Youdao, Inc.

Good day and welcome to UDAO's first quarter 2026 earnings conference call. Today's conference is being recorded. At this time, I would like to turn the conference over to Mr. Jeffrey Wong, investor relations director of UDAO. Please go ahead.

Thank you, operator. Please note that the discussion today will contain forward-looking statements related to the future performance of the company, which I intend to qualify for the safe harbor from liability. as established by the U.S. Climate Security Litigation Reform Act. Such statements are not guarantees of the future performance and are subject to certain risks and uncertainties, exceptions, and other factors. Some of these risks are beyond the company's control and could cause actual results to differ materially from those mentioned in today's press release and this discussion. A general discussion of the risk factors that could affect UDOT's business and its financial results is included in certain company filings with the U.S. Securities and Exchange Commission. The company does not undertake any obligation to update this forward-looking information, except as required by law. During today's call, management will also discuss certain non-GAAP financial measures for comparison purposes only. For the definitions of non-GAAP financial measures, and Reconciliation of Gap to Non-Gap Financial Results, please see the 2026 Fourth Quarter Financial Results news release, issued earlier today. As a reminder, this conference is being recorded. A webcast replay of this conference call will also be available on UDOT's corporate website at ir.edu.com. Joining us today on the call from UDOT's Senior Management, Dr. Feng Zhou, our Chief Executive Officer, Mr. Lei Jing, our President, Mr. Peng Xu, our Senior VP, and Mr. Wayne Lee, our VP of Finance. I will now turn the call over to Dr. Zhou to review some of our recent highlights and strategic directions.

