AVCJ&WeRide's Tony Han

Q: When will fully autonomous robotaxis achieve commercial rollout in China?

Tony: In the past, people often gave timelines, but these kept getting pushed back. The bottleneck used to be software reliability, but algorithms are quite advanced now. It's the hardware capabilities-the price of sensors, computer power, redundancies in vehicle platforms-that cannot keep up. To realize L4 autonomous driving, you need different parts of the industry to advance together. If this can happen quickly, 2025 is possible for commercialisation. If not, it will be postponed again. We have done a lot of work coordinating upstream and downstream partners to push the industry forward. It's like we wanted to be a software company, then we found half our time was spent on hardware, and we also spend a lot of time in car factories with first-tier suppliers.

Q: That means regulation is not a bottleneck...

Tony: No, it's not. In China and the US, regulators are encouraging autonomous driving.The problem is the reliability of software and hardware. Our robo-minibus started entirely unmanned operations in Guangzhou Bio Island in January. WeRide was the first in China to achieve unmanned daily operations. Robotaxis are taking longer because in the passenger car field we don't have industry leaders like [electric bus manufacturer] Yutong. They have designed a minibus for L4 autonomous driving and put it into mass production; no one has done the same for passenger cars in China.

Q: WeRide's recently announced collaboration with Bosch suggests a move into L2 [where vehicles change lanes, overtake, and park automatically; often defined as advanced driver assistance systems, ADAS] and L3 [conditional automation]. What's behind that?

Tony: The progress of the entire industry is very complicated. Fully unmanned autonomous driving is not easier than going to the moon. It is the highest peak in terms of achievement and, working towards that, you can incubate your technology in other scenarios. As a company, we must give confidence to the capital markets; we must be able to generate revenue, and perhaps even profit.

Q: Would you advocate Waymo's L4-from-day-one strategy over Tesla's preferred approach of moving from L2 to L4 through repeated iteration?

Tony: My view has been entirely consistent on this: moving from L2 to L3 to L4, you cannot bypass the man-machine co-driving stage. It's hard to define clearly when people should define when people should give way to algorithms. And by the time this has been established, L4-focused companies should have already succeeded. I think in just a few years, it will become apparent which is the fastest way to L4.

Q: Telsa has argued that its approach will work on a much larger scale than that of Waymo...

Tony: Waymo's approach can be scaled as well. This hasn't happened yet for economic reasons. On the contrary, the Tesla model of L3-L4 co-driving is fundamentally flawed. Without a 11:56 Preview solution, there are likely to be many accidents when the company does roll out L4 services at scale.For now, WeRide's priority is achieving scale with our L4 autonomous driving minibus.

Q: How does WeRide differentiate itself from other L4 robotaxi players?

Tony: Our road has been bumpy, and we have proved ourselves in the face of all kinds of setbacks. I believe our technology and operations are both industry-leading. We've been running a commercial robotaxi pilot programme in Guangzhou's Huangpu district for more than 1,000 days, having started a year earlier than our competitors. There are videos of our robotaxis operating in cramped urban villages during rush hour. We also have full-stack capability, with entirely unmanned robo-minibuses and robo-sanitary vehicles.

Q: What was the bumpiest part of the road?

Tony: In 2019, the industry endured a capital winter. It was hard to raise money, and we were eight months away from having to shut down. Our capital-raising specialist left the company. I came from a chief technology officer role and had never run a funding round from beginning to end. I was having six meetings a day with investors - if anyone showed interest in the business, I would meet them in person.

Q:You raised capital then and you did so again at the beginning of this year at a valuation of USD4.4bn. Is the road now smooth?

Tony: I don't think so. As a company, it's dangerous to think that having overcome setbacks, everything will now be fine.And in the current environment, no entrepreneur should feel safe. We should focus on how to survive and thrive despite uncertainty-and this means embracing uncertainty, not fighting against it or denying it.

Q: How are you embracing uncertainty?

Tony: Our collaboration with Yutong is a good example. We started out looking to work with passenger car manufacturers, but then Yutong proposed that we put our technology into its minibus -a commercial vehicle. And when we saw the minibus, it was everything we dreamed it would be in terms of suitability for L4 autonomous driving. We didn't take long in deciding to pursue this opportunity; the difficulty was getting the whole team to accept it. Working with Yutong did not cause us to deviate from our main objective to have pure unmanned vehicles in operation. We are one of few companies to have test licenses for pure unmanned vehicles in the US and China - and at the height of the pandemic, our minibuses delivered daily necessities to locked-down districts in Guangzhou. Embracing uncertainty helped us to stand out.Companies have long-term plans and short-term plans. You shouldn't be too rigid in short- term planning.

Q: Do you risk losing focus by having a broader set of product lines than others in the market?

Tony: All our vehicles are designed to run on intra-city roads. If your algorithm is rudimentary, then you have a different one for each product. WNe look to develop an algorithm like Linux.In essence, most operating systems used in EVs are Linux-a single set of algorithms suitable for multiple platforms. That's how we can support multiple product lines with limited manpower.

Q: The surge in hard-tech innovation has prompted more university professors to launch start-ups based on their research projects. As a professor-turned-entrepreneur, what do you think makes someone from an academic background suited to running a business?

Tony: It is often said that if Elon Musk was in China, he wouldn't raise any capital because he's arrogant, doesn't treat investors well, fails to meet agreed deadlines, and doesn't have strong controls over spending. Generally, investors prefer serial entrepreneurs.The three leading "new force" electric vehicle manufacturers in China-Nio, Xpeng, and Li Auto-were all founded by serial entrepreneurs. It's hard for professors to become serial entrepreneurs; typically, they don't achieve tenure until after the age of 30. However, it is natural for them to want to start businesses and push artificial intelligence and other technologies towards commercialisation. Hopefully, as this trend gathers momentum, we will see more students become founders and ultimately become serial entrepreneurs.

‍