Soaring 30%! What did BYD Have up its Sleeve on "Intelligent Driving for All" ?

On February 10, 2025, BYD, as it did with the "DM 5.0" technology released last May, once again sparked a new wave of "Intelligent Driving for All". Focusing on "electricity being cheaper than fuel" (significantly improving hybrid range and greatly reducing fuel consumption), BYD once again sparked a new wave of "Intelligent Driving for All". For specific details, please take a look at the minutes of the release event: BYD (Minutes): The Era of Intelligent Driving for All Has Arrived.

The day after the release event, BYD Company Limited (1211.HK) fell 0.7%, seemingly indicating that funds had somewhat priced in this round of intelligent driving expectations in advance. However, as of February 19, BYD had risen 10.7% compared to before the intelligent driving release event.

Coincidentally, Dolphin Research follows up on this BYD "Intelligent Driving for All" centered release event to examine BYD's intelligent driving layout and attempt to understand the following core questions:

1. How valuable is BYD's "Intelligent Driving for All"?
2. Did the solutions presented at this BYD release event exceed market expectations?
3. What is the significance of BYD's "Intelligent Driving for All" strategy?

Specifically:

1. How valuable is BYD's "Intelligent Driving for All"?

To analyze the value, let's first look at BYD's intelligent driving solutions and hardware/software configurations:

① What configurations and models are equipped with BYD's intelligent driving?

From the information of this release event, BYD's "God‘s Eye" is divided into three versions: A/B/C.

1) High-end: God‘s Eye A is the high-level intelligent driving three-laser version (DiPiIot600), mainly equipped on the flagship models of the Yangwang brand;

2) Mid-high-end: God‘s Eye B is the high-level intelligent driving laser version (DiPiIot300), mainly equipped on some models of the Denza brand and high-end models of the BYD brand;

3) Mid-low-end: God‘s Eye C is the high-level intelligent driving three-eye version (DiPilot100), mainly equipped on the remaining models of the BYD Dynasty and Ocean series.

In this release event, BYD focused on introducing and updating the intelligent driving solution of the Eye of God C version, with a greater focus on creating "Intelligent Driving for All" in 2025, still concentrating on the sales base below 200,000 yuan.

② Looking at the hardware and software configurations of BYD's three intelligent driving versions (combined with third-party research information):

1) High-end and mid-high-end versions - support for urban NOA functionality

① God's Eyes A (DiPilot 600):

a. Hardware:

Sensors: Uses 3 LiDARs (total cost estimated at 5000 yuan), 12 cameras (total cost estimated at 1200 yuan), 12 ultrasonic radars (total cost estimated at 300 yuan), and 5 millimeter-wave radar solutions (cost estimated at 500 yuan).

Intelligent driving domain control: God's Eyes A's intelligent driving chip mainly adopts NVIDIA's dual Orin X solution (computing power at 500-600 Tops), with the total cost of intelligent driving domain control estimated at 8000-12000 yuan.

Total hardware cost: The total cost of intelligent driving hardware is expected to be close to 20,000 yuan.

b. Software:

It is expected to use algorithms provided by Momenta or Huawei, rather than self-developed solutions.

Intelligent Driving Features:

Supports high-speed NOA + urban NOA + e³ parking, the no-map version of intelligent driving that can be used nationwide will be launched in December 2024.

Equipped Models: Million-level models

The flagship model of the Yangwang brand.

② God's Eyes B (Dipilot 300)

Hardware:

Sensors: Uses 1-2 LiDARs, with two configurations (cost estimated at 2000-4000 yuan), 12 cameras (total cost estimated at 1200 yuan), 12 ultrasonic radars (total cost estimated at 300 yuan), and 5 millimeter-wave radar solutions (cost estimated at 500 yuan).

Intelligent driving domain control: God‘s Eye B mainly adopts NVIDIA's single Orin X solution (computing power at 200-300 Tops), with the total cost of intelligent driving domain control estimated at 5000-7000 yuan.

Total hardware cost: The total cost of intelligent driving hardware is expected to be close to 10,000-13,000 yuan.

In terms of intelligent driving software:

It is also expected to use algorithms provided by Momenta rather than self-developed solutions.

