Good evening, everyone. Welcome to Momenta's first half 2026 interim results conference call.

This is Simon Shao, Senior Director of Corporate Investment and Financing Team. Before we start, I would like to quickly remind everyone that today's discussions will contain forward-looking statements.

These statements involve risks and uncertainties, and actual results may differ materially from our current expectations. Please know that numbers stated in management's prepared remarks are in RMB terms, and we will discuss non-IFRS measures today, which are more thoroughly explained and reconciled to the most comparable measures reported in the company's earnings release and filings with the Hong Kong Stock Exchange.

Joining us on the call today for the management presentation are our Founder and CEO, Xudong Cao, and our Chief of Staff, Anren. Additionally, our Senior Vice President, Huan Sun, will join us later for the Q&A session.

With that, it is my great pleasure to turn the call over to our CEO, Xudong. The floor is yours.

Good evening, everyone. I'm Xudong, CEO of Momenta.

Looking back at the first half of 2026, we remained deeply committed to our core mission: better AI, better life. We stayed firmly on our data-driven technology path, fully dedicated to building a world-leading foundational platform based on what we call our world model.

Serving as our unified technical base, the world model allows us to build a profound understanding of the physical world. Through its three layers of pre-training, simulation, and reinforcement learning, we can horizontally integrate our solutions across mass production, robovan, and robotaxi, as well as other future-oriented scenarios.

These technological advancements have translated directly into our business outcomes. To give you a high-level overview, our financial performance in the first half of the year was exceptionally strong.

Our total revenue reached RMB 1.6 billion, delivering a robust year-over-year growth of 76%. Our adjusted gross margin stood at an impressive 75%, which remains well above the industry average.

At the same time, we have made significant strides toward profitability. Our adjusted net loss for the period was just RMB 14.1 million, narrowing by 97% compared to the same period last year.

However, I want to emphasize that immediate profitability is not our primary objective at this current stage. Moving into the second half of the year, we will continue to strategically invest in R&D, computing infrastructure, as well as the scalable commercialization and operations of our robot businesses.

This remarkable improvement clearly demonstrates the operational efficiency and scalability of our business model. Underpinning both our financial success and our technological edge is our immense data moat.

Coupled with our continuously expanding computing infrastructure, this strong foundation has driven scalable and high-quality growth in our mass production business. We have delivered 105 mass-produced models, and our cumulative design wins have reached 219 vehicle models across 26 OEM customers.

Our cumulative installations of solutions for mass-produced vehicles have officially broken through the 1 million unit mark. This firmly secures our position as the number one third-party supplier in the urban NOA market.

Furthermore, we are proud to say our footprint has now expanded to 10 countries and regions. Beyond passenger vehicles, our commercial progress is well under way.

Our robovan has initiated small-scale trial operations in Suzhou, while our robotaxi is now testing or operating in six cities across three countries. Collectively, these comprehensive commercial operations have already empowered us to accumulate over 13 billion miles of real-world operational data.

Looking forward, our latest generation R7 world model are being gradually deployed in mass-produced vehicles starting in the third quarter. By translating the triple capabilities of pre-training, simulation, and reinforcement learning into real-world applications, R7 will deliver an intelligent driving experience that is much more human-like and, above all, much safer.

We confidently expect NOA penetration, particularly urban NOA, to continue its upward trajectory in the quarters to come. Against this highly favorable market backdrop, our mass production business solutions achieved exceptional growth.

As of June 30th, we have established mass production partnerships with 26 OEM customers, spanning across domestic, international, and joint venture brands. Our cumulative design wins have now reached 219 vehicle models.

To give you some context, that is a solid increase of 49 models compared to the end of 2025. More importantly, we currently hold a backlog of 114 secured yet-to-be-delivered models in our pipeline.

This massive backlog provides us with exceptional visibility and virtually guarantees the sustained growth of our mass production business moving forward. Meanwhile, we successfully delivered 37 new mass-produced vehicle models in H1, bringing our cumulative total to those 105 models on the road today.

We maintain our position as the number one third-party supplier in the urban NOA solution market. Drilling down into the numbers, we added approximately 321,000 new units during the first half, with robust 84% year-over-year growth.

As I highlighted in my opening remarks, our cumulative installations officially surpassed the 1 million unit milestone. More importantly, the pace at which we achieve every 100,000 new installations is accelerating rapidly.

It took us nearly 24 months to reach our first 100,000 units. Today, it takes us only 36 days.

This exponential acceleration not only validates the explosive market demand for our solutions, but also perfectly illustrates the compounding power of our business and the data flywheel. Our commercial strategy for our mass production business can best be described as a blanket coverage strategy, in Chinese [putiangaidi?]

. In essence, this means maximizing our footprint to cover as many OEM customers and diverse vehicle models as possible.

The logic behind this is clear: as competition in the automotive market intensifies, banking on a single vehicle model or hoping for a one-off blockbuster hit is far from enough to sustain long-term business viability. Furthermore, in today's highly complex and volatile environment, even our OEM clients find it incredibly difficult to accurately forecast the sales volume of any specific model.

