Earnings Call Transcript - WRD Q1 2026

Operator, Operator

Good morning, and good evening, ladies and gentlemen. Thank you for standing by, and welcome to WeRide's First Quarter 2026 Earnings Conference Call. Please note that today's event is being recorded. Please note that the Chinese interpretation is for convenience purposes only. In the case of any discrepancy, management statements in their original language will prevail. Joining us today, WeRide's Founder, Chairman and CEO, Dr. Tony Han, and CFO and Head of International, Ms. Jennifer Li. Before we continue, I would like to refer you to the safe harbor statement in the company's earnings press release, which also applies to this call as today's call will include forward-looking statements including WeRide's strategy and future plans. These forward-looking statements are made under the safe harbor provision of the U.S. Private Securities Legislation Reform Act of 1995. Forward-looking statements involve inherent risks and uncertainties. The company's actual results could differ materially from those stated or implied by these forward-looking statements as a result of various important factors, and please refer to the Risk Factors section of the company's Form 20-F filed with the SEC and announcement on the website of the Hong Kong Stock Exchange for a full disclosure of these risk factors. The company does not assume any obligation to update any forward-looking statements, except as required under applicable law. Please note that all numbers stated in the 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 SEC and the Hong Kong Stock Exchange. The company's unaudited financial and operating results are released earlier today via Newswire and can be found on the company's IR website. With that, I will now begin the company's video presentation. Now I would like to pass the floor to the company's founder, Chairman and CEO, Dr. Tony Han. Please go ahead, sir.