Thank you, Jeffrey, and thank you all for participating in today's call. Before we begin, I would like to remind everyone that all numbers are denominated in Myanmar B, unless otherwise stated. Yodal delivered a solid start in 2026, and net revenues were RMB 1.3 billion, up 3.8% year over year. Operating profit was RMB 57.5 million, marking our seventh consecutive quarters of operating profitability. While operating margin improved sequentially by 0.5 percentage points to 4.3%, year-over-year operating profit declined 44.7%, primarily reflecting our proactive investments in core strategic initiatives, including AI, as well as a high comparison base from the restructuring of learning services in the same period last year. Net operating cash outflow narrowed significantly by 60%. 3.6% year-over-year to RMB 93.1 million. Supported by successful AI product launches in Q1 and a strong pipeline ahead, we remain focused on delivering full-year improvements in profitability and cash flow in 2026. We continue to advance the AI technologies that drive our business growth. Just this week, we released Confucius 4, our open-source learning large language model. Its most important new feature is multi-model inputs, enabling industry-leading capabilities in solving and teaching K-12 subjects that require visual understanding, such as geometry. We also released EmotiVoice 2, our open-source, high-fidelity AI text-to-speech model, with advanced features including cross-lingual voice cloning. In addition, we launched Confucius Translation 4, our latest AI translation model, delivering industry-leading performance across 40 languages. With that, let me walk through the performance of each business line during this quarter. The revenues from the learning services segment were IMB $627.5 million, up 4.2% year-over-year. Youdao Lingshi maintained strong momentum with growth, growing by over 20% year-over-year in Q1. Product-improved innovation remained a key driver of this growth. Powered by our proprietary Confucius LRM, we launched English AI essay grading this quarter, further enhancing our differentiated AI-powered learning experience. The feature provides personalized, high-quality feedback reports in approximately one minute, improving learning outcomes for students while increasing operational efficiency for teaching assistants. Early adoption has been encouraging, with approximately 10,000 essays graded by AI to date. Our programming courses maintain strong momentum in the first quarter, with growth billions growing by over 20% year-over-year, supported by ongoing product enhancements and the strategic expansion of our user acquisition channels. In addition to business growth, our students continue to achieve outstanding results in top tier competitions, winning one gold, one silver, and two bronze medals at the 43rd National Olympiad in Informatics Winter Camp. In addition, one student was selected for the Chinese national team and won a gold medal at the 2026 International Winter AI Olympiad. These results underscore the depth of our teaching capabilities and the strength of our programming education ecosystem. Within learning services, our AI-driven subscription services continued their robust growth trajectory. In the first quarter, total sales exceeded RMB 100 million, representing year-over-year growth of over 70%. We also continue to iterate our proprietary Confucius LLM with a focus on high utility learning and productivity scenarios, further enriching our AI agent mix. This quarter we launched two new AI agent products. The first is Lobster AI, a personal AI desktop assistant designed for productivity and secure deployment. Lobster AI enables enterprises and individual users to deploy powerful, customized AI agents while maintaining data privacy. Since its open-source release, it has gained strong traction among the global developer community and surpassed 5,000 stars on GitHub. The second is Yudao Baopu, an AI-native knowledge base designed for complex knowledge synthesis. Powered by a dynamic reasoning architecture, Yudao Baoku can decompose complex queries, perform multi-round verification, and provide precise citations. It helps users transform large volumes of materials into structured multi-model outputs, including chart-rich presentations and mind maps, helping users improve knowledge work productivity. In addition to launching new AI-native products, we continued to upgrade our core applications. The AI simultaneous interpretation feature in UDAW dictionary and UDAW desktop translation saw user engagement increase by over 100% year-over-year. This growth was driven by two key upgrades. First, the deployment of Confucius 3 translation LLM. which reduces the latency by approximately 50%. And second, the evolution of the feature from a translation tool into a more autonomous AI agent, enabling more natural interactions and deeper contextual understanding. Our technical capabilities were further validated at the 14th National Interpretation Contest, where Yudao won championships in eight of the 16 AI-tracked language categories, demonstrating the strength of our AI translation systems. In the first quarter, our online marketing services maintained strong momentum, generating RMB $611.1 million in net revenues, up 20.9% year-over-year. Growth was primarily driven by increased demand for performance-based advertising, supported by our continued investments in AI technology. Gaming remained a core advertising vertical and continued to demonstrate resilience and steady growth. At the same time, we captured emerging opportunities in fast-growing sectors, particularly AI applications and short-form dramas. By integrating advanced AI capabilities with vertical-specific marketing scenarios, we achieved over 50% year-over-year advertising revenue growth in each of these emerging sectors. On the product front, we continue to leverage our vertical advertising LRM to enhance product and service quality. In Q1, we launched an upgraded version of Infinise, our one-stop AI platform for KOL marketing. The upgrade focused on two key areas. First, workflow synergy. Symfony now enables brands to manage the full collaboration lifecycle, from top-tier influencers to KOCs, through a streamlined online workflow that significantly shortens collaboration cycles. Second, AI-powered self-service. The platform automates influencer recommendations and content creation, lowering entry barriers while improving execution efficiency. Since the upgrade, Infonese has received positive feedback from KOLs and marketers. To date, nearly 60,000 influencers globally have registered on the platform, providing a solid foundation for future expansion. Gross margin for online marketing services was 29.6% in the first quarter, largely stable year-over-year and up 1.8 percentage points sequentially. marking the second consecutive quarter of sequential improvement. Turning to our smart devices segment, net revenues were RMB 109.4 million in the first quarter, down 42.6% year-over-year. We continue to exercise operational discipline in this segment, prioritizing SKU Health, inventory management, and profitability over near-term volume growth. At the same time, our products continue to receive strong external recognition. This quarter, the Yudao Tutoring Pen was honored as the best educational hardware solution at the 2026 EdTech Awards and was the only Chinese product to receive this distinction. In addition, Yudao Space X was recognized as an AI benchmark by Wall Street CM, reflecting continued recognition of our AI capabilities and educational value. Looking ahead, we remain firmly committed to our AI-native strategy by continuously refining our vertical LLMs for learning and advertising and expanding our AI agent matrix. We are enhancing how users learn, work, and market while creating new opportunities for sustainable growth. As we continue to improve user experience, we remain focused on driving continued improvements in profitability and cash flow in 2026. With that, I'll hand the call over to Supong for a deeper dive into our financial results. Thank you.