That is exactly why our strategic foundation relies on the law of large numbers. By diversifying our presence across a massive matrix of OEMs and vehicle models, we effectively neutralize individual market risks.

This is exactly how we will ensure our long-term victory in this ever-changing and highly unpredictable automotive landscape. During the first half of this year, we continued to deepen our partnerships with customers.

Vehicles equipped with our solutions have delivered a growing number of best-selling models. Let me share a few standout examples.

Among our domestic OEM customers, one leading partner's flagship large SUV is currently achieving monthly sales of approximately 10,000 units with Momenta solutions. Similarly, their mid-to-large SUV is recording about 14,000 units with our solutions in monthly sales.

Furthermore, the MG7 recorded sales of over 45,000 orders in just 19 days, and the Roewe D7 secured over 10,000 blind booking orders in a mere 7 days. Turning to our global OEMs and joint venture partners, we are experiencing equally strong momentum.

To name a few: the new BMW iX3 secured over 30,000 new orders within just two days, and the new generation Mercedes-Benz GLC SUV EV received more than 12,000 reservations within just 24 hours. Additionally, within the joint venture new energy SUV sector, models powered by Momenta solutions have consistently dominated the top five spots on the sales charts over recent months, with the Buick Electra E7 and Bozhi 3X serving as prime examples.

This robust market validation proves our value. I firmly believe that the upcoming launch of our R7 solution will further empower our customers to continuously boost their vehicle sales performance.

Securing high-caliber customers and building long-term partnerships with them is no small achievement. This is particularly critical for our business given the lengthy automotive industry product cycle.

Take Mercedes-Benz, our partner of more than nine years, as a perfect example of combined strategic investment and deep-rooted business partnership. Mercedes-Benz first engaged with and invested in Momenta back in 2017, followed by multiple subsequent rounds of investment.

It also acted as a cornerstone investor in our IPO. Like I mentioned before, our intelligent driving solution has already been deployed on the all-new electric CLA.

The all-electric GLC and GLE SUV made their debut recently. In addition, we will roll out our robotaxi services together in Abu Dhabi.

Mercedes-Benz applies extremely rigorous standards when selecting trusted advanced driving solution providers. Momenta has consistently delivered state-of-the-art solutions with remarkable speed to meet their requirements.

Turning to our roadmap for next-generation capabilities, we have already initiated L3 R&D projects in collaboration with multiple OEMs. In the first half of this year, we officially announced alongside Audi that the E7X model will serve as launch model equipped with Momenta's Level 3 mass production solutions.

Provided that regulatory frameworks are in place, we expect to achieve mass production of L3-capable models within 2027. We firmly believe the commercialization of L3 will serve as a major catalyst for the whole industry, further accelerating the penetration rate of advanced intelligent driving.

Regarding our global strategy, overseas expansion has clearly become a defining growth trend for us. As of today, vehicle models equipped with Momenta's solutions have successfully expanded to 10 countries and regions around the world, including markets such as Norway, the United Kingdom, the United Arab Emirates, Singapore, Australia, New Zealand, and Thailand.

Our global expansion is advancing on two fronts. Beyond supporting Chinese vehicle models in their overseas expansion, we are also actively progressing international projects directly with several global OEM customers.

By combining the global rollout of Chinese models with the future mass production of these international projects, we are confident that our market penetration and brand influence on the global stage will be significantly strengthened. Next, let's move to our robovan business.

We established a strong foundation in 2025 by achieving our minimum viable product closed loop in Suzhou. Building on that, we have made highly positive progress in the first half of this year by officially launching small-scale trial operations.

Empowered by our technological edge, our roadmap moving forward is clear: for the remainder of 2026, our goal is to complete commercial closed-loop validation across two major domestic cities, supported by the operational launch of our version 1.5 mass-produced vehicles. Moving into 2027, we will enter a phase of rapid replication.

We will expand our footprint to dozens of key domestic cities and take our first global step by entering the UAE market. This will be fueled by the delivery of multiple version 1.5 models and our next-generation version 2 vehicles.

Finally, by 2028, these efforts will culminate in fully scaled operations across China and a robust international presence spanning the Middle East, Europe, Japan, South Korea, and Southeast Asia. All of this will be powered by a mature, comprehensive vehicle matrix, firmly securing our position as the industry leader, aiming to become the number one robovan platform globally.

Building on this momentum, we have already initiated commercial deployment in the robovan sector, securing preliminary commercial partnerships with multiple key customers. Specifically, we are providing solutions for express delivery network customers, including industry leaders such as JD.com, SF Express, and China Post, as well as franchise outlets of major players like ZTO, STO, YTO, and J&T Express, just to name a few.

Furthermore, our commercial scenarios are continuously expanding, now covering diverse use cases such as nighttime delivery, fresh produce, beverages, and fast-moving consumer goods. Similarly, our robotaxi business is making significant breakthroughs.

By the first half of 2026, we secured crucial licenses and expanded our global footprint, with testing or operations now active across six cities in three countries. Focusing first on our domestic layout, we have successfully obtained licenses in four major Chinese cities.

