Pony AI

wrote a post ·

Pony AI Q4 Earnings Call: Target to deploy Robotaxi in over 20 cities with more than 3,000 vehicles by 2026, achieving revenue growth of over threefold

On March 26, autonomous driving company Pony AI $Pony AI (PONY.US)$ / $PONY-W (02026.HK)$ released its Q4 and full-year financial report for 2025, ending December 31, 2025: Q4 Robotaxi business revenue reached $6.7 million, a year-over-year increase of 160%; Full-year 2025 Robotaxi business revenue was $16.6 million, a year-over-year increase of 129%; Passenger fare revenue in Q4 grew by over 500% year-over-year, while the annual increase approached 400%.

Pony AI founder and CEO James Peng stated during the earnings call that he is confident of tripling Robotaxi revenue by 2026, with operations expanding to over 20 cities across domestic and international markets. Following the commercial launch of the seventh-generation Robotaxi, the company has achieved positive unit economics (UE) in Guangzhou and Shenzhen. The current Robotaxi fleet size exceeds 1,400 vehicles, with cumulative users surpassing one million.

After the earnings release, Pony AI executives including CEO James Peng, CTO Tiancheng Lou, and CFO Haojun Wang attended the earnings conference call to discuss key financial results and address analyst inquiries.

[Highlights from the Conference Call]

– By 2026, the company aims to increase Robotaxi revenue to more than three times the level achieved in 2025.For the full year of 2025, the Robotaxi business has become the core growth driver for the company. In Q4, Robotaxi revenue grew by 160% year-over-year, while annual revenue increased by 129%. Notably, passenger fare revenue in Q4 surged by over 500% year-over-year, reflecting a significant acceleration in real-world paid travel demand. The total number of registered users has surpassed one million, nearly triple the figure from the same period last year.

The seventh-generation Robotaxi is rapidly achieving mass production and commercial deployment. The current Robotaxi fleet size has exceeded 1,400 vehicles, surpassing the 2025 target of building a fleet of 1,000 vehicles.By the end of 2026, the company expects its fleet size to exceed 3,000 vehicles.。

Shenzhen operations have become an important validation case for business performance,with daily order and revenue figures reaching record highs.As of March 2026, daily per-vehicle revenue reached a peak of RMB 394 per vehicle, with average daily orders per vehicle hitting 25. In the first two months of 2026, paid orders in Shenzhen exceeded the total volume for the entire year of 2025.

The company proposeda 'China + Overseas' dual-engine expansion strategyand plans to deploy Robotaxis in over 20 cities domestically and internationally by 2026, with nearly half of its operations in overseas markets. Domestically, the Greater Bay Area and emerging first-tier cities are included in the expansion plan. In the Chinese market, the company continues to deepen its presence in first-tier cities' operational areas,entering Hangzhou and Changsha, two emerging first-tier cities, in March. Overseas, business has expanded across Europe, the Middle East, East Asia, and Southeast Asia, reaching a population of 100 million. Recently, through collaboration with Uber and Verne, it entered Croatia,with plans to launch Europe's first commercial Robotaxi service。

– The 'co-built fleet model' will become an essential driver for the next phase of business expansion. By advancing fleet deployment and operations in partnership with ecosystem partners, the company aims to enhance capital efficiency, optimize revenue structure, and accelerate the large-scale rollout of Robotaxis. Pony AI has deepened collaborative partnerships with Toyota, Verne, and Ruqi Mobility, achieving the co-built fleet model. The Platinum Intelligence 4X Robotaxi jointly developed by Pony AI and Toyota began mass production in February, with plans to deploy 1,000 vehicles within the year.

– Technically,Robotaxi has become the first commercially viable application of Physical AI, continuously supporting commercial expansion. The company’s technological moat in the L4 domain enables Robotaxi to provide stable services in high-value and high-complexity scenarios, translating into differentiated experiences that users are willing to pay for, allowing the company to adopt a value-driven pricing strategy while improving unit economics.

– Financially,the company achieved its first-ever single-quarter profit in Q4,primarily driven by strategic investment gains. With over $1.5 billion in cash reserves at year-end, the company is well-positioned to fund R&D, market expansion, and large-scale commercialization efforts.