Tony Han, Founder, Chairman and CEO

Hello, everyone, and thank you for joining us today. We started 2026 with strong momentum as a global leader in autonomous driving. In the first quarter, the total revenue of WeRide reached RMB 114 million, 58% up year-over-year. These results are driven by our accelerating global robotaxi deployment, growth across our broader autonomous driving business and the great success of our L2++ level ADAS solution. By the end of April, our global robotaxi fleet grew to around 1,300 vehicles, representing one of the largest global robotaxi fleets globally. At the same time, our Level 4 autonomous driving fleet, including robovan and the robobus, has grown to around 2,800 units. They have been deployed to or tested in 12 countries and over 40 cities worldwide. We believe the steady and significant progress this quarter reflects not only the maturity of our technology, but also the growing operational experience for L4-level fleet in multiple cities. First of all, I want to point out that we have made a major technical breakthrough through GENESIS, our closed-loop world model-based simulation engine, which boosted our model evolution pace by several folds. We can now train AI models for autonomous driving with synthesized corner cases, which may be very rare or even imaginary. With a compact AI model leveraged on our GENESIS, we have achieved four consecutive championships in the China Urban Intelligent Driving Competition. This is unprecedented. The best previous record is held by Huawei ADAS system, which got two consecutive championships. Today, many companies talk about world models and simulation platforms. But WeRide is one of very few companies that have publicly demonstrated footage of a highly realistic autonomous driving world model at scale. We have already released a GENESIS demonstration video on YouTube where viewers can check the model's ability to reconstruct and simulate visually realistic driving environment, strictly following physical laws as well. GENESIS can generate holistic virtual driving environments consistent to the desired locations, including traffic laws, pedestrians, weather conditions and complex interactions between vehicles and the surrounding objects. More importantly, we can add this environment component to simulate highly challenging scenarios such as aggressive driving behavior, heavy traffic, extreme weather, European road conditions and many other long-tail corner cases that are very difficult and extremely expensive to replicate in the real world. By leveraging synthetic data and large-scale simulation, GENESIS improved training and validation efficiency by thousands of times compared to traditional road testing. It also significantly reduces the crucial dependence on large-scale testing fleet and accelerates the deployment process in new operational regions globally. GENESIS is not just a capability. It is a unified simulation and AI training platform supporting applications from L2 ADAS to L4 robotaxi. And it is the very same backbone that makes us one of the only companies in the world to have already achieved successful commercialization of both L4 and mass-production L2 vehicles. The newly developed GENESIS now has paved the way for WeRide to the physical AI world. With GENESIS, the ADAS system developed by WeRide is comparable to the performance of FSD 14.3 in California. I personally own two Teslas and I drive with FSD 14.3 every day. I look forward to entering our global urban intelligent driving competition directly facing FSD 14.3 from Tesla, how one day we can meet in the U.S. or Europe and give our consumers a head-to-head comparison. Today, we are seeing our technology leadership translate into real global commercial scale. Let me walk you through the key operational and commercialization milestones we have achieved this quarter. First, in China, our robotaxi business continued to make strong progress in scale, operational efficiency and commercialization depth. By the end of April, our domestic robotaxi fleet expanded to about 1,000 vehicles, while our service area in Guangzhou increased by 97% compared to the end of 2025, including additional downtown districts. On the demand side, average daily orders per vehicle domestically reached 17 trips during Q1 with peak periods reaching 28 trips per vehicle. Registered robotaxi users also doubled almost every year. We believe these metrics continue to demonstrate growing user adoption and improving unit economics as robotaxi commercialization scales. In this quarter, we also continued to deepen our partner ecosystem. In April, we expanded our collaboration with Lenovo in autonomous driving computing platforms with a joint target to deploy 200,000 autonomous driving vehicles globally over the next five years. Together with Geely Farizon, we plan to deliver 2,000 upgraded, purpose-built robotaxi GXRs in 2026. We believe this partnership further strengthened our manufacturing scalability and deployment capability globally. Now turning to international markets. We also continue to see strong momentum in both new market launches and commercialization progress. In Singapore, we launched the country's first public autonomous driving service together with Grab. Since the initial deployment plan began in the second half of last year, the fleet has built a trustworthy operational track record in a highly regulated international market. In the Middle East, we have launched fully driverless commercial robotaxi operation in Dubai together with Uber and Dubai's RTA. This is the city's first fully driverless commercial robotaxi service. Meanwhile, Abu Dhabi service coverage expanded to around 70% of the city's core area. Across the Middle East, together with Uber, we remain on track to deploy at least 1,200 robotaxi across Abu Dhabi, Dubai and Riyadh by 2027, as additional ODDs are to be added. In Europe, we entered Slovakia in March, our fourth European market and continued progressing towards fully driverless commercial operation in Zurich. Overall, we continue to believe WeRide remains the most globally deployed autonomous driving company today with deployments across 12 countries and permits in 8 markets. This global footprint not only diversifies our revenue but also demonstrates operational and regulatory capabilities that are hard for followers to replicate. It also keeps us on track towards our long-term vision of deploying tens of thousands of global robotaxi globally by 2030. Beyond global robotaxi, our ADAS business is also seeing growing commercial traction. Our current version of ADAS system WRD 3.0 has been adopted by nearly 30 vehicle models, including vehicles from leading OEMs, such as GAC and Chery. In April, GAC Aion launched presales for the Aion N60, the first mass-production vehicle with WeRide's solution. Leveraging the generalization capability of GENESIS, our WRD 3.0 solution is now supporting three major chip platforms: NVIDIA DRIVE, Qualcomm Snapdragon and SiEngine StarLight. We believe this level of multi-chip compatibility is crucial because it enables faster mass production, greater cost optimization and broader OEM adoption. At the same time, we continue expanding internationally with partners, including partners bringing WeRide-powered ADAS solutions to consumers globally. Finally, for robobus, we also continue making progress globally. We are collaborating with the Geneva public transport operator, TPG, on our autonomous bus deployment. Meanwhile, we are preparing for the robobus operation with partner branches open for the third consecutive year. To summarize, the first quarter of 2026 was about substantiating the technology leadership of WeRide into commercial scalability through growing robotaxi operations, expansion into new international markets, growing deployment pipelines and continued unrivaled winning momentum in ADAS. We believe we are executing well with our global strategy, and we continue to see a clear path toward long-term growth. With that, let me turn the call over to our CFO, Jennifer Li, for a deeper view of the financial results this quarter. Thank you.

