Conversation with PENG Jun from Pony.ai: An "Anti-consensus" Autonomous Driving Company

In 2016, when Peng Jun founded Pony.ai, he had anticipated that it would take at least 10 years of effort for Robotaxis (autonomous taxis) to move from a distant vision to a reality and achieve large - scale implementation. This required waiting for the maturity of technology and laws and regulations, as well as considering social acceptance.

Today, whether it's Waymo shuttling through the streets of San Francisco or Pony.ai's fleet taking orders in the bustling downtown area of Nanshan, Shenzhen, it has been proven. Autonomous vehicles have already integrated into the social traffic flow and started commercial operations.

However, what Peng Jun might not have expected 10 years ago was that a key issue restricting the rapid expansion of his fleet was a series of trivial maintenance and operation tasks after the absence of drivers.

When humans drive, tasks such as charging, car washing, vehicle maintenance, and even helping customers carry luggage are all done by the drivers. But when autonomous vehicles hit the road, these seemingly insignificant tasks have become problems.

Peng Jun, the CEO of Pony.ai, was interviewed by 36Kr Auto recently. They had a discussion on topics such as the implementation and operation of Robotaxis, the choice of L2 mass - production business, why there was no rush to layout embodied intelligence, and the selection of technical routes.

Peng Jun told 36Kr that Pony.ai has established an operation team and explored a set of operation and maintenance standards to provide logistical support for autonomous vehicles. "These human resources include both remote safety personnel and ground support and logistics staff."

In the future, even if the fleet size expands rapidly, the vehicle - to - personnel ratio will not increase significantly (i.e., breaking the traditional growth logic of more vehicles requiring more people). From this perspective, Peng Jun believes that this is a blind spot in capabilities that today's ride - hailing companies or automakers may overlook when entering the Robotaxi field.

After a 10 - year long - distance run, combined with the rapid maturation of the supply chain due to automotive intelligence, unmanned vehicle technology has entered the stage of substantial commercial operation in various fields. A representative example is unmanned logistics vehicles. Companies such as White Rhino, Neolix, and Jiushi Intelligence have successively received capital investment and have become the business cards of China's automated logistics.

As one of the markets with the greatest potential for unmanned vehicles, Robotaxis have attracted automakers, ride - hailing platforms, and other companies to take turns to enter the field and make layouts. Tesla, XPeng Motors, Geely, etc. have all announced clear unmanned vehicle operation plans.

A mainstream voice in the industry believes that automakers have a complete vehicle engineering ability and a technical system from assisted driving to autonomous driving that can achieve a data snowball effect. They will be the core players in the Robotaxi field. However, Peng Jun told 36Kr that the manufacturing capacity in China has exceeded demand. For Robotaxis, if a company has never done it before, it is almost starting from scratch.

Peng Jun believes that for Robotaxis, technical ability determines whether a company can start (from 0 to 1), that is, whether it can do this thing, while operation ability determines efficiency. Currently, Pony.ai has explored a gradually clear business model.

Along the way, the development of Robotaxis has not been all smooth. It has long faced commercialization difficulties. Many peer autonomous driving companies have shifted their focus to the L2 intelligent driving supplier business to find more revenue - generating channels.

Pony.ai is one of the few companies that has consistently adhered to the Robotaxi front. Currently, Pony.ai's seventh - generation vehicles have achieved single - vehicle profitability in Guangzhou and Shenzhen. After the financial model is verified, Pony.ai plans to rapidly increase the fleet size to 3,500 vehicles this year.

Looking back on the choices made along the way, Pony.ai has mostly made "anti - consensus judgments".

Peng Jun said that Pony.ai once briefly tried the L2 assisted driving mass - production business, but soon found that it was destined to be a low - profit industry. "Due to the low technical threshold and non - standard experience, automakers hold the power of discourse, and intelligent driving companies are prone to get involved in price wars."

Facing the recent aggressive layout of Robotaxis by automakers and ride - hailing platform companies, Peng Jun sharply said, "It's always easy to make announcements." "Has Tesla achieved what it has been shouting about for 10 years?"

Automakers generally hope to follow Tesla's example and use their existing end - to - end algorithm capabilities as homologous technologies to support both assisted driving and Robotaxi businesses.

However, Peng Jun provided a different technical judgment from the industry mainstream in the interview. Pony.ai did not follow the technical rhythm of large language models to build an integrated autonomous driving algorithm with super - large parameters. Instead, it adopted an algorithm strategy of multiple small models to improve operation efficiency and reduce computing power dependence.