Intelligent Driving Features:

Supports high-speed NOA + lightweight urban NOA + e4 parking, the no-map version of intelligent driving that can be used nationwide will be launched in December 2024, but there is still a gap compared to the first tier of intelligent driving solutions from Momenta.

Equipped Models: Expected to be equipped on models priced at 200,000 yuan and above

Will be equipped on some models of the Denza brand and high-end models of the BYD brand, according to the information from the press conference, the first batch of models to be launched includes Seal/Sea Lion 07 EV high configuration models, Song L EV, Han EV intelligent driving version mid-to-high configuration models equipped with God‘s Eye B, with the minimum model price expected to be around 200,000 yuan.

2) Mid-to-low-end version: Supports high-speed NOA + automatic parking function

③ God's Eyes C (Dipilot 100) - The strategic core in 2025 still focuses on "Intelligent Driving for All"

Hardware:

Sensors: Uses a pure vision solution, does not equip LiDAR, with 12 cameras (total cost estimated at 1200 yuan), 12 ultrasonic radars (total cost estimated at 300 yuan), and 5 millimeter-wave radar solutions (cost estimated at 500 yuan) On the intelligent driving domain controller: The God‘s Eye C mainly uses NVIDIA ORIN (100 TOPS) chips or Horizon J6m chips, with the total cost of the intelligent driving domain controller expected to be between 3,000 to 4,000 yuan.

Total hardware cost: The total cost of intelligent driving hardware is expected to be around 6,000 to 7,000 yuan, which is an estimated net increase of about 4,000 to 5,000 yuan compared to the previous Dipilot 10/30 solution.

On intelligent driving software:

It is expected to adopt a self-developed solution, possibly initially using rule-based algorithms for rapid implementation.

On intelligent driving functions:

Supports high-speed NOA (with a takeover frequency of over 1,000 km) + automatic parking function (with a success rate of up to 99%) + city memory navigation to be pushed by the end of the year, bringing the starting price of high-speed NOA to the 70,000 yuan level.

Models equipped: Expected to be equipped across the entire range of models priced around 70,000 to 200,000 yuan, with added features at no extra cost.

The minimum model price supporting God‘s Eye C is the Seagull (intelligent driving version starting at 78,800 yuan), marking the first time in the industry that high-speed NOA is available in the 70,000 yuan market. It is expected that God‘s Eye C will be equipped across the entire range of models priced within 70,000 to 200,000 yuan.

Summary: The strategic focus for 2025 is not on breakthroughs in high-end models but rather on "intelligent driving penetration."

1) Mid-to-high-end models:

From the perspective of BYD's mid-to-high-end model intelligent driving configurations, it is expected that models capable of achieving urban road NOA will still be priced above 200,000 yuan, and the intelligent driving software solution providers will still mainly rely on third-party Momenta solutions, rather than self-developed solutions, indicating a gap in intelligent driving performance compared to the first-tier players.

In comparison, XPeng has already made urban NOA available in the 150,000 yuan level (Mona M03 high-end intelligent driving version, XPeng P7+ standard), and BYD's mid- to high-end models do not have a significant competitive advantage. It is expected that BYD's strategic core this year is still not focused on high-end breakthroughs but rather on achieving the basic intelligent driving functions for sales volume.

① Intelligent driving technology performance: BYD's mid-to-high-end intelligent driving technology is not self-developed, and the Momenta solution does not perform as well as the domestic first-tier players like Huawei, XPeng, and Li Auto;

② Pricing of urban NOA models: BYD's expected equipped models will still be above 200,000 yuan, lacking the cost-performance advantage compared to XPeng.

The next-generation intelligent driving chip from NVIDIA, the NVIDIA Thor, is expected to be released in 2025 (with a computing power of 1,000 Tops+). Currently, car manufacturers planning to equip this chip include ZEEKR, Li Auto, etc. (expected to be on vehicles by Q3 2025), which will accelerate the R&D progress of high-end intelligent driving in mid- to high-end models (higher computing power on the vehicle side, better intelligent driving performance). However, currently, BYD has no such plans for mid- to high-end models in 2025.