This rapid progress is a strong testament to both our technological maturity and our highly effective regulatory communication. Let me highlight these key milestones: in Shanghai, we secured the driverless demonstration operation license in Pudong this January.

We plan to officially launch public commercial operations jointly with EnjoyGo within 2026. Remarkably, it took us only eight months from starting road tests to securing this driverless license.

This sets a new industry record and fully validates the safety and readiness of our technology. In Suzhou, we obtained the driverless demonstration application license in Xiangcheng District.

This was the very first license of its kind in Jiangsu Province, marking a critical step in our commercial layout across the Yangtze River Delta. In Wuxi and Shenzhen, we have continuously expanded our testing footprint.

We secured manned road test licenses in Wuxi this January, and most recently in Shenzhen this July, which proudly marks our official entry into the core of the Greater Bay Area. Beyond our domestic achievements, our overseas expansion has been equally remarkable, proving the strong global generalization capability of our technology.

In Abu Dhabi, we secured our road testing license in September 2025. Impressively, it took only three months from the arrival of our first vehicle to obtain this license, setting another local industry record for speed.

Currently, we are actively pushing for the operational license and expanding our ODD. We are on track to launch public commercial operations within 2026 in close collaboration with our Middle Eastern partner, Lumo.

In Germany, just this July, we reached a massive milestone by securing a nationwide Level 4 testing license from the KBA, the German Federal Motor Transport Authority. This allows us to conduct Level 4 autonomous driving tests on urban roads across the entire country.

We are incredibly proud to be the very first Chinese company to receive this qualification. This achievement not only validates the international competitiveness of our technology, but also lays a rock-solid foundation for our deep strategic layout in the European market.

Looking ahead, we have a very clear roadmap for scaling our robotaxi business over the next few years. By firmly leveraging our data-driven map-free approach, we are rapidly expanding across both domestic and international markets.

This strategy gives us the distinct advantages of massive data accumulation, significantly lower vehicle costs, and a much faster path to profitability. Let me share more details: by the end of 2026, we target to secure testing licenses in over 10 cities globally, with our fleet size reaching hundreds.

Moving into 2027, we plan to expand our footprint to over 30-plus cities worldwide, scaling our fleet to thousands. More importantly, from a technology standpoint, we expect our algorithmic safety to achieve a level that is 10 times safer than average human drivers.

Building upon this foundation of safety, 2028 will mark the year of massive scaled operations for us. We will successfully transition our robotaxi fleet from pilot trial capacities into a normalized, everyday commercial mobility service.

This milestone will signify our ultimate leap from technological leadership to large-scale commercial operations. All of these incredible commercial milestones, from mass production to our global robotaxi fleet, are driven by one unified technological core.

As we accelerate toward large-scale commercialization of L4 autonomous driving, we remain firmly committed to our data-driven technology path. Today, I want to clearly define Momenta's core positioning in this new era.

Our core positioning is building a global-leading world model foundation platform. This single, powerful foundation empowers our entire ecosystem.

It horizontally integrates our solution across mass production, robovans, robotaxis, and robotrucks. Looking even further ahead, it lays the very groundwork for what we call real intelligence, bringing embodied AI into the physical world.

We truly believe autonomous driving is the first massive application of physical AI. To achieve large-scale L4 commercialization, our world model platform relies on three core layers: pre-training, simulation, and reinforcement learning.

First, pre-training: building on massive data from over 1 million vehicles, 13 billion driving miles, and over 100 million golden data clips, our model is able to directly internalize physical laws. Second, simulation: we significantly boost our validation efficiency by focusing on scarce long-tail scenarios and applying scientifically weighted mapping.

Currently, we run 10 million simulated kilometers every single day. We intelligently categorize our datasets by reproduction difficulty: extremely difficult, moderately difficult, and easy, applying corresponding mapping weights.

Through this scientific approach, our 10 million simulated kilometers are actually equivalent to a staggering over 4 billion kilometers of real-world road testing. This perfectly recreates long-tail hazards and drastically reduces our physical testing costs.

Third, reinforcement learning: through a continuous reward mechanism, our AI doesn't just copy good human drivers; it actively evolves to surpass them in both safety and efficiency. Ultimately, our world model is driven by a powerful positive flywheel across four dimensions.

First, data scaling: our real-world driving data will grow from 10 billion kilometers in April to 24 billion kilometers by the end of this year, creating an unmatched closed-loop data barrier. Second, compute scaling: in the cloud, we are expanding from 20,000 to up to 60,000 GPUs this year.

Third, model scaling: fueled by this data and compute power, our model parameters are expanding rapidly. Finally, business scaling: bigger models deliver better product experiences, which secures more design wins and drives revenue.

More vehicles generate more data, and more revenue funds more compute. It is a perfect self-accelerating cycle.

Powered by this unified platform, we are now rapidly expanding our commercial footprint from mass production into our robovan and robotaxi operations globally. With that, I will conclude the strategic and business overview.

I'll now hand the call back to our moderator. Thank you.

Thanks, Xudong. Now, to discuss our financial results for the first half of the year, I will hand the call over to our Chief of Staff, An Ren, please go ahead.

Thank you, Simon. Hello, everyone.

I am Anren, Chief of Staff of Momenta. Please note that we will discuss some non-IFRS financial measures today.

You can find detailed explanations in our earnings release and our filings with the Hong Kong Stock Exchange. As Xudong highlighted earlier, I am very pleased to report that we delivered an exceptionally strong financial performance in the first half of the year.

Our results are characterized by robust top-line growth coupled with resilient pricing. For the first half of 2026, our total revenue reached RMB 1.6 billion, representing a significant year-over-year increase of 76%.

Breaking down that top line: our NRE revenue reached RMB 995 million. This was fueled by our growing customer base, especially among global OEMs, combined with higher delivery volumes thanks to our improved efficiency.

Meanwhile, our licensing revenue reached RMB 607 million. This growth was driven by a rapid increase in urban NOA penetration in the first half of this year.

As the market leader, we also accelerated mass production across more vehicle models, directly driving our rapid sales growth. In a highly competitive market environment, our licensing ASP has remained highly stable.

This powerful combination of surging revenue and a stable ASP clearly demonstrates the premium value that our OEM customers place on our technology. Moving further down the income statement, our profitability profile continues to strengthen.

For the first half of the year, our gross profit reached RMB 1.2 billion, representing a stellar 75% growth year-over-year. Now, turning to our bottom line, we made substantial progress.

Our adjusted net loss narrowed by an impressive 97% year-over-year to just RMB 14 million. In short, our business is scaling rapidly, our margins are expanding, and we are on track to move toward profitability in the long term.

However, echoing Xudong's earlier point, maximizing short-term profitability is not our immediate focus. Moving into the second half, we will sustain our strategic investments in R&D, computing infrastructure, and the commercial scaling of our robot businesses.

We are continuing to step up our investments in R&D. For the first half of the year, our R&D expenses reached RMB 1.2 billion, representing a 19% increase year-over-year.

At the same time, our combined selling, general, and administrative expenses grew by approximately 50% year-over-year as we continue to scale our commercial operations globally. What this tells us is that our overall operating leverage is continuously improving.

Beyond the income statement, this operational excellence also extends to our working capital. This efficiency is directly reflected in our cash flow.

During the reporting period, our operating net cash outflow narrowed significantly to RMB 381 million, representing a robust 46.8% improvement year-over-year. Our accounts receivable days have continued to improve steadily and modestly to 105 days.

These metrics are clear indicators that our organic cash generation capability is strengthening rapidly. Furthermore, our balance sheet remains exceptionally strong.

We currently hold approximately RMB 15 billion in cash reserves after IPO. This robust liquidity provides us with an abundant runway and deep financial resources, ensuring we are fully funded to support our long-term strategic investments and technological R&D.

That concludes our prepared remarks for today. Thank you all for your time and attention.

Operator, we are now ready to open the call for questions.

旭东总、孙总、安总，感谢接受我的提问啊。首先恭喜公司取得亮眼的上半年业绩。那我们看到今年上半年其实汽车的整体需求放缓，然后智驾呢也是弱于预期的，那Momenta是如何在这样的环境下保持了非常高的一个收入增长？也想请教一下怎么看待行业长期的竞争格局，那公司想怎么去维持第三方供应商市场第一名的领先位置？I'll translate it by myself. Good evening.

Thanks for taking my questions, and congratulations to the company on your outstanding performance in the first half of the year. Given the overall slowdown in automotive demand in the first half of this year, how does Momenta maintain high revenue growth in such an environment?

I would like to ask for your views on the long-term competitive landscape of this industry and how the company maintains its leading position as the number one in the third-party supplier market. Thank you.

非常感谢，非常好的问题啊。这边是那个旭东啊，对。就首先我先先回答第一点啊，就是我们在一个特别大的市场里面，这个市场的话这个智驾的天花板也非常高。25年城市NOA的渗透率的话是超过了10%，那今年上半年的话城市NOA的渗透率的话大概率已经超过了15%，而且还在快速地增长啊。那我们获得好的业绩回报的话，其实简单来说就是好的技术、好的产品和好的交付啊。这三点都做好的话是非常非常稀缺的啊。这个技术的话后面有机会就再展开介绍。那个产品的话这个我可以讲一下啊，就我们去年的话这个R6就引入了这个强化学习，今年的话我们的R7引入了世界模型。在在技术和这个产品上的话是持续地坚持创新，让消费者的体验的话持续地提升，更加地安全，更加地安心，做到好用、爱用，这个这个离开的一个程度，嗯。那在交付上的话我们也在不断地提升我们的交付质量啊。去年我们交付了40款车，今年上半年的话我们就已经交付了37款车。整个交付的话都是围绕着以客户为中心，然后去通过这个标准化、流程化、自动化的方式去提升我们的交付质量和效率，对。那最终的话就是持续的领先一定来自于好的技术、好的产品和好的交付，嗯，对。

Now is the translation. Thank you, Brenda.