– Robotruck and licensing/application businesses continue to unlock growth potential: The cost of the next-generation Robotruck autonomous driving suite has decreased by 70%, with operations expected to commence in 2026, driving revenue growth in the second half of the year; shipments of Autonomous Driving Controllers (ADC) are projected to grow sixfold year-over-year by 2025.

[Key Highlights from the Earnings Call]

Pony AI Founder and CEO James Peng:

– 2025 is a pivotal year for Pony AI to complete key commercial validation. The company has made significant progress in revenue growth, fleet expansion, user base increase, and unit economic model optimization. In terms of Robotaxi, the fourth quarter saw substantial growth in both Robotaxi business revenue and passenger fare income. The acceleration of mass production and commercial deployment of the seventh-generation Robotaxi pushed the fleet size beyond 1,400 vehicles, with cumulative users surpassing the one million mark. This also helped Guangzhou and Shenzhen successively achieve positive unit economics (UE).

– In 2026, the company will fully enter a high-growth channel. We are advancing our Robotaxi business at full speed by leveraging fully driverless capabilities, policy support, mass production systems, scaled operational experience, and mature ecosystems. Since achieving UE positivity in February, Shenzhen has maintained robust growth momentum, with daily net revenue per vehicle reaching a record high of 394 yuan. Backed by ample funding from its Hong Kong IPO and with Toyota's Platinum Intelligence 4X seventh-generation Robotaxi entering mass production, the company is confident about exceeding a fleet size of 3,000 vehicles by year-end and tripling autonomous ride-hailing service revenue compared to 2025 levels. By the end of the year, we plan to deploy Robotaxis across more than 20 cities domestically and internationally, replicating this successful model worldwide.

– The company is executing a dual-engine expansion strategy of 'China + overseas' this year. In the domestic market, the company continues to deepen its presence in first-tier cities, moving towards bustling downtown areas, and expanding into Hangzhou, Changsha, and other Greater Bay Area cities. During the Spring Festival, we met strong holiday travel demand in Shenzhen’s Nanshan and Bao’an districts and have entered the busy Guangzhou University Town area. Internationally, the company is accelerating the replication of its technology and operational systems, proven in China's complex urban environments, to Europe, the Middle East, and Asia. This includes launching the first paid Robotaxi service in Doha, Qatar, and advancing fully driverless testing in Dubai, driving international operations from pilot phases toward revenue generation.

– The construction of an ecosystem and co-built fleet model will be crucial for the company to accelerate expansion, reduce costs, and improve capital efficiency. An increasing number of partners are adopting the co-built fleet model, contributing funds to purchase vehicles and sharing operational revenues. Toyota was the first participant in this model, and through collaboration, we have secured 1,000 units of the seventh-generation Platinum Intelligence Robotaxi, jointly promoting the commercial deployment of a thousand-vehicle fleet. Additionally, we are collaborating with BAIC and GAC, leveraging OEMs' mature supply chains and after-sales networks to further reduce vehicle costs, with plans to co-build fleets overseas. To reach a broader user base, we have partnered with Tencent to integrate into WeChat's mobility service platform, enabling hundreds of millions of WeChat users to conveniently hail our autonomous ride-hailing services.

– For Robotruck operations, the company covers major logistics routes from key transportation hubs and ports to consumer destinations across China. It launched the fourth-generation autonomous truck, reducing the cost of the autonomous driving suite by 70%, with plans to begin mass production and operation of this model this year. Currently, the company has deployed fully driverless heavy trucks at Jiangmen Port in Guangdong, completing tests of 1+N fully autonomous convoys under extreme weather conditions, thoroughly validating technical capabilities. The next step is to deploy Robotrucks in more ports and mining areas.

– Technology licensing and applications achieved strong growth, with sales of autonomous driving domain controllers increasing sixfold. Application scenarios expanded to include low-speed delivery, autonomous cleaning vehicles, logistics, and humanoid robotics, with strong customer demand continuing to drive business growth.

Pony AI Founder and CTO Tiancheng Lou:

– 2025 is a critical year for the company to accelerate the transformation of technological capabilities into commercial outcomes. As the first application in the physical AI field to truly achieve commercial deployment, Robotaxi has validated its market value through real-world operational data and user growth. The company has long adhered to an L4-native technology roadmap and continues to invest heavily in the PonyWorld world model. This ensures that technological leadership is not only an engineering milestone but also the core engine of commercial success.

– In terms of fleet scale, the company has established mature multi-OEM mass production synergy capabilities. The seventh-generation Robotaxi is currently advancing mass production across multiple vehicle platforms, including GAC, BAIC, and Toyota, demonstrating the strong generalization ability and platform adaptability of the autonomous driving tech stack, laying the foundation for a fleet size exceeding 3,000 vehicles. This multi-OEM collaboration network accelerates scaling, deepens local cooperation, and enriches Robotaxi model options. Our Robotaxis operate around the clock in multiple cities domestically and internationally, performing robustly in complex urban road conditions, peak travel times, and adverse weather. For example, operations in Croatia cover most of downtown Zagreb from day one in complex urban conditions rather than simple fixed routes, fully proving the robustness of our technology and the vast potential for domestic and international expansion. We are confident in achieving the goal of deploying in over 20 cities globally.

– Efficiency improvements are mainly reflected in cost and operations. On the hardware side, the seventh-generation Robotaxi leverages a platform-based design with higher cost-performance components to optimize and continuously lower BOM costs. On the operations side, a solid safety record helps reduce insurance costs and improves remote assistance efficiency, ensuring the company scales on a per-vehicle profitability basis. Additionally, the highly generalized technological advantage supports the establishment of a comprehensive and scalable operational process, creating ideal conditions for advancing the co-built fleet model. This allows us to expand fleet size faster and with better capital efficiency.

– In terms of user experience and long-term competitiveness, the company has already demonstrated service differentiation advantages in high-value, high-complexity scenarios. It can meet the high-frequency commuting needs of Shenzhen's core urban areas, maintain continuous operations during Beijing’s heavy snowstorms, and the seventh-generation Robotaxi also delivers smoother driving, effectively alleviating motion sickness concerns for users. The Robotaxi has proven its technological capabilities in real-world travel scenarios. Technological leadership is not only a milestone at the engineering level but also the core engine of commercial success. Users voluntarily choose and naturally repurchase due to the excellent experience, without needing discounts to drive demand.

– Looking ahead to 2026, the company will continue to increase investments in AI talent recruitment and infrastructure, promote the iterative upgrade of PonyWorld’s world model, further reduce BOM costs, and sustain rapid commercial growth through technological leadership.

Pony AI Co-founder and CFO Wang Haojun:

– 2025 marks the first year of large-scale commercialization for the Robotaxi business, which continues to serve as the company’s core growth driver, achieving remarkable revenue growth. In the fourth quarter, revenue from autonomous ride-hailing services reached $6.7 million. For the full year 2025, revenue hit $16.6 million. The main growth drivers behind this acceleration were passenger fare services, with fourth-quarter passenger fare revenue surging over 500% year-on-year, and annual growth nearing 400%. This reflects an accelerating release of real payment-based travel demand.

– Even more encouraging is that, just four months after the launch of the seventh-generation Robotaxi, Guangzhou and Shenzhen—two of China’s most valuable cities—successively achieved positive unit profitability. This not only proves the feasibility of the unit profitability model but also demonstrates the positive network effects emerging from increased fleet density: shorter wait times, higher vehicle utilization rates, and rising orders per vehicle, all contributing to continuously improving overall user experience and demand. Robotaxi has entered a new phase where operational efficiency drives revenue growth. In the first two months of 2026, paid orders in Shenzhen have already surpassed the total for the entire year of 2025.

– Looking ahead to 2026, we are highly confident in the growth prospects of the Robotaxi business as we push for further increases in autonomous ride-hailing service revenue. We expect Robotaxi revenue in 2026 to grow at least threefold compared to 2025. At the same time, we will adopt a strategy of 'one core leading, dual engines driving development' to rapidly expand into over 20 cities domestically and internationally by collaborating with partners through a 'fleet co-building model,' adding high-margin recurring revenue streams. This model will reduce our capital expenditure requirements for fleet deployment while further enhancing operational efficiency, allowing us to expand into new regions faster and more efficiently. Overall, the high gross margin characteristics of the Robotaxi segment are fundamentally improving our revenue quality, which will continue to boost the group’s overall profitability.

– Leveraging the mature Robotaxi technology stack, the Robotruck and technology licensing and application businesses are also unlocking growth potential. The cost of the next-generation Robotruck autonomous driving suite has decreased by 70% compared to the previous generation, with mass production deployment planned to start in 2026, and revenue growth expected to accelerate further in the second half of the year. On the technology licensing and application side, strong demand for autonomous driving domain controllers (ADC) is evident, with ADC shipments increasing sixfold year-on-year in 2025. Customer demand has expanded to multiple scenarios including low-speed delivery, autonomous cleaning, logistics, and robotics, providing the company with more scalable revenue sources.

– In terms of financial fundamentals, we achieved our first quarterly GAAP-level net profit, showcasing the success of our full ecosystem strategic investments. As of December 31, we held $1.5 billion in cash reserves, providing a solid capital foundation. We will prudently yet aggressively increase investments in R&D, large-scale deployment, and market expansion. We are confident that these investments will accelerate the large-scale commercialization process and lead to faster revenue growth in 2026. The industry remains in a critical phase of capturing market share and strengthening technological and operational barriers. With ample cash reserves, optimized costs, and superior technology, we are confident in sustaining revenue growth and continuously improving long-term profitability.

[Q&A Session]

Bank of America Securities Lee Myeong-Hoon

Regarding your target of surpassing a fleet of 3,000 Robotaxis by the end of 2026, could you share the capacity ramp-up and deployment plans? Additionally, after achieving profitability per vehicle in Shenzhen and Guangzhou, what is your outlook for future unit economics?

Dr. Peng Jun

I believe achieving positive profitability per vehicle is a significant victory not only for us but for the entire industry. It proves that our technology can be effectively validated in the real world, showing that Robotaxis are not only technically feasible but also profitable when scaled. After achieving this milestone in Guangzhou, we have again reached it in Shenzhen, validating the replicability of the model. This achievement stems from our focus on enhancing user experience rather than relying solely on discount strategies.

From a policy perspective, we are indeed seeing favorable tailwinds for the Robotaxi industry. Domestically, with the coordinated efforts of central and local governments, more cities are opening up Robotaxi services, and existing markets are issuing more licenses to support larger fleet sizes. Globally, many countries are drawing on the progress made in China and the US to remove regulatory barriers and accelerate deployment. The positive momentum at the regulatory level gives us full confidence to replicate our successful model in more domestic and international markets.

Therefore, to seize this opportunity, we are accelerating capacity ramp-up and expanding operations into more cities.

Regarding capacity ramp-up: Over the past two months, we have focused on the production of Toyota Platinum Intelligence 4X Robotaxis, along with the continued rollout of BAIC and GAC vehicles. We are confident that the fleet size will exceed 3,000 units by the end of the year.

Regarding expansion plans: First, we will continue to deepen our presence in core downtown areas; second, we will expand into new markets such as Hangzhou and Changsha, and plan to cover more cities in the Greater Bay Area; third, given that current pricing domestically is lower compared to overseas, yet still delivers positive unit economics with ongoing optimization, we expect profit margins in overseas markets to be even more attractive. We plan to enter at least 20 cities to establish a stronger first-mover advantage. Our co-built fleet model will also optimize capital expenditure efficiency, accelerate fleet expansion, and grow Robotaxi revenue to at least triple this year.

Morgan Stanley Tim Hsiao

Regarding Pony AI's latest operational scope under the dual-engine strategy mentioned by management, given the company’s expansion strategy into over 20 cities this year, could you share the specific list of cities and the proportion between domestic and overseas markets? Additionally, facing geopolitical tensions in the Middle East, has the company encountered any challenges or resistance?

Dr. Peng Jun

From a strategic perspective, we will use our successful experience in China as a blueprint for overseas expansion. Our technology and business model have been fully validated and can be quickly replicated in overseas markets. In fact, we expect that nearly half of our target to enter 20 cities by 2026 will be located overseas, including regions such as Asia, Europe, and the Middle East.

In terms of market penetration strategy, our approach of co-building fleets with industry leaders significantly optimizes capital expenditure efficiency. This model not only helps us scale efficiently but also assists us in building a robust localized network. Currently, we are collaborating with international giants like Uber, Bolt, and Stellantis. One example is that we launched the first commercial Robotaxi service in Europe together with Uber and Verne.

Looking ahead, we will explore more cities in Europe and further expand our presence in the Asian market, such as in South Korea and Singapore.

Regarding the last point of your question, the Middle East market. First, this remains one of our key focus areas at present. So far, the current geopolitical tensions have had almost no direct impact on our business. We continue to advance our operations in the GCC region. Next, we plan to launch a paid service in Doha, Qatar, in collaboration with Mowasalat; in Dubai, we are also prepared to fully initiate unmanned operations once approval is granted later this month.

CITIC Lyon You Yang

Pony AI's world model and autonomous driving technology stack have already been deployed in multiple cities both domestically and internationally. I would like to ask how this technology generalizes and adapts to new environments. What role does the world model play in accelerating urban expansion?

Dr. Lou Tiancheng

First, the essence of driving lies in interactions and negotiations with surrounding dynamic traffic participants. Whether it’s in Guangzhou, Shenzhen, or Zagreb, this essence never changes. The difference between various cities and countries lies in the probability distribution of similar scenarios rather than the fundamental nature of the challenges. Here are a few typical corner cases: for example, aggressive lane changes without checking mirrors, or a fallen bicycle suddenly appearing on the road—these scenarios can happen anywhere.

Our technology has undergone the most rigorous validation—we have achieved full-scale operation under all time periods (including peak hours) and weather conditions in densely populated core areas of major Chinese cities. This means that when we enter a new city like Zagreb, we are not starting from scratch. The capabilities of the systems we deploy already cover all potential scenarios in the new market. Therefore, we can directly commence operations in Zagreb’s high commercial value downtown area.

To address the second part of the question, the world model serves as a key enabler in accelerating this process. It accurately models the interaction dynamics and negotiation behaviors between vehicles and surrounding traffic participants, and generates large-scale simulation scenarios reflecting specific traffic patterns of new markets. Through reinforcement learning conducted in these simulated environments, our system continuously refines its driving strategies. This allows us to avoid collecting massive amounts of real-world test data every time we enter a new city, enabling efficient validation and fine-tuning.

The core driving our expansion into over 20 cities is very clear: a multi-OEM collaboration network provides us with a diversified vehicle platform tailored to local needs; a standardized operational system, from remote assistance to fleet management, ensures our processes are highly standardized and highly replicable; our technical support covers a wide range of Operational Design Domains (ODD), enabling us to operate in complex urban environments rather than being limited to low-difficulty specific simple routes.

In summary, we are highly confident in achieving the goal of deploying Robotaxi services in over 20 cities domestically and internationally by the end of 2026.

UBS Group Fang Xinyu

My question is about the co-built fleet model. Regarding the new vehicles planned for 2026, could you please elaborate on how this model will be applied? And how should we view the benefits this model brings to the company and value chain partners?

Dr. Wang Haojun

As everyone has seen, we have achieved the key milestone of turning单车盈利 positive in Guangzhou and Shenzhen. Following that, we observed numerous partners across the entire ecosystem eager to enter the Robotaxi sector, and Pony AI has become their preferred partner. Under this co-built fleet model, partners are responsible for funding vehicle purchases while integrating into various aspects of the Robotaxi value chain, such as ground operations, vehicle maintenance, and charging. This creates a win-win situation: partners can generate continuously growing revenue from deployed vehicles, while we can rapidly expand our fleet size without tying up substantial amounts of our own capital. This year, nearly half of the new vehicles will be implemented through this model, with Toyota being the primary adopter. This model not only enhances capital efficiency during our expansion but also opens up new revenue streams for us—through revenue sharing or ongoing income from 'AI virtual driver' technology licensing fees, combined with our self-operated passenger fare revenue, which will collectively drive our 2026 Robotaxi revenue to more than triple that of 2025.

In addition to our current lineup of partners such as Toyota, Ruqi Mobility, and ATBB, we actually expect more partners to join this year.

Huatai Securities He Pianpian

Regarding the 1,000 Robotaxis contracted with Toyota, what are your plans for deployment? Are there further plans for scaled expansion and strategic initiatives with Toyota in the future?

Dr. Peng Jun

Toyota is far more than just an ordinary partner for us. Since 2019, Toyota has become our largest strategic shareholder. This means our cooperation has long surpassed a simple supplier relationship and evolved into deep strategic synergy.