Jennifer Li, CFO and Head of International

Thank you, Tony. Hello, everyone. Before we dive into the financials, I want to highlight that all figures are in RMB, comparisons are year-over-year, unless stated otherwise. 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 SEC and Hong Kong Stock Exchange. Now let's discuss our first quarter financial performance. We delivered total revenue of RMB 114 million in the first quarter, representing an increase of 38%. Product revenue increased 116% to RMB 20 million, mainly driven by increased deployment of robotaxi and other L4 vehicles. Service revenue increased 49% to RMB 94 million. This revenue growth mirrors the solid commercial progress we made in this quarter, together with our proven track record in execution and deployment. Considering the seasonal impact of Chinese New Year and Ramadan in the Middle East, our business performance surpassed our original internal targets. Group level gross profit increased 56% to RMB 40 million in the first quarter, with the group level gross margin of approximately 35%. We maintained top-line expansion without sacrificing margin discipline and demonstrated the inherent profitability of our autonomous driving business as we further scale. This was also supported by the increased exposure to our ex-China market where we continue to see the structurally stronger margin profile as we expand into new international territories. Operating expenses were RMB 459 million with R&D expense accounting for 77% of the total operating expenses. Our operating expenses are stable in absolute RMB amount compared to the same period in 2025. To break down further, R&D expense increased by 12% to RMB 353 million in Q1 2026. Excluding share-based compensation, R&D expense grew 16% to RMB 322 million. This consistent investment in R&D underpins our technology roadmap and ensures we stay at the forefront of autonomous driving innovation. Administrative expense decreased by 33% to RMB 83 million in the first quarter. Excluding share-based compensation, administrative expense decreased by 17% to RMB 51 million. The decrease was primarily driven by lower professional services fees, mainly related to audit and legal compliance services, and partially offset by increased personnel costs for the expansion of our team as a growing business. Selling expense increased by 63% to RMB 23 million in Q1 2026. Excluding share-based compensation, selling expense increased by 81% to RMB 22 million. The increase corresponds to ongoing expansion for our business and underscores our commitment to support growth appropriately. Our net loss came at RMB 359 million in the first quarter. On a non-IFRS basis, the net loss increased slightly by 11% to RMB 326 million in the first quarter. The slight uptick was largely driven by ongoing R&D spending as we continue to invest ahead of scale in our long-term technology leadership. As of March 31, 2026, we had total capital reserves of RMB 6.22 billion, comprising RMB 6.18 billion in cash and cash equivalents and time deposits, RMB 29 million in investment and wealth management products and RMB 18 million in restricted cash. We maintained short-term bank loans of RMB 294 million to support daily operations. We have well positioned our capital base to match our cash deployment needs with headroom to support ongoing growth and strategic initiatives. Under the USD 100 million share repurchase program authorized by our Board of Directors on March 23, 2026, we have repurchased approximately 24.4 million Class A ordinary shares, including in the form of American Depositary Shares, as of market close on May 12 for a total consideration of approximately USD 61.4 million. This reflects our firm belief in the company's long-term value and growth potential. Moving forward, we proceed with confidence and a well-defined focus. By end of 2026, we remain on track to deploy 2,000 robotaxi worldwide, marking the first milestone in our journey towards hundreds of thousands by 2030. As we continue to scale, our approach is to enter new regions and cities which have proven effective and replicable. Supported by strong technological leadership, operational know-how and increasingly robust global rollout, we're prepared to lead the next phase of autonomous driving. With that, operator, we are now ready to take some questions.

Operator, Operator

The first question comes from Stanley Wan of Morgan Stanley.

Stanley Wan, Analyst, Morgan Stanley

This is Stanley from Morgan Stanley. I have two questions. First, on your robotaxi expansion, are you on track to meet your expansion roadmap in 2026? Could you update us on the latest expansion progress in China, the Middle East and the rest of the world? My second question is on WeRide's overseas business. With the company's core advantage being its possession of overseas operating licenses in 8 countries and in light of the global expansion of players like Waymo and Zoox and also the accelerating regulatory approvals, how do you view the key growth drivers, profitability path and timeline for tangible contributions from your overseas operations in 2026 to 2027?

Tony Han, Founder, Chairman and CEO

Okay. So I will take the first question, and then I will let Jennifer answer the second question about profitability. So first question is about whether we are on track with our robotaxi expansion. I think overall, we are very optimistic and confident in our global robotaxi expansion. And we have been making steady progress all the time. I will share some numbers I have already told the market: at the end of April, our global robotaxi fleet reached approximately 1,300 vehicles. That is, to my best knowledge, one of the biggest robotaxi fleets globally. In China, our robotaxi fleet has grown to approximately 1,000 vehicles with solid demand, healthy daily order volume and a growing user base. I want to share a number: our average daily orders per vehicle exceeded 17 trips during the first quarter. That is an amazing number. Our overseas robotaxi fleet has expanded to approximately 300 vehicles across multiple markets. In the Middle East, we recently launched our fully driverless commercial operation in Dubai on the Uber platform. Today, if you really want to ride a fully driverless robotaxi outside of the U.S. and China, to my best knowledge, the only way is through Uber hiring WeRide's robotaxi in Abu Dhabi or Dubai. That's our current stage. Regional tensions have created some short-term softness in utilization, but we remain very confident. More important, we remain firmly committed to the Middle East, and our long-term investment and operational presence are well received by local governments. We remain on track toward our commitment to deploy 1,200 vehicles across Dubai, Abu Dhabi and Riyadh. In Europe, we are expanding our footprint. In Switzerland, we obtained the region's first driverless commercial permit and continue progressing towards public operations in Zurich. We also recently launched our national autonomous driving program in Slovakia, making our entry into another new European market. More broadly, we believe WeRide has established a meaningful first-mover advantage through years of global expansion. Overall, I think this global operation and regulatory footprint is becoming an increasingly important differentiator as autonomous driving commercialization accelerates worldwide. So that's my answer to your questions about our global roadmap and our plan. About profitability, I'll hand over to Jennifer.