Autonomous driving is a massive market with very high potential. We see that in 2025, the penetration rate of urban NOA in China was roughly more than 10%, and we see that in the first half of this year, this penetration rate has been rising towards more than 15%, which is showing a very rapid growth.

Delivering strong performance essentially boils down to consistently building a good technology, building a good product, and good execution. I will later on introduce how we have managed to perform very well in technology, but for the product itself, for last year, Momenta has already introduced our R6 model, which is our reinforcement learning model, and R7 model introduced this year is a world model.

Driven by continuous technological and product innovation, we provide end users with a safe, seamless experience that they truly value and rely on. In terms of delivery and execution, our execution efficiency has also been further improved.

Last year, we delivered roughly 40 vehicle types, and for the first half of this year alone, we delivered around 37 types. We are customer-centric.

Every day we think about how to deliver better customer experience, better customer service. I think about it, our senior executives think about it, and our whole team thinks about it.

Thank you.

谢旭东总，非常清楚。然后我有一个follow-up的问题，就是我们其实还是非常认可Momenta提出的世界模型基座，然后也想问一下，就是这块最新的进展如何？A quick follow-up question on me: we highly recognize the world model foundation proposed by Momenta and would like to learn more about this, its latest progress. Thank you.

这个非常好的问题啊，感谢你的问题。就我们这个R7世界模型的话已经在这个最后的交付阶段了，然后呢进展也非常好。主要是是在两个方向上取得了非常好的scaling，一个是model scaling，另外一个是data scaling。那依托于我们这个世界模型以及在model scaling和data scaling上的进展，我们R7世界模型的话陆续会给重要的客户开始交付了。目前这个R7在很多场景上都取得了我们上一代相比上一代我们的R6有3到5倍的一个提升，并且在这个成本上，我指的是在硬件成本上，我们的客户的话并不需要额外承担额外的成本。我们沿用了R6的传感器和R6的芯片就能够提升这个，通过软件算法的升级就能有3到5倍的一个提升，嗯。那下半年预计会有很多车型搭载我们的这个R7世界模型，包括了这个大众ID. Era 9X，还有凯迪拉克的XT5，还有这个奔驰的GLC、GLE，还有宝马的iX3。

Thank you for your good question. We have made a lot of progress in our R7 model, which has been centered around, first of all, model scaling, and secondly, data scaling.

Leveraging advances in this model and data scaling, we have started delivering the R7 model to key customers. R7 has achieved a 3 to 5 times improvement over R6 across many scenarios.

From a cost perspective, R7 imposes no additional burden on our customers as its sensor configuration and chip configuration inherits that of the R6. We expect that a broad lineup of vehicles to adopt the R7 model in the second half of the year, including, but not limited to, the Volkswagen ID.

Era 9X, Cadillac XT5, Mercedes-Benz GLE, and BMW iX3, etc. Thank you.

谢谢各位领导。我的问题呢主要是关于我们出海的一个进度啊，包括出海啊的出口啊海外车型的落地速度。目前呢我们国家的海外业务进一步发力放量的话主要瓶颈是客户呢还是产品认证，还是本地化能力，还是成本？请管理层稍微介绍一下。Could management provide update about the current progress in the global expansion, specifically regarding the deployment of the both export vehicle and overseas made models? Additionally, we're looking for further scale of the overseas business.

What do you see the key bottleneck right now? Would it be the customer onboarding, or product certification, localization capability, or just the product competitiveness?

Thank you.

谢谢Bin总的问题，这个题可以由我们的孙环总来回答。 好的，感谢Bin总的问题。对，关于出海这个问题呢，我们目前观察到就是海外高阶智驾由于法律法规的一些限制，进展确实是比国内要缓慢一些。但我们也看到啊，例如这个欧洲DCAS法规也要开始松动了，预计在27年会开放城市NOA。城市NOA在海外市场开放后，包括说像特斯拉和中国的主机厂，他们会释放这个更高阶的这个智驾功能。海外市场的这个城市NOA的竞争啊也会开始加剧。目前Momenta在海外市场已经做了一些布局。一方面我们已经有多款车型出口海外市场了，截至目前我们搭载Momenta方案的这个车型已经覆盖了10个国家和地区。另外一方面呢，因为特斯拉和中国主机厂进入海外市场，海外的主机厂对于城市NOA的这个需求啊也会愈发地紧迫。Momenta因为和海外这个客户啊有长期的这个合作关系，以及我们在国内海外有大量的线下的数据的沉淀，这样也可以帮助我们的算法很快地适配到这个海外的道路的环境，加速我们和海外客户的这个在当地市场的车型的这个落地量产。

Thank you for question. I will address it through three points.

The progress of advanced intelligent driving in overseas market has historically been slower than that in China, primarily due to stricter legal and regulatory frameworks. However, we are beginning to see positive shifts.

For instance, regulations such as DCAS, Driver Control Assistance Systems, in Europe are starting to relax, and we anticipate that the regulatory environment will open up for urban NOA around in 2027. Second, once urban NOA is legally permitted in these international markets, both Tesla and Chinese OEMs are planning to roll out their NOA features.

Consequently, we expect the competition for urban NOA in overseas markets to intensify significantly. Third, Momenta has already made substantial strategic progress globally to capture this opportunity.

We already have multiple vehicle models exported globally. To date, vehicles equipped with Momenta solutions have established a footprint across 10 overseas countries and regions.

Secondly, as Tesla and Chinese OEMs expand into international markets, traditional global OEMs are feeling an increasingly urgent need for urban NOA capabilities. Leveraging our longstanding partnerships with these global customers, along with our massive accumulation of real-world driving data both domestically and internationally, Momenta is uniquely positioned.

We can rapidly adapt our algorithms to global environments, thereby accelerating the mass production and deployment of localized models for global OEMs in their respective markets. Thank you.

哎，首先恭喜那个Momenta上半年出色的业绩。我有两个问题，第一个问题是目前公司累计的这个定点已经达到219个，然后还有114个在手。随着这个OEM车型迭代和推新的频率越来越快，公司对项目的筛选标准是否有变化？以及目前的资源配置来看，公司认为合理的手在手项目规模大概是什么水平？First, congratulations to Momenta on your outstanding first half results. I have two questions.

First one is on project pipeline. You have 219 vehicle nominations, including 114 yet to enter production.

As OEMs accelerate model refresh cycle, are your criteria for selecting new projects also changing? Given your current resource space, what would you consider a reasonable level of outstanding projects that you can support?

Thank you.

好的，谢谢你的问题。关于这个目前的这个pipeline呢，应该说是比较充足的。刚才无论是旭东还是环都介绍了我们在拿项目的时候的一些精神和原则吧，然后也是比较充足的。然后从criteria的角度呢，我们是遵循大数定律，就是我们希望能够有更高的市占率来避免需要去预测某一款车型的销量是否波动，因为连OEM自己都无法预测，所以拿到更多的市占率就意味着稳健的持续的增长。然后特别的也想分享一下就是说一个盈利性和现金流方面的这个成绩吧。就是刚才旭东和我也都有介绍过，就是这半年的亏损在大幅地收窄，然后经营性净现金流也有大幅的改善，这个是我们很早就预料之内的这个事情啊。我们也一直有强调说自动驾驶天然就持续烧钱这个观点是一个误区。作为一家创业公司呢，能把自己的盈利节奏和经营性现金流管理好是对客户最大的负责。不管你拿什么pipeline，客户一定都希望你能够快速地交付，然后迭代好产品，所以有好的盈利和经营性现金流才有底气持续果断地投入研发迭代产品，而不是依赖投资人给的现金流。Momenta在判断商业机会是否能规模化、决定不同市场的优先级以及找到最佳timing上一直有着比较强的战略能力，这个从Day 1就避免了把钱花错地方。第二呢就是我们在pipeline上要求团队持续地能够服务高门槛高标准的客户，能够靠不断提升产品体验来规避价格战，然后同时我们也健康地把应收周转天数控制在105天左右。第三呢就是也是想强调一下公司的主线文化，因为我们的客户、股东和员工都非常欣赏和受益于公司的系统建设能力。建主线、复用主线、工具化、自动化，这个是我们每一天的习惯。然后大量的pipeline就要求更高效的交付，我们的交付能力呢从400人忙两年一个车型进化到数十人三个月，依靠的就是这套能力。然后关于这个pipeline、关于财务表现、盈利性等等，我觉得都是Momenta持续增长的核心目标吧，但是我们不会说把快速的盈利当成一个短期的目标，会持续地加大投入、高效投入。

Thank you, Junming, for your good question. First, to address your question regarding pipeline, I am very confident to reiterate that from the statements that you have previously heard from Xudong and Huan, we have a very strong and abundant pipeline on the road.

What is worth mentioning is that the pipeline and orders of Momenta follows the law of big numbers, which means that we do not solely rely on a single type of car models, rather than given the dominance and increasing market share that we have, we cover a diversified portfolio of different car types and car portfolio of car type portfolios, therefore enable us to weather through different volatility in the very volatile car auto markets, and therefore sustain our stable growth down the road. Actually, the sharp narrowing of our losses and improved operating cash flow were very well within our expectations so far.

We always argue that autonomous driving inherently burns cash is a total misconception. For a startup, managing cash flow and a path to profitability is our ultimate responsibility to our customers.

It gave us the independence to decisively invest in R&D and iterate products without constantly relying on investor funding. This comes down to three things: first, our strategic judgment.

From Day 1, Momenta has been highly capable at sizing market, setting priorities, and timing our entries. This ensured we didn't burn money in the wrong places at the wrong time.

Second, our pricing power and financial discipline. We serve top-tier customers by constantly elevating the product experience, which helps us bypass price wars.

Also, we maintain a very healthy accounts receivable days of around 105 days. Third, our mainline culture, which is deeply rooted in systematic capacity building.

Building, reusing, tooling, and automating mainline is our daily habit. This capability allowed us to shrink vehicle delivery efforts from 400 people over two years down to just dozens of people in three months.

This is a scalable software mindset, not a project mindset. Today, the culture has spread across every department, and our rule is very simple: do it the first time, get a result, then build a system; but to repeat it, build the system first to guarantee the result.

Finally, to be clear, short-term profitability is not our immediate goal, and we will never compromise our tech leadership for quick financial gains, but rather than high-efficiency, aggressive investments will also continue to come.

Thank you for your detailed answer. My second question was actually on cash flow and profitability, and you've already addressed it.

I have no further questions. Thank you.

喂，好的。感谢。我是中信的科技分析师连轶希。那么想请教一个问题。前段时间公司也组织了RoboWeek活动，也分享了L4的业务进展。那我们也去坐了Momenta的robotaxi和robovan，产品体验也非常的好。所以想给旭东总提问啊，您如何规划Momenta的L4业务进程？Thanks for week and had opportunity to test drive the latest robotaxi and the robovan product recently. The experience was excellent.

My question for Xudong is: how will management develop Momenta's L4 business going forward? Thank you.

这个问题由旭东来回答一下，嗯。好的。

好的，这个非常感谢你的问题啊，非常好的问题。就是之前我记得有一次也交流过类似的问题啊，就说这个自动驾驶的话，一开始就可能有两种做法。有的人的话就是假设我们最终的目标是登月的话，有一种路径的话就说珠穆朗玛峰离月亮比较近，那我就一开始去登珠峰。但我们的做法的话更像是说登月的本质到底是什么，然后解决这个本质的问题的话需要做什么。那需要造火箭，那我们就来造火箭。所以的话就是我们选择的路径是两条腿战略，一条腿是量产自动驾驶，另外一个是完全无人驾驶。那个量产自动驾驶提供数据流，就相当于在给我们造火箭。登月的话是我们实现大规模L4的终局。那我们没有一开始选择去登珠峰啊，是一开始选择去造火箭。那这个的话一开始可能会更难，但是呢经过了很长一段时间的积累，然后我们积累了我们的数据飞轮，积累了我们的海量数据，最终形成了我们的护城河，对。那这个是我我们的这个整个大的战略或者整个大的思路啊，做做一个类比。那我们再来看这个Robo的整个的市场。那我们的判断的话，整个Robo的市场还是非常大的。5年后看啊，就比如说去到了30年或者31年的时候，整个robotaxi容量的话应该能到100万到150万台。如果每每年每单车的毛利润可以做到1万美金的话，还是非常大的一个市场。另外robovan的话，我们预估可能5年后的这个容量回到全球范围来看的话可能会到300万。那每年即使单车毛利润保守估计的话2000美金，加在一起的话这个robotaxi和robovan的话总的体量也会有200亿美金的一个体量。所以呢我们瞄准的是真正能够规模化的Robo的领域啊，不去这个小规模市场去竞争，嗯。然后呢就是要做到这一点，那我们觉得最重要的大概就是把握好节奏，持续地创新。然后并且在执行层面上的话有两个关键点要做好，那第一个关键点的话就是安全。那无人驾驶追求的是十万中无一的安全标准，那我们的目标的话就是通过每年10倍每年10倍的提升，那最终能够至少实现10倍于人类的安全性。10倍于人类的安全性才能够实现规模化的Robo，不管是robovan也好还是robotaxi也好。另外一件事情的话就是做好跟整个监管的配合，不管事中国的还是海外的这个监管。我们相信只有好的监管、严格的监管，并且跟监管有很好的合作，才能够实现良币驱逐劣币，对整个的行业的发展的话会带来很大的促进作用，嗯。那总结来看，从最终的实际结果来看的话，我们凭借着这个量产项目的沉淀，然后呢以及我们这个R7世界模型的非常强大的泛化能力，那我们在在国内也包括了海外，海外的话是德国的全域，还有在UAE的阿布扎比，我们都拿到了L4的牌照，对。

Thank you very much for the question. I once used an analogy during another event for the different paths for building up autonomous driving from Day 1.

There are basically two types of passes. Many people set out with an ambitious goal to land on the moon.

Then they figure out that since Mount Everest is the closest point to the moon, they decide to climb it. Yet to reach the moon, what you really need to do is to build a rocket.

From my perspective, delivering L2 mass-produced solutions is exactly how we could build that rocket rather than going to the moon directly. Landing on the moon represents our end goal of large-scale L4 driverless deployment.

By contrast, climbing Mount Everest is analogous to building a dedicated fleet, a path that will prove difficult to sustain for commercialization and technical iteration over the long run. For a long time, we have focused on accumulating massive volumes of real data and activating our moat, the data flywheel.

Today, our rocket is ready, and we can accelerate the progress towards landing on the moon. As a result, you will begin to see tangible advances in our L4 initiatives.

Five years from now, in the global markets, we project a robotaxi market volume of 1.5 million units with an annual gross margin of US$10,000 per vehicle. For robovan, we assume a market volume of 3 million units with an annual gross margin of US$2,000 per vehicle.

Combined, this represents a total addressable market exceeding US$20 billion. Accordingly, we are targeting truly scalable segments rather than competing in small, low-barrier markets.