2) Mid-to-low-end models: "Intelligent driving penetration" will become the largest increment in 2025. At the same time, from the marginal changes in the price range of models equipped with BYD's intelligent driving in three configurations compared to 2024, the price range for mid-to-high-end intelligent driving versions (Dipilot 300/Dipilot 600) remains unchanged compared to 2024; however, the most significant drop in the intelligent driving price range is seen in the mid-to-low-end intelligent driving version (Dipilot 100), where the pricing range for equipped models drops from 200,000-300,000 yuan in 2024 to the 70,000-level to 200,000 yuan in 2025. This is also the basic market for BYD's sales, and the "intelligent driving drop" of low-priced models will become BYD's largest increment in 2025.

2. Outcome of this press conference

Did the plan exceed market expectations?

From the information presented at the press conference, there were no surprises in the progress of intelligent driving technology, while the key to exceeding expectations lies in pricing:

① The starting price for models equipped with high-speed NOA has dropped to the 70,000 yuan level: BYD's lowest-priced model, the Seagull (equipped with the God‘s Eye C intelligent driving version starting at 78,800 yuan), and the Qin Plus will also feature an intelligent driving version, with a starting price lower than the market expectation of 90,000-100,000 yuan.

② Models upgrading intelligent driving functions achieve additional features without price increases: The pricing for upgraded intelligent driving models remains largely unchanged from the previous 2025 models. Dolphin Research expects BYD to mainly achieve this through a) scale advantages, b) self-developed components for the Dipilot 100 scheme, and c) reducing supplier costs, among other methods.

③ Wide coverage of models, standard across the entire series: All models are equipped with advanced intelligent driving as standard. Dolphin Research expects that, based on the production ramp-up pace, the proportion of models equipped with intelligent driving versions is expected to reach over 50% of total annual model sales, with BYD's annual delivery volume of intelligent driving models expected to be at least 2-3 million.

Overall, in Dolphin Research's view, the value of BYD's intelligent driving lies primarily in cost reduction rather than breakthroughs in technology-driven "Intelligent Driving for All".

Therefore, to evaluate the value of BYD's intelligent driving press conference, the focus is not on breakthroughs in advanced intelligent driving technology but rather on equipping mid- to low-end models with basic intelligent driving functions—high-speed NOA and automatic parking, which are also the foundation of BYD's sales.

BYD achieves this through ① scale advantages; ② self-developed components for the Dipilot 100 scheme; ③ reducing supplier costs (expected to decrease by 15-20% this year), bringing high-speed NOA down to the 70,000 yuan price range for the first time, driving down costs to enable basic intelligent driving functions to penetrate the market, unlike XPeng, Li Auto, Huawei, and other new car manufacturers that pursue "Intelligent Driving for All" driven by technology.

While discussing Intelligent Driving for All, Dolphin Research emphasizes that cost-driven intelligent driving penetration (like BYD) — equalizing the cost of intelligent driving- and technology-driven intelligent driving upgrades (like XPeng) — leading intelligent driving technology- have very different commercial logic and strategies. Please continue to follow Dolphin Research's subsequent analysis.

3. What is the significance of BYD's "Intelligent Driving for All" strategy?

① 2025: Achieve an offensive strategy for sales volume through "Intelligent Driving for All", realizing both defensive and expansion of market share:

In a previous in-depth article titled BYD: The Final Battle!, Dolphin Research mentioned that BYD formed a leading advantage through the generational leap of DM 5.0 hybrid technology, achieving an expansion of the hybrid market base in its 2024 strategic approach.

However, competitors like Geely began to release similar technologies in the second half of 2024, and by the end of 2024, they started large-scale production (Galaxy Starship 7 EMI), successfully creating popular models. BYD's first-mover advantage in hybrid technology has been largely caught up by competitors, and BYD's sales in December 2024 and January 2025 fell short of market expectations, showing a downward trend in market share. BYD urgently needs significant upgrades to new models to defend and further expand its market share.

Therefore, at the beginning of 2025, during this smart driving launch conference, BYD proposed the "Intelligent Driving for All" strategy, directly introducing high-speed NOA to the 70,000 RMB price range (smart driving penetration), combined with the fifth-generation DM technology (range of 2100km+ and fuel consumption of 2.9L when out of battery), forming a dual advantage of "smart driving + ultra-low usage cost" in the 70,000 to 150,000 RMB vehicle segment.