Jennifer Li, CFO and Head of International

Yes, I'll answer the second one. Stanley also mentioned the advantage of having operating licenses in 8 countries. When comparing the total number of robotaxi deployed and driverless permits held by any autonomous driving company outside China and the U.S., WeRide ranks first among these companies. We are very encouraged by the momentum of our international expansion so far. As you may recall from the last earnings release, international revenue already accounted for approximately one-third of the total group revenue last year. Our Middle East subsidiary was already profitable at the net level. Tony already talked through our key developments in some of the regions this year and we will announce some more exciting deployments in due course. This year, we expect international revenue to grow even faster and contribute an even larger share of group revenue, supported by the positive economics. I would like to elaborate a little bit more about our city selection criteria. People always ask, you guys already have presence in about 40 cities, so do you want to expand to more cities or focus on the current ones? Our strategy is to focus on scale in existing cities as well as enter additional cities selectively. We consider whether the market has the potential to support scalable and commercially attractive operations over the long term. On the monetization potential side, we pay close attention to the overall gross booking opportunities as well as gross bookings per mile because both scale and unit price matter significantly for long-term robotaxi unit economics. Besides, we focus heavily on whether there's a realistic path to scale for all the cities we are currently deploying. We do see potential for each city to deploy thousands of autonomous vehicles. Europe remains a key focus for us this year. We will share news on more deployments in more cities in Europe hopefully soon. That concludes my answer.

Operator, Operator

Next, we have Ming-Hsun Lee from Bank of America.

Ming-Hsun Lee, Analyst, Bank of America

I also have two questions. First, how does the recent report about China holding new self-driving approvals impact WeRide? Second, it's about the technology difference. There's been a consistent industry-wide debate over the LiDAR and HD Map approach that Waymo adopts versus the camera-only solution by Tesla. What is WeRide's view?

Tony Han, Founder, Chairman and CEO

I will take them one by one. First, regarding the recent pause on new self-driving approvals in China, many have noticed the incident of another company in Wuhan and the resulting reports. Our view is that this is more of a short-term regulatory adjustment rather than a structural change to the industry. From our discussions with central and local governments, and through careful investigation, WeRide has a very good safety and operational record and both central and local governments have given us strong support. We were assured that support for autonomous driving remains very strong. This pause is primarily a hold on new approvals, but current autonomous vehicles remain operational and our orders are increasing, as I mentioned earlier. As autonomous driving moves from early pilots towards larger scale commercialization, it is natural for regulators to place great focus on safety, and we think this is a responsible attitude. From our perspective, this is ultimately positive for the long-term development of the industry: companies with very good safety records should be rewarded, and those with poor safety records should be penalized. We expect regulation to become more differentiated over time based on factors such as safety performance, operational track record and technical capability. For WeRide specifically, we remain confident because we have accumulated meaningful real-world operational experience both in China and internationally. We operate and test across more than 40 cities in 12 countries. Overall, we view the regulatory pause as a necessary step for the industry and believe it will ultimately lead to a more sustainable competitive landscape. On the second question about LiDAR and HD Map versus camera-only approaches: my view is simple. I spent many years in Missouri, the 'show-me' state, so I prefer to see results rather than debates. For L4 robotaxi, Waymo has deployed thousands of robotaxi and operated safely in many U.S. cities. If HD maps and LiDAR can be used as part of the solution, why not. WeRide has long adopted a multi-sensor redundancy oriented technology path. We believe that combining camera and LiDAR builds a strong, robust autonomous driving system. This approach is broadly aligned with leading global robotaxi players like Waymo. At the same time, we are closely watching the progress of vision-only approaches, including Tesla's work. FSD 14.3 has made very good progress, and we admire Tesla's achievement. However, it is still not a fully driverless robotaxi system. In our ADAS systems, using a one-stage end-to-end world model, we achieved four consecutive championships in China, beating other ADAS solutions. We are aware of the advantages of camera-only solutions based on foundation models and world models, and we are good at it. We have also prepared extensively for complex traffic scenarios in cities like Madrid and Zurich. With our GENESIS model, we believe we can combine HD Map approaches with world-model-based vision approaches. One of WeRide's strengths is flexibility across different architectures; we can make a marriage between LiDAR-plus-camera-based systems and vision-first approaches. For reliable and safe robotaxi deployment at the current stage, HD Map remains important. Gradually, vision and world-model approaches can help maintain map freshness and handle map changes. We plan to cherry-pick the advantages of each approach and integrate them into our core system. Given our ADAS programs and our L4 deployment operational experience, we believe WeRide is among the few companies capable of doing this.