Thank you, Dr. Zhou, and hello, everyone. Today, I will be presenting some financial highlights from the first quarter of 2026. We encourage you to read through our press release earlier today for further details. For the first quarter, total revenue, RMB 1.3 billion, or US dollar, 195.4 million, representing a 3.8% increase from the same period of 2025. Now, revenue from our learning services, RMB 627.5 million, or US dollar, 91 million, representing a 4.2% increase from the same period of 2025. Net revenue from our smart devices was RMB 109.4 million, or US dollar, 15.9 million, representing a 42.6% decrease from the same period of 2025, primarily due to the decline in demand for smart devices in first quarter of 2026. Net revenue from our online marketing services were RMB 611.1 million, or US dollar, 88.6 million, representing a 20.9% increase from the same period of 2025. The year-over-year increase was mainly attributable to the increased demands for performance-based advertisements through the third-party's internet properties, which was driven by our continued investment in AI technology. For the first quarter, our total gross profit was RMB $602.3 million, for the U.S. dollar, $87.3 million, largely slight compared with the same period of 2025. Gross margin for learning services was 60.2% for the first quarter of 2006, compared with 59.8% for the same period of 2005. Gross margin for smart devices was 39.9% for the first quarter of 2006, compared with 52.3% for the same period of 2005. Gross margin for online marketing services was 29.6% for the first quarter of 2006, compared with 30.5% for the same period of 2015. For the first quarter, our total operating expense was RMB $544.8 million, or US dollar $79 million, compared with RMB $510.2 million for the same period of last year. Looking at our expense in more detail, sales and marketing expense for the first quarter of 2016 was RMB $382.2 million, compared with RMB $357.6 million in the first quarter of 2015. Research and development expense for the first quarter of 2026 were RMB $115.4 million, remaining stable with the same period of 2025. Our operating income margin was 4.3% in the first quarter of 2026, compared with 8% for the same period of last year. For the first quarter of 2026, our net income attributable to the ordinary shareholders were RMB $38.6 million, or US dollar $5.6 million. compared with RMB 76.7 million for the same period of last year. Non-GAAP net income attributable to the ordinary shareholder for the first quarter was RMB 44.9 million, or US dollar 6.5 million, compared with RMB 81.7 million for the same period of last year. Basic and diluted net income per ADS attributable to the ordinary shareholder for the first quarter of 2026, or RMB 0.33, or US dollar 0.05 and RMB 0.32 or US dollar 0.05 respectively. Non-cash basic and diluted net income per EDS attributable to the ordinary shareholder for the first quarter was RMB 0.38 or US dollar 0.06 and RMB 0.37 or US dollar 0.05 respectively. For net cash using operating activity was RMB 93.1 million. All-US dollar, 13.5 million for the first quarter. Looking at our balance sheet, as of March 31st, 2016, our contract liability, which mainly consists of deferred revenue generated from our lending services, were RMB 667 million. All-US dollar, 96.7 million, compared with RMB 847.7 million as of December 31st, 2025. At the end of the period, our cash, cash equivalents Current and non-current restricted cash and short-term investment totaled $515.2 million, with your starter $74.7 million. This concludes our prepared remarks. Thank you for your attention. We would now like to open the call for questions. Operator, please go ahead.

Thank you. We will now begin the question and answer session. To ask a question, you may press star then 1 on your telephone keypad. If your question has already been addressed and you'd like to remove yourself from queue, please press star then 2. Once again, that's star than one if you have a question. Today's first question comes from Brian Gong at Citigroup. Please go ahead.

Yes, thanks management for taking my question. And congratulations on decent results. So my question is about our AI. So we have noticed that Youdao launched Lobster AI and Youdao Baoku in the fourth quarter. Could management share the strategy regarding your AI applications? Thank you.

Thank you, Brian. AI applications are clearly gaining momentum in 2026, driven by the explosive growth of both AI chat and AI coding in recent months. So our focus is on capturing this opportunity in the areas that we have strong capabilities in. Education, productivity, and advertising. So we are approaching this opportunity in AI from several dimensions. The first dimension is models and algorithms. So it is increasingly clear that beyond foundational models, there are significant opportunities in not pre-training but post-training, fine-tuning, reinforcement learning, and development of vertical and specialized purpose-built models. So this is where we are focused at. So our goal is to, in the model area, is to basically build specialized models that deliver unique intelligence for our users and customers. And this has already become one of our key differentiators in education and also in advertising. For example, we recently released Confucius 4, our open source education RRM. So one of its most important feature is vision input. that has been specially trained for education scenarios. So what this does is this enables strong capabilities in solving and explaining K-12 problems that requires a vision input. For example, the geometry questions, geometry problems. So this direct supports our K-12 learning products as math and geometry and all these different visualized problems are really, really important for students. So similarly, we recently released the Confucius translation for our latest translation model. It supports real-time voice translation across 40 languages and operates at less than one-tenth of the cost of general purpose large language model. So making it highly suitable for large scale commercial deployment of these really, really popular kind of live translation and voice interpretation services, which has become more and more popular. So the second dimension is applications. So Lobster AI and Yodobo, who are both exciting new products, So compared with our early AI products, these two are a little bit special. They are designed to be more intelligent, more agentic, and more capable of handling long-running, complex, high-value tasks for our users. So Lobster AI is a personal AI desktop system that can support a wide range of use cases, creative exploration to productivity in professional settings. So, Field Art Baoku, in contrast, is a more specialized tool that focuses on deep research and personal knowledge management. So both products have significant long-term potential. So going forward, we will continue to upgrade our AI applications to make them more intelligent, grow their user base and explore monetization opportunities. So beyond these two new products, our existing applications also continue to perform well. So AