division and with external suppliers is intense.

I. The Three Pillars of BYD's Intelligent Driving Research and Development

BYD's intelligent driving research and development falls under its Product Planning and Automotive New Technology Institute, commonly referred to as the "Planning Institute."

Earlier this year, BYD reorganized the Planning Institute into the "BYD Automotive New Technology Institute" and merged its various intelligent driving research and development teams into this new organization. The two main forces are the Intelligent Driving Development Center, based on the Electronic Integration Department, and the Intelligent Driving Development Team of BYD's Fifth Business Unit.

January 2024 marked a crucial turning point.

During the Vision Day event, China's largest new energy vehicle manufacturer clearly outlined its intelligent strategy of "Vehicle Intelligence" and accelerated the timeline for integrating intelligent driving capabilities into its vehicles.

Under the umbrella of Vehicle Intelligence, BYD boasts

its own platform-level architecture for vehicle intelligence, the Xuanji Architecture;

a unified brand for intelligent driving computing platforms using "DiPilot + Digital Series";

and a mass-produced intelligent driving system, "The Eye of Heaven."

At that time, The Eye of Heaven had already been mass-produced and integrated into BYD's Auto and Denza brands.

The intelligent driving systems in the Auto U8 and Denza N7 primarily stem from BYD's collaboration with Momenta based on the Dipai Zhixing framework, while BYD's in-house intelligent driving algorithms are also under development.

In mid-2024, to accelerate the development of in-house algorithms, BYD established the "Tianxuan" team within the New Technology Institute.

This team joined the previously merged Intelligent Driving Development Team of the Fifth Business Unit, named "Tianlang."

Together, the New Technology Institute now boasts three teams dedicated to the development of intelligent driving.

The Intelligent Driving Development Center is led by Han Bing, Deputy Dean of the New Technology Institute and Director of the Intelligent Driving Center; Tianxuan is headed by Xu Lingyun, former Director of Intelligent Driving at HiPhi; and Tianlang is led by Li Feng, former Director of the Software Center at BYD's Fifth Business Unit.

All three teams report to Yang Dongsheng, Senior Vice President of BYD and Dean of the Automotive New Technology Institute.

II. Unified Hardware and the Economies of Scale from 4 Million Annual Sales

When BYD decided to begin mass-producing advanced intelligent driving systems in 2023, it opted for standardized hardware specifications, such as camera configurations.

This decision offers several benefits: facilitating unified data collection, reducing procurement costs, and enabling easy switching between suppliers.

Initially, this strategy impacted some suppliers' entry into BYD's supply chain.

For example, DJI Auto (now Zhuoyu Technology) had unique hardware featuring binocular vision as the primary front-facing sensor. However, as BYD's in-house developed CloudRide technology began to be widely integrated, the binocular system, with its ability to preview road conditions, regained prominence in BYD's premium vehicles.

BYD further unified its intelligent driving computing hardware during Vision Day.

The intelligent driving computing platforms are now collectively referred to as DiPilot xxx, without specifying the specific chip brand and model:

NVIDIA Drive OrinN and Horizon Journey 5, with up to 100T of computing power, are collectively known as DiPilot 100;

Single OrinX, with up to 300T of computing power, is designated as DiPilot 300;

Dual OrinX, with 508T of computing power, is designated as DiPilot 600.

Since the mass production of The Eye of Heaven intelligent driving system in Auto and Denza, BYD has developed and manufactured its intelligent driving domain controllers, including their aluminum housings.

"With BYD's procurement volume, even the cost of aluminum procurement is cheaper than that of Tier 1 suppliers shipping tens or hundreds of thousands of units annually," commented a supplier.

DiPilot also delineates the focus areas for the three intelligent driving teams:

The Tianxuan team initially focuses on the development of high-computing platforms, namely DiPilot 300 and above;

Tianlang focuses on intelligent driving research and development for low-to-medium computing platforms;

The Intelligent Driving Development Center (IDC), responsible for mass production and delivery, acts as an internal supplier to the first two teams while also considering external supplier solutions.

IDC, which integrates the former Electronic Integration Department, boasts a larger team and is also responsible for the development of domain controllers, including the next-generation NVIDIA Drive Thor domain controller.

Based on this structure, BYD is concurrently developing intelligent driving capabilities on NVIDIA, Horizon, and Black Sesame platforms. Among them, NVIDIA's platform, with its mature Drive Orin, currently receives the most attention from BYD's internal teams.

HiEV has learned that Tianlang's in-house developed algorithm based on Orin N could potentially enter mass production as early as November, marking BYD's first breakthrough in mass-producing in-house intelligent driving algorithms.

III. Bringing Advanced Intelligent Driving to Vehicles Priced Under 100,000 RMB

In 2023, BYD introduced advanced intelligent driving capabilities through its Auto and Denza brands.

To date, advanced intelligent driving versions equipped with The Eye of Heaven have been launched for the Auto U8, Denza N7, Han EV, Song L EV, Sealion 07 EV, and Dolphin.

Additional models such as the Denza N9 and Tang L are expected to follow soon.

Essentially, most of BYD's best-selling models priced above 200,000 RMB now offer advanced intelligent driving capabilities. Internally, BYD aims to extend this capability to models priced under 100,000 RMB by 2025.

In less than two years, BYD has virtually built its advanced intelligent driving capabilities from scratch and integrated them into its best-selling models, thanks largely to the maturity of the intelligent driving supply chain. The intense competition in the intelligent driving market, driven by the capital influx in previous years, has also contributed to this achievement.

Furthermore, BYD's intelligent driving team has benefited from the exit of some automakers from the fiercely competitive market, allowing BYD to absorb experienced teams with mass-production capabilities.

Xu Lingyun, head of Tianxuan, has worked for XPeng Motors, DJI Auto, and HiPhi, while part of SAIC Motor's in-house R&D team joined the Fifth Business Unit after the integration of the R Auto brand. R Auto's intelligent driving system, mass-produced on the R7, features advanced capabilities based on the OrinX platform.

In addition to in-house algorithms, BYD is also planning to develop its own intelligent driving chips, starting with low-to-medium-end products.

A source close to BYD's supply chain commented, "Don't underestimate BYD's speed and strategic resilience."

Solemnly declare: the copyright of this article belongs to the original author. The reprinted article is only for the purpose of spreading more information. If the author's information is marked incorrectly, please contact us immediately to modify or delete it. Thank you.