Operator, Operator

Next, we have an analyst from Huatai Securities.

Analyst, Analyst, Huatai Securities

I have two questions. First, we are seeing more and more Level 2 portfolios announcing plans to enter Level 4 market. What's your view on this? Second, can you give us more details on multi-chip platform compatibility for Level 2? How do you achieve chipset compatibility versus other approaches?

Tony Han, Founder, Chairman and CEO

Thank you for these two questions. These are technical and strategic questions, so I'll answer both. First, regarding companies with ADAS roots moving into L4, I want to emphasize that, to my best knowledge, only a very small number of companies are doing full L4 robotaxi and L2++ systems at the same time. I suggest a practical standard: if a company claims to be an L4 robotaxi company it should have a fleet of at least 100 driverless robotaxi open to public operation for more than half a year without any significant or serious accidents. For ADAS companies, the standard is mass-production deployment in vehicles. There is still a significant gap between advanced ADAS and a true L4 robotaxi system. The challenge is fundamentally about robustness, operational capability and scalability. For example, many ADAS systems target a mile-per-critical-intervention figure of around 1,000 miles, but for L4 robotaxi, the operational metric for miles per critical intervention is above one million kilometers (or the equivalent), so there is a three order-of-magnitude difference. Improving by 10x per year is an extremely formidable task. It takes many years to build reliable L4 fleets. I have a 'golden test' rule: make a driverless fleet of at least 100 vehicles, operate it for half a year, and see what happens. Ambition is good, but achieving L4 requires proven operations over time. Second, regarding multi-chip platform compatibility: our secret sauce is GENESIS. GENESIS creates AI models that can be trimmed for different chipsets and different compute complexities. So far, we have rolled out systems based on NVIDIA DRIVE platforms, Qualcomm-based systems (for example, Qualcomm 8650), and we are working on SiEngine StarLight and other platforms. With our GENESIS-based approach, we can build a world model and then support different onboard AI models across a spectrum of hardware. This gives us a competitive advantage, and to my knowledge, few other companies support such a wide spectrum of chipsets for both ADAS and global robotaxi solutions.

Operator, Operator

Next, we have Kai Xiao from CICC.

Kai Xiao, Analyst, CICC

I have a question on the multi-chip platform strategy you mentioned. Can you share why this strategy is important for your L2++ ADAS business?

Tony Han, Founder, Chairman and CEO

The main reason is that different OEMs have different hardware preferences. Some OEMs want extremely cost-effective chips, while others want very high compute platforms. They also have different supply chain requirements. A flexible multi-chip architecture allows us to support a broad range of vehicle platforms without redesigning the system each time. GENESIS helps us generate a spectrum of onboard AI models which accommodate different top requirements, such as Qualcomm-based platforms or NVIDIA DRIVE platforms. Multi-chip vendor flexibility enables faster mass production and better cost optimization. Different chip platforms offer different costs and supply chain advantages. If an ADAS company or autonomous driving software company can offer a spectrum of solutions across multiple chipsets, OEMs are more likely to work with you. We believe this flexibility helps secure more production and more vehicle types. Also, exports of Chinese automobile models are growing globally, and many overseas models currently lack a strong ADAS system. We have already secured more than a number of vehicle models to supply ADAS systems for them, so next year you may see WeRide ADAS in cars in Southeast Asia, the Middle East, and some markets in Europe. Please stay tuned.