In terms of mass production of Robotaxis: since 2019, we have jointly launched multiple Robotaxi models based on Toyota's platform. By 2026, we will add over 2,000 vehicles, nearly half of which will be based on the all-new Toyota bZ4X, the seventh-generation model. This vehicle was co-developed by us with Toyota China and GAC Toyota and is already in official mass production on Toyota’s production line. It can be said that there is a very strong synergy between us—Toyota's top-tier manufacturing capabilities and first-class vehicle platform, combined with our L4 autonomous driving technology and operational experience, have achieved perfect synergistic benefits.

Considering our cooperation, beyond joint development, Toyota is also the first partner to adopt our 'co-built fleet' model, where they fund the fleet, helping us achieve efficient capital expansion. This fully demonstrates Toyota's strong confidence in Pony AI. Moving forward, we will join hands to start from first-tier core cities domestically and advance commercial implementation.

HSBC Bank Ding Yuqian

The first question is about the competitive landscape: how do you view automakers entering the Robotaxi field? The second question relates to competitive advantage: currently, market trends are shifting towards scale, and as more players enter, what is Pony AI's most unique competitive advantage?

Dr. Peng Jun

I'll answer the first question. Recently, we have indeed seen many announcements from new players declaring their entry or plans to enter the Robotaxi space, including automakers, ride-hailing companies, tech giants, and even startups. Overall, the influx of new players precisely shows that everyone sees the long-term potential of this space, and we certainly welcome it as we work together to grow the market. However, the reality is that L4, especially Robotaxi, is an extremely complex system engineering challenge that requires an integrated solution. As I mentioned earlier, the Robotaxi industry relies on five key pillars: technology, policy, mass production, operations, and ecosystem collaboration. These elements are interconnected, and simply pouring in resources won’t accelerate development. Over the years, we have actually developed unique advantages across various aspects of the Robotaxi industry. Regarding our specific competitive barriers, I’ll let Tiancheng elaborate further.

Dr. Lou Tiancheng

Technologically speaking, I don’t believe automakers automatically gain an advantage in L4 Robotaxis just because they excel at manufacturing or have made progress in L2 systems. L2 and L4 are fundamentally different—they are not two stages on the same path. For L2 systems, improving MPI (miles per intervention) might paradoxically increase actual risks. Semi-autonomous driving tends to create a dangerous illusion that the system is 'basically fine,' until it suddenly fails at a critical moment when human drivers are often out of sync and unable to take over in time. This is why L2 cannot naturally lead to L4, especially when discussing the scaled operation of fully driverless fleets.

One of Pony AI’s unique advantages is our long-term investment in world models and L4-native virtual driver training methods. This matters because L4 Robotaxis must significantly outperform human drivers in safety, and this cannot be achieved merely by mimicking human driving behavior. To reach such a level of safety, the system must continuously evolve through large-scale trial and error in virtual environments, making world models essential. In other words, the key to training L4 virtual drivers lies in building a virtual environment with robust simulation-to-reality (sim-to-real) capabilities, particularly regarding vehicle interactions. This is also why achieving L4 requires years of accumulated AI investment rather than relying primarily on collecting more real-road data.

The second unique advantage is that we have a real fleet of Robotaxis in operation, which continuously helps us identify the gaps between our world model and the real world. The hardest part of L4 isn't the first 99%, but the last 1%—those rare but safety-critical corner cases. To safely handle these scenarios, looking at individual trajectory data points is far from enough; what truly matters is understanding the various behaviors of other vehicles and pedestrians when interacting with AI drivers under different intentions. This is precisely why the world model is so crucial: only the world model can fully cover the complete combination space of different intentions in corner cases, and this coverage is the fundamental guarantee of L4 safety. At the same time, only data from fully driverless Robotaxis can continuously narrow the gap between the world model and the real world. Real-world Robotaxi operations allow us to observe and understand the actual behavior patterns of vehicles and pedestrians in these scenarios, whereas human driving data cannot capture interactions with Robotaxis. This also explains why new entrants typically start with small fleets. Regulators understand this logic, so they are naturally very cautious during the early stages of license issuance.

Overall, automakers entering the field confirms the enormous opportunity in this sector, but L4 Robotaxi cannot be achieved by simply extending L2 capabilities. Pony AI's unique advantages come from two aspects: one is the world model specifically designed for L4, and the other is our real fleet of Robotaxis that continuously optimizes the world model. These two form a closed loop, driving the continuous evolution of both the model and the product.