On one hand, this will significantly accelerate the penetration of fuel vehicles in this price range (currently, joint venture fuel vehicles in this price range have no energy consumption advantage, relatively higher usage costs, and lack smart driving features), hastening the elimination of fuel vehicles. On the other hand, it will launch an offensive against new energy vehicle companies lagging in smart driving within this price range, further achieving market share expansion.

Currently, in the 70,000 to 150,000 RMB sales base, there are no new energy models at the same price point equipped with smart driving or only equipped with low-level smart driving (capable of only the most basic functions like ACC, LCC, etc.). Competitors, due to low gross margins and insufficient economies of scale, either find it difficult to catch up in the short term or incur significant losses while trying. The essence of this smart driving penetration strategy is that while the smart driving capability is insufficient, the core principle is to increase quantity without increasing price, reflecting BYD's unique cost control and procurement bargaining power under its scale salesThrough high-speed NOA+ automatic parking, BYD has created a phase-leading time difference with the concept of "Intelligent Driving for All" (at least nearly six months ahead).

With the support of this strategy in 2025, the overall sales and performance certainty is strong. The previously set sales target of 5 million is not difficult to achieve under this offensive approach, and it is likely to generate better-than-expected performance.

② The trend of software algorithms gradually shifting from external supply to self-research:

Although BYD has the largest stock of new energy vehicles in China, there are not many models equipped with advanced smart driving hardware (models equipped with Dpilot 100 and smart driving hardware account for no more than 5%). Only vehicles with advanced smart driving hardware can complete a series of closed-loop processes such as high-quality driving data collection, storage, transmission, and cloud uploading. Therefore, the effective data volume of BYD itself is still relatively small and does not have an advantage among new energy vehicle companies.

From the current smart driving solutions of car companies, the end-to-end route is still hopeful to truly achieve L4- L5 high-level autonomous driving compared to the rule-based route. The core of the end-to-end approach is essentially the competition of training computing power + high-quality driving data + smart driving technology talent.

Dolphin Research believes that BYD's move to standardize high-level smart driving hardware across its entire lineup is essentially aimed at collecting high-quality data. At the same time, utilizing Momenta's white-box delivery smart driving solution accelerates the iteration of self-researched end-to-end large models, catching up on urban NOA self-research and accelerating the trend of software algorithms for mid- to high-end models shifting from external supply to self-research.

③ Medium to long-term self-research of smart driving software is the core competitiveness:

BYD's "Intelligent Driving for All" is still achieved through cost reduction driven by scale effects rather than technological breakthroughs, which does not constitute a fundamental barrier. Competitors in the 100,000-200,000 yuan new energy vehicle segment will continue to follow suit, just like the hybrid DM 5.0 technology. The first-mover advantage in 2024 will be caught up by new energy competitors in 2025 (such as Geely's Thor hybrid). The key to BYD's moat in the medium to long term still lies in the car company's ability to self-research smart driving software, and BYD still needs to accelerate the self-research of urban NOA smart driving software.

In the medium to long term, BYD may still face:

a) Leading smart driving companies like XPeng, leveraging their technological lead and model downscaling (especially with BYD's smart driving driving a significant upgrade in the smart driving supply chain, which may further lower the cost of smart driving hardware) will continue to downscale urban NOA to the 100,000-150,000 yuan price range;

b) The smart driving catch-up from new energy vehicle brands under independent car manufacturers such as Geely, Changan, Chery, and Great Wall in the 100,000-200,000 yuan segment. After BYD loses its temporary technological lead, it will still need to adopt "price for volume" strategies again.

Therefore, the market currently has relatively no disagreement on BYD's performance certainty in 2025, while the biggest disagreement is the uncertainty BYD will face in 2026 (especially after the smart driving downscaling is caught up by competitors), which still needs to rely on breakthroughs in intelligent self-research (especially urban NOA algorithms) to maintain the sales base and assist in high-end breakthroughs, as well as achieve smooth progress in overseas expansionThis article concludes the interpretation of BYD's purpose for this "Intelligent Driving for All". In the next article, Dolphin Research will focus on analyzing the potential valuation reshaping opportunities brought by this wave of "Smart Driving Penetration" for BYD and the industry. Stay tuned