Operator, Operator

Next, we have Xinyu Fang from UBS.

Xinyu Fang, Analyst, UBS

I have one question about the strategic balance and prioritization between China and international business. As Tony mentioned, there has been news about tightening scrutiny for autonomous driving permits domestically and WeRide has been making steady advancement in international markets with better unit economics. Could management please share your thoughts on the balance and priority of domestic and overseas operations lately?

Jennifer Li, CFO and Head of International

Xinyu, I'll take your question. We see both China and our international markets as strategically important for WeRide. In the near- to medium-term, certain international markets offer a clearer and faster path to commercially attractive robotaxi economics thanks to favorable pricing, partnerships and favorable regulation. Demonstrating sustainable profitability early on is critical for our industry because healthy cash flow enables self-sustained growth in local markets. We started building a significant international model in 2021, ahead of many peers, and since then we have gained hard-to-replicate expertise in global deployment, regulation, localization and fleet operations. Meanwhile, China remains a key long-term market due to the size of the ecosystem and infrastructure; it's our home, and we continue to maintain a very strong focus there as well. Overall, we see a powerful flywheel effect: as we scale across more cities and countries, we get more data and validation to improve performance and regulator trust, which accelerates permits, expansion and commercialization. So both are very important. We see huge potential for international revenue to grow rapidly this year, and we are on track to achieve our full-year revenue targets.

Operator, Operator

Next, we have Tianyu Lu from Citic Securities.

Tianyu Lu, Analyst, Citic Securities

I have two questions. First, what's Uber's current shareholding and how should investors think about the relationship between WeRide and Uber? Second, could you share your go-to-market model across different markets?

Jennifer Li, CFO and Head of International

Overall, Uber holds more than 5% based on their latest public filing. Uber is a strategic shareholder and a key partner for WeRide. We view this as a strong endorsement of WeRide's technology and our deployment and commercialization strategy. Operationally, we're deploying with Uber in more cities outside the U.S. than many other AV players, which reflects the depth of our relationship and our ability to execute at scale. We expect to expand to more cities with Uber this year and to remain a trusted partner as they build their global strategy. Our go-to-market approach is diversified and tailored to each market. In China, we offer our own app and channels to build awareness and operational capabilities; you can get our robotaxi on WeRide Go and through partners such as Amap and Tencent mini-program. In Southeast Asia, we partnered with Grab. In Europe, we collaborate with local platforms, operators and public transport operators, including SBB in Zurich and TPG in Geneva. This accelerates deployment and permitting. Regarding the asset-light business model for robotaxi: outside China, we already implement asset-light models in all markets. In some cases, robotaxis are purchased by platforms or local customers; in other cases, they are owned by third-party fleet owners. We have executed both models successfully. The asset-light model allows WeRide to focus on technology and operations while leveraging local capital to scale efficiently. In China, our priority is to continue improving unit economics and operational efficiency so that revenue share becomes appealing for third-party owners. Over time, China should also transition toward more asset-light structures as utilization strengthens.

Operator, Operator

Our last question comes from May Lou from HSBC.

May Lou, Analyst, HSBC

I have one question. What's the trajectory for robotaxi vehicle cost reduction?

Jennifer Li, CFO and Head of International

Cost reduction remains one of our key focus areas and we continue to see meaningful progress. There are three main drivers. First is hardware and system integration. With our latest platform and our purpose-built vehicle GXR, we are moving toward pre-integrated and standardized solutions and the BOM cost continues to drop year over year. Second is scale and supply chain optimization. As we move to a few thousand unit deployments, we achieve better cost efficiency across components and manufacturing. Third is operational efficiency. As the fleet scales and utilization increases, operational costs per vehicle decline. For example, in the last quarter we already announced that our remote safety officer ratio improved to 1:40 from previously about 1:10 or 1:20. Internationally, we are following the same trend, though not yet at 1:40, but moving in that direction. Remote operational efficiency has a very meaningful impact on per-vehicle TCO and overall unit economics. Additionally, when we enter new markets there may be upfront localization and integration costs which can temporarily increase per-vehicle cost in the early stage. As deployment scales within each market, those costs are amortized and the overall impact becomes limited and manageable. So cost reduction is a combination of hardware, supply chain, and operational efficiency.

Operator, Operator

Due to time restraints, I will conclude the call today. Thank you for your participation in today's conference. This concludes the program. You may now disconnect.