Nomura Securities Joel Ying

How does management view the impact of NVIDIA's release of its open-source L4 autonomous driving model during this year's CST period?

Dr. Lou Tiancheng

I believe the key to this question lies in distinguishing between a 'model' and a 'real product.' An open-source autonomous driving model might be a good starting point, but it is not the final product. There is a significant gap between having a model and having a fleet of Robotaxis that have undergone safety verification, received government approval, and can operate commercially at scale. Bridging this gap is precisely where our core advantage lies.

At Pony AI, our advantage stems from years of full-stack self-research capabilities and extensive experience in deploying real L4 Robotaxis. This includes not only software or algorithmic models but also covers vehicle architecture, sensor and redundancy design, domain controllers, operational systems, validation frameworks, and the commercial capabilities required to run a true Robotaxi company. For example, customizing sensors and designing redundancies directly impacts safety, manufacturing capabilities, and the bill of materials (BOM). Our collaboration with OEMs is far from a simple plug-and-play solution; it’s a deep partnership that brings better system integration, higher reliability, and lower overall costs.

Therefore, we believe NVIDIA's advancements contribute to the development of the entire ecosystem. However, we firmly believe that the entry barrier for this industry remains extremely high.

Of course, NVIDIA is also an important partner for us in the domain controller area, and we have maintained a close cooperative relationship.

CITIC Securities Lu Tianyu

How does the company plan to utilize the funds raised from last year's Hong Kong IPO? Given the acceleration of fleet deployment goals, will the cost and expenditure projections for 2026 be revised upwards?

Dr. Wang Haojun

As I mentioned earlier, we have a substantial cash reserve. As of December 31, 2025, thanks to the over 800 million USD raised from the Hong Kong IPO, our cash reserves reached 1.5 billion USD, providing solid long-term capital support for our continued growth in the coming years.

Our achievements in the mass production, deployment, and per-vehicle profitability of the seventh-generation vehicles in 2025 have established us as clear industry leaders. Looking ahead, we aim to accelerate revenue growth to further expand our leading position and propel the entire industry into its next phase. Therefore, we will strategically increase our investments.

We have significantly expanded our operational vehicle fleet in first-tier cities in China, particularly in Shenzhen and Guangzhou, and successfully entered new markets such as Hangzhou, Changsha, and internationally, Croatia. To achieve our goal of covering more than 20 cities domestically and internationally by the end of the year, we will focus on increasing investments in business expansion, operations, and marketing.

As we scale up our deployments, we will adopt a dual-driven model of 'co-built fleets' and 'self-operated vehicles' to expand our fleet. Meanwhile, we will recruit AI talent and invest in AI infrastructure to further enhance the capabilities of our 'virtual driver.' This ensures that we continue to meet the high public expectations for safety, reliability, and high-quality service, delivering trustworthy Robotaxi services.

As James mentioned, this is a critical window to expand market share, and necessary investments are crucial to solidifying our technological and operational moat. We believe that strategically increasing investments is a wise and value-oriented trade-off. Through rigorous capital allocation, combined with the advantages of the co-built fleet model, these investments will translate into faster revenue growth, broader city coverage, and a larger fleet size, further establishing our industry leadership.

CICC (China International Capital Corporation) Xiao Kai

I have a question regarding the current rise in raw material prices, such as memory chips. What impact does this have on the company’s production plans and costs?

Dr. Wang Haojun

As I mentioned in my previous remarks, whether it is the bill of materials (BOM) for the entire vehicle or the autonomous driving kit (ADK), the impact of rising memory prices is limited. This is mainly due to our forward-looking supply chain strategy and the inventory synergies brought by the autonomous driving controller (ADC) business. Long before the market saw price increases and supply shortages, we had already secured sufficient supplies of memory chips. This move has provided a solid guarantee for our production target of over 3,000 Robotaxis this year.

Thanks to the aforementioned supply chain measures and the release of economies of scale, our plan to reduce the ADK's bill of materials by 20% in 2026 compared to 2025 is proceeding as scheduled. Additionally, with our ongoing software and hardware optimization, there is further room for overall material cost reductions in the future.

29K Views