Regarding the "fear of maps" state in the assisted driving industry, Pony.ai also did not follow the trend. Peng Jun admitted that the company will still adopt the light - map route in the long term.

"You are familiar with the roads you often drive on, so you drive easily. You will feel tired when you go to places you've never been to. This is very normal. So why not use maps?" Peng Jun said. "Even Tesla uses maps."

In the hot field of embodied intelligence, Pony.ai did not enter the arena immediately but chose to observe.

"This is another thing that will take at least 10 years," Peng Jun gave a more direct judgment. "There will always be opportunities to do it, but we still need to see clearly and think clearly."

Peng Jun, CEO of Pony.ai

The following is the edited transcript of the conversation between 36Kr Auto and Peng Jun of Pony.ai:

On the implementation of Robotaxis: Automakers and ride - hailing platforms are not good at it

36Kr Auto: I remember that the industry initially said that Robotaxis (autonomous taxis) would be implemented around 2020. But judging from the progress of Waymo or your company, it seems that this has only been achieved today, about five or six years behind schedule. Where do you think the gap comes from?

Peng Jun: I think it was those who didn't understand that made such claims. I've always believed that Robotaxis require at least 10 years of effort.

Just like Elon Musk saying every year that Robotaxis are coming, but he hasn't achieved it in 10 years. These are the claims of those who don't understand because they clearly don't understand what L4 - level autonomous driving is. Those who made loud claims, such as Cruise under GM, have now failed.

I think those who really understand know its complexity. From technological development to the maturity of laws and regulations and public acceptance, it takes at least 10 years of effort.

36Kr Auto: Now that Robotaxis are in operation, many problems have been magnified. For example, sometimes Waymo's vehicles stop at traffic lights or drive onto railway tracks. Are these the most difficult parts of implementation?

Peng Jun: These problems cannot be exhausted. They will always emerge, and we can only address them one by one. In this world, many things are almost perfect, but to achieve the last one - ten - thousandth or one - hundred - thousandth, we still need to keep working on it.

36Kr Auto: What kind of mechanism should be used to solve these problems?

Peng Jun: There are several systems. First, just like treating a disease, early detection. After detection, we can quickly incorporate general solutions into the development cycle and add them to the boundary conditions. The world model we built can accurately model the vehicle's surrounding environment, including the kinematic model of the vehicle itself and the kinematic models of surrounding traffic participants. By continuously enriching the scenario data in the world model, when the system detects a new special situation, it can quickly include it in the training samples of the world model, enabling the model to learn to handle such boundary conditions and thus improving the generalization ability of the autonomous driving system.

Moreover, in such situations, a good fallback condition must be set.

36Kr Auto: What is your fallback mechanism like?

Peng Jun: There are many. First, there is technical fallback. For example, we have a lot of redundancy in the design. All sensors are redundant. The entire vehicle control, including acceleration, braking, steering, power, electricity, and network, is self - redundant.

Secondly, we have a whole detection and fail - over system. You can imagine it as a three - level degradation mechanism. Our main system is for normal driving. Usually, 99.9% of the time, it's the main system. If the main system really fails, we actually have a mechanism to pull over. If it's on the highway, it can even pull over and find the nearest exit. So it's a degradation mechanism.

If this level also fails, in the worst - case scenario, the vehicle will stop safely within the lane lines. It still has sensors, but the third - level situation is not ideal as it will block the road.

However, there is also remote monitoring, which will detect problems in time and summon ground support personnel to the scene for assistance.

36Kr Auto: You've set a target of 3,500 vehicles this year. How was this calculated? Why 3,000 instead of 5,000 or 10,000?

Peng Jun: Production is relatively easy. Production and inventory are prepared according to demand, based on the development of the market, licenses, and the prediction of the domestic and international situations. There are approximately this many.

Especially many people in the automotive industry who have never been involved in Robotaxis forget that taxi drivers and ride - hailing drivers do many things besides driving, such as charging, maintenance, cleaning, and handling many customer - related matters, like carrying bags. All these things are done by drivers casually. For unmanned vehicles, efficient solutions are needed for all these things, and relevant infrastructure also needs to be built.

This is why ride - hailing platforms claim to have operation capabilities and want to do L4 - level operations. In fact, their so - called operation is different from the actual operation required for Robotaxis. Whether it's ride - hailing platforms or taxi platforms, they don't need to do the things I mentioned above because they only need to manage the drivers, and the drivers do all these things.

But these things are absent in the Robotaxi field, so companies that do operation in other fields don't have a natural advantage.

For automakers and mobility platforms, Robotaxis are a new species and cannot be simply grafted. People say it's a platform - based business, but in fact, the platforms only understand a different set of operations and don't even know what Robotaxis are all about.

36Kr Auto: How does Pony.ai plan to handle these things today?

Peng Jun: We need to do all these things ourselves first, establish standards and platforms. There are actually many technical solutions. For example, how to make the vehicles come back together for charging and how to optimize the peak and valley electricity consumption.

How can you charge and clean 20 vehicles simultaneously? Plugging in the charger only takes half a minute. How can you improve the efficiency of other tasks? There are many things to do. We are essentially establishing a set of standards. Of course, we can also find third - parties to do these things in the future, but for now, we need to establish our own standards.

36Kr Auto: Currently, there are 1,000 vehicles. How many vehicles are matched with one person for ground operations? What is the vehicle - to - personnel ratio?

Peng Jun: When adding up all the ground support staff, remote maintenance personnel, and monitoring personnel, the vehicle - to - personnel ratio is already very low, and the labor cost accounts for a very small proportion of the total operation cost.

36Kr Auto: Is this model based on the current fleet size or a long - term plan?

Peng Jun: In the long run, it may increase slightly, but not significantly.

36Kr Auto: What proportion do you think these invisible operation tasks account for in the barriers of Robotaxis?

Peng Jun: They don't determine whether a company can enter the field (0 or 1), but they determine efficiency. If one person can manage 30 vehicles while another can only manage 20, the costs will be different. So they don't determine entry, but they determine many aspects of the business.

Technology determines whether a company can enter the field. It can be said that 99% of the companies are eliminated. But after having the technology, these things are also very important because they determine efficiency.

On the L2 business: Once price competition starts, it becomes a red - ocean market

36Kr Auto: There is a saying in the industry that the mass - production of intelligent driving is too difficult and takes too long for engineering. Is that why Pony.ai has always focused on Robotaxis and hasn't ventured into the L2 business?

Peng Jun: Mass - production is actually easier, and the L2 market is closer and shorter - term, but it's a red - ocean market. In the end, it's an industry with no profit after intense competition. The Robotaxi market is definitely much larger but also more distant.

From an industry perspective, the technical threshold for L2 is relatively low, and most companies have the R & D and production capabilities. The technical paths and implementation methods are similar, and it's difficult for users to intuitively perceive the performance differences between different products.

There is also no unified standard. Whether the system needs to be taken over every 10 kilometers or 100 kilometers doesn't make a significant difference to ordinary people because in essence, the user (driver) is the fallback. In this situation, automakers hold the power of discourse, and market competition gradually focuses on price, continuously squeezing the industry's profit margin. We noticed this early on.

36Kr Auto: Around what time did you notice this?

Peng Jun: In 2021 and 2022, when many people started to shift their focus to the L2 business.

36Kr Auto: The revenue scale of the L2 business is still quite large. I see that Pony.ai's target for Robotaxi revenue this year is only about $100 million. But if you were to do the L2 business, getting a mass - production project from an automaker could reach this scale.

Peng Jun: The market for Robotaxis is getting bigger and will continue to grow in 10 years. The market for L2 may shrink in the next two years, and its shipment volume won't increase much. Currently, the penetration rate is 60%. Even if it reaches 100%, it's only a doubling of growth. But the price per vehicle is dropping rapidly, so the market is getting smaller, not bigger.

On automakers entering the Robotaxi field: They need to start from scratch

36Kr Auto: Taking first - tier cities in China as an example, there are about 100,000 ride - hailing vehicles. How many Robotaxis would be needed to cover the whole city?

Peng Jun: Our current goal is not to cover the whole city but to reach at least 10% - 20% of the number of ride - hailing vehicles. I think it's achievable.

36Kr Auto: At this time, does the production process or manufacturing standards of the vehicles need to be updated, or can they remain the same as they are now?

Peng Jun: We are constantly updating. There are several capabilities, first of all, operation capabilities and production capabilities. The production standards, capabilities, and durability are also constantly improving. Currently, for our thousands of vehicles, we are fully following automotive regulations.