That said, thorough preparation is essential throughout this journey. Scaling up must follow a measured pace, and any groundbreaking innovation requires step-by-step validation.

At the tactical execution level, our overall advancement framework revolves around two core pillars: one is safety, and the other one is regulation. On the safety side, for driverless vehicles, we aim for an exceptionally high safety standard of near-to-zero failure performance.

We aim to improve our safety performance by almost 10 times every year, so that in order to reach the real scalable Robo, it has to be at least 10 times much safer than human driving levels. For regulation, in my view, stringent regulation and controlled rollout speeds are positive developments.

High standards create high entry barriers, which, in turn, foster a market environment where high-quality players crowd out those inferior ones. We embrace regulation, and we maintain deliberate control over our rollout speed.

As a result, drawing on overseas safety compliance expertise cultivated from our L2 mass production projects and the strong generalization capability of our models, and we have secured nationwide L4 testing licenses in Germany as well as a permit in Abu Dhabi. Thank you.

好的，各位领导好，感谢给我这个提问的机会啊。那我的问题是关于这个价格战还有价格压力的。因为我们看到国内车型目前其实价格非常地有激烈的价格战，然后但是我们看到今年上半年咱们的这个ASP还是有一个环比的增长。想请公司帮我们展望一下之后的价格战会不会对咱们的单车软件许可定价还有整体的毛利率造成一定的压力呢？然后公司将会怎么样捍卫自身的一个溢价能力？So my question is: given the intense price competition taking place among the domestic OEMs, will this trend exert downward pressure on your licensing ASP and overall gross margin? How would you defend your premium pricing power going forward?

Thank you. 这个问题请环来回答，嗯。

好的，感谢您的提问，对。这是一个非常核心的问题，我来分三个点来回答。第一个是关于当前单车的均价ASP和这个定价权。对，就国内市场的这个价格战确实是给供应链带来了不小的压力。未来呢大家可能会看到我们的这个ASP单车ASP会有一定的调整，但不是说我们这个丧失了这个定价权，而是基于量产的激增的这个阶梯做定价，做这个volume discount，把这个规模化带来的边际成本的降低与主机厂去共享，是一个健康的商业共赢的模式。这是第一点。第二呢我想特别强调是Momenta绝对不会为了拿单盲目地去跟风做这个底线的降价。智驾呢是关乎生命安全的。便宜但是不安全的這個系统会直接反噬车企的品牌。我们的溢价能力呢是由真实的业绩支撑。在一些合作车型上，价值1万多的这个高阶智驾选装包选装率能够到80%以上。这个证明呢只要是体验跨越了信任的阈值，消费者还是非常愿意为极致的安全和体验去买单的。这个是第二点。第三呢我们想这分享一下未来的收费模式。我们现在也正在与客户积极去探讨像C端收费的订阅模式。行业的一些先行指标已经证明了这条路的可行性。随着产品体验的跨越式提升，特斯拉，举例子，特斯拉V14版本在今年上半年的这个订阅率已经飙升到50%左右。说明这个颠覆式的体验是可以带来用户持续付费的意愿的。同时呢我们也正在向L3迈进。随着奥迪E7X一些合作车型的推出，以及明年相关法规的落地，为智驾的商业模式可以从单车的license像订阅模式提供出一些基础。这也给将来单车收入质的飞跃带来一些新的可能性。对，我分享如上。

Thank you for the question, Joey, and I will address this in three points. Regarding our current ASP and pricing power, we have seen the intense price competition in the domestic market, and it has indeed placed pressure across the supply chain.

Moving forward, you might see some adjustments in our ASP, but I want to be clear that this does not indicate a loss of pricing power of Momenta. Rather, it reflects our tiered pricing strategy and tiered pricing structure, or what we call volume discounts, as our mass production scales exponentially.

Sharing the marginal cost reductions derived from economies of scale with our OEM partners is a healthy, win-win commercial strategy. Secondly, I want to reiterate that Momenta will never chase deals by blindly following the market and cutting prices without bottom lines.

Autonomous driving is fundamentally a matter of life safety. A system that is cheap but unsafe will inevitably backfire and severely damage an OEM's brand reputation.

Our premium pricing power is backed by tangible performance data. On certain models we partner with, the advanced ADAS optional package priced at over RMB 10,000 continues to achieve an impressive take rate of over 80%.

This clearly proves that once the product experience crosses the threshold of user trust, consumers are more than willing to pay a premium for ultimate safety and exceptional driving experience. Thirdly, regarding our future monetization models, we are actively exploring subscription models targeting end customers with our OEM clients, and leading industry indicators have already validated this path.

Driven by a generational leap in product experience, the subscription rate for Tesla's V14 version surged to around 50% in the first half of this year. This demonstrates that delivering a disruptive experience is the absolute key to unlocking sustained consumer willingness to pay.

At the same time, we are relentlessly advancing towards L3 autonomy. With the upcoming launch of partnered models like the Audi E7X coupled with the anticipated rollout of relevant regulations next year, intelligent driving's business model is transitioning from licenses to subscription, enabling a step change in the per-vehicle revenue.

Thank you.