BYD and Huawei have taken successive actions: traditional auto insurance can no longer cover the risks of new intelligent driving systems
In 2026, BYD and Huawei successively made decisions to take responsibility for intelligent driving.
On May 28th, BYD took the lead in announcing that during the user's compliant use of the urban navigation function, if an accident occurs for which the vehicle is responsible, the official will bear the direct economic losses such as vehicle repair, third - party losses, and personal injuries. Moreover, it is free, with no upper limit, and will not affect the next year's commercial insurance premium.
Less than a month later, Huawei followed suit.
On June 22nd, "Yinwang" under Huawei announced the addition of guarantee and service rights for the high - order function package of ADS, covering personal and property losses caused by accidental accidents during the use of Qiankun Intelligent Driving ADS assisted driving on legal roads and in legal areas.
On the surface, it seems that car companies are providing a technical endorsement for high - order intelligent driving. But in a broader industrial context, it touches on a more fundamental question: When the system starts to participate in driving, who should ultimately bear the responsibility for accidents?
The industry is breaking the deadlock, and the regulatory authorities are also taking action to promote the reform of new energy vehicle insurance and intelligent driving insurance.
Since this year, Shenzhen has introduced the "Ten Measures for New Energy Vehicle Insurance in Shenzhen", proposing to explore new models such as comprehensive intelligent driving insurance and "insurance + technology + risk control". Beijing has launched the development and application of commercial insurance for intelligent connected new energy vehicles, attempting to establish a new risk guarantee arrangement for the scenarios of "human - machine co - driving" and even future "machine - driving".
The signal is very clear: For autonomous driving to truly achieve large - scale implementation, it is no longer possible to wait to establish a risk mechanism that can identify responsibilities, share risks, and compensate for losses.
Autonomous driving requires a new insurance mechanism
Historically, every major productivity change has been accompanied by new institutional arrangements.
The steam engine changed the production organization mode, and the modern factory system was thus formed; the railway changed the transportation system, and countries began to establish railway regulatory rules; the Internet changed the way of information flow, and legal systems such as data protection, network security, and e - commerce were also improved accordingly.
Autonomous driving also follows this historical law.
From the perspective of social operation, what really determines whether autonomous driving can be implemented on a large scale is not just whether the technology is mature enough, but whether a matching institutional system can be formed behind this technology.
To put it more bluntly, it's about who is responsible for what, who will pay for accidents, and what to do if one can't afford to pay.
Autonomous driving is not just a simple upgrade of automotive technology, but a reconstruction of driving responsibilities. Traditional cars are driven by humans, while autonomous vehicles are mobile terminals operated by enterprises, controlled by systems, and scheduled by algorithms.
Once an accident occurs, whether the system makes a wrong judgment, whether the algorithm has defects, whether the operator has fulfilled its management obligations, and whether the vehicle manufacturer should bear product liability will all become new responsibility issues.
As the responsibility structure changes, the risk - sharing mechanism also changes accordingly.
Imagine that an enterprise operates a fleet of millions of Robotaxis. Even if the single - vehicle accident rate decreases, the overall risk exposure will increase as the scale expands. At this time, it is difficult for a single enterprise, a single user, or a single accident - responsible party to bear the losses alone.
This is exactly the premise for insurance to intervene. It transforms the risks that individuals or enterprises can hardly bear alone into a socialized mechanism that can be priced, shared, and compensated.
In the traditional automobile society, a closed - loop of traffic regulations, traffic police liability determination, insurance claims, and repair and rescue has been formed. In the future autonomous driving society, a new type of infrastructure covering liability determination, risk sharing, and loss compensation is also needed.
If the maturity of autonomous driving technology marks the rise of a "new productivity", then the subsequent laws and regulations are the reshaping of "new rights and responsibilities"; and insurance is the "new finance" pillar that undertakes these changes in rights and responsibilities.
Therefore, autonomous driving insurance cannot simply be an extension of traditional vehicle insurance. It needs to break through the old framework centered on driver responsibility and gradually evolve into a comprehensive risk management tool that integrates system responsibility, software responsibility, algorithm responsibility, network security risks, and operational risks.
So, what exactly should the insurance that matches this new productivity of autonomous driving look like? Has it emerged yet?
The existing "intelligent driving insurance" is only a transitional solution
Unfortunately, the current insurance logic is still designed around humans.
Under the current legal and regulatory framework, once an accident occurs, the first step for the traffic police is to determine liability, judging whether the driver has violated regulations or made operational mistakes. Then the insurance company will pay according to the liability ratio.
As autonomous driving evolves to L3 and above, the driving subject begins to shift from "humans" to "legal persons" such as automakers, technology companies, or operators. Under this transition, traditional vehicle insurance is no longer applicable either in product form or in scope of coverage.
The market has not stood still. There have been many preliminary explorations and practices in the industry around intelligent driving risks.
The "intelligent driving insurance" launched by XPeng Motors has cooperation partners including five insurance companies such as PICC Property and Casualty Company Limited, Ping An Insurance, China Pacific Insurance, Zhonghua Insurance, and Sunshine Insurance. Huawei's Hongmeng Smart Mobility has added intelligent driving worry - free service rights for all models of AITO, covering multiple scenarios such as intelligent parking, lane cruising, and navigation assistance. SERES has cooperated with Ping An Property Insurance to launch an intelligent driving liability insurance guarantee plan.
But strictly speaking, as of now, the "intelligent driving insurance" rights and interests launched by most car companies have not been filed with the regulatory authorities in the form of independent insurance products.
Mostly, after car owners purchase basic vehicle insurance, car companies provide an additional right, service commitment, or guarantee plan. Some are compensated by car companies themselves, while others are packaged into existing vehicle insurance or value - added services through cooperation with traditional insurance companies.
Overall, this type of insurance product is still in a transitional stage.
For car companies, consumers generally have concerns about the safety of intelligent driving. The existing "intelligent driving insurance" is like a reassurance pill. Endorsing with insurance is essentially a regular service to promote the popularization of intelligent driving.
Insurance companies actively participate in it. On the one hand, they want to obtain vehicle operation data. On the other hand, they are also preparing for the reconstruction of the future vehicle insurance actuarial model.
Insurance companies are in a dilemma when facing this new species of autonomous driving.
The most direct manifestation is the increase in claim payment pressure.
Smart cars are equipped with a large number of sensor devices. After an accident, the repair cost increases, the claim amount also increases accordingly, and the insurance premium will naturally rise. Car owners feel that "it is expensive to buy insurance and difficult to renew it"; insurance companies face rising claim costs and more difficult risk judgments.
At the same time, insurance companies are also stuck on the two mountains of data and liability determination.
In the scenario of "human - machine co - driving", the accident chain involves the driver, the car company, the software system, and the hardware equipment. After an accident, vehicle operation data, system logs, and black - box information are needed to restore and assign responsibility.
But most of these data are in the hands of car companies. It is difficult for traditional insurance companies to obtain them completely, timely, and at low cost, and they lack sufficient technical capabilities to analyze complex software and algorithm logic.
The awkwardness of the current so - called "intelligent driving insurance" is that it addresses users' concerns about intelligent driving risks, but has not really established a mature liability determination and risk pricing mechanism.
The core of intelligent driving insurance is not just the claim payment after an accident, but the risk identification before the accident, the data traceability during the accident, the repair control after the accident, and the long - term liability pricing.
Traditional insurance companies have licenses and funds, but they don't understand technology and don't have data.
It is precisely the car companies themselves that truly master the underlying code, have real - time traffic data, are most capable of identifying, pricing, and controlling this risk, and are willing to take responsibility for intelligent driving.
It is recommended that car companies enter the market to break the deadlock
BYD and Tesla have set examples.
Although both are involved in insurance, there are essential differences in the ways car companies play the game.
For the "intelligent driving insurance" - related services launched by some car companies in the market, car companies are the providers of rights and interests, cooperation partners, or referral channels.
BYD has a licensed property insurance entity, and Tesla also includes vehicle data in insurance pricing through Tesla Insurance in some states in the United States and is qualified to design and define vehicle insurance products.
This means that around intelligent driving, one can break out of the traditional vehicle insurance framework and connect vehicle data, driving behavior, software capabilities, the repair system, and insurance pricing - use data to re - price risks, use the repair system to control claim costs, and make the safety benefits brought by intelligent driving explicit.
The core of pricing lies in data.
Usage - Based Insurance (UBI) is a major innovation. It is insurance priced based on usage behavior. Through the Internet of Vehicles, intelligent terminals, or in - vehicle devices, data such as driving behavior, vehicle status, and driving environment are collected, and then the insurance premium is dynamically adjusted according to the real risk.
Tesla has abandoned the rigid pricing of traditional vehicle insurance, which is renewed annually and uses a one - size - fits - all approach, and created its own safety scoring model - Safety Score. Based on the real - time connected UBI model, data is read in real - time through in - vehicle hardware, and the car owner's insurance premium is dynamically adjusted with a 30 - day observation period.
In Tesla's latest Safety Score mechanism, as long as the car owner turns on FSD, the system defaults that this driving segment is a "perfect drive (100 points)". In some areas such as Illinois, for individual car owners who frequently use FSD during commuting, their monthly insurance premium has dropped from nearly $380 in the past to more than $50.
When car companies engage in insurance, it is also part of the full - life - cycle management of vehicles, forming a closed - loop of vehicle manufacturing, vehicle use, accident occurrence, damage assessment, repair, and then back to product design.
BYD's insurance layout reflects this path. BYD enters the vehicle insurance sector through its own insurance entity and integrates insurance into the automotive ecosystem. Insurance purchase, renewal, claim settlement, repair, and user service can all be connected through the official APP, intelligent cockpit, and authorized repair network.
On the one hand, it reduces intermediate links and lowers sales and service costs; on the other hand, it channels claim settlements into the official repair network, reducing disputes over damage expansion and assessment. More importantly, it forces the front - end R & D and manufacturing, helping car companies continuously optimize vehicle design and reduce repair costs from the source.
Another important aspect is the monetization of the safety of intelligent driving.
Intelligent driving reduces accidents, makes users' travel safer, and benefits society by reducing traffic losses. Insurance companies also benefit from the decrease in claim payments. As the technology provider, car companies find it difficult to directly obtain returns from the safety value they create.
Industry insiders point out that "intelligent driving insurance is the most realistic business model for autonomous driving, and the safety value can be directly monetized."
When car companies participate in insurance pricing and risk - taking, the decrease in the accident rate is no longer just a technical indicator, but is directly translated into reduced claim payments, optimized insurance premiums, and operating income.
Tesla's approach is to make FSD a variable that affects insurance premiums. The more users use the system, the better the risk performance, and the more likely the insurance price will decrease. BYD's approach is to reduce transaction costs, repair costs, and users' vehicle - using costs through direct insurance capabilities, official channels, and the repair system.
It can be predicted that in the era of smart cars, insurance will no longer be an additional financial product after selling cars, but will gradually become the core entry point connecting users, data, repair, and the after - market.
In this sense, whoever can form a closed - loop for insurance is more likely to take the initiative in the next - stage intelligent driving competition.
Conclusion
Of course, the threshold for insurance is not low. Actuarial science, risk control, solvency, and compliance supervision are all hard skills that car companies cannot bypass and cannot rush.
But there is more than one way - joining hands with insurance companies and leveraging each other's strengths can also lead to a symbiotic path.
In the future landscape, no one is an island. Car companies, insurance companies, and technology enterprises will eventually converge in technology and data to jointly build an anti - risk mechanism for the era of autonomous driving.
This article is from the WeChat official account "50 People in Vehicle - Road - Cloud", author: Hu Xiaofeng, published by 36Kr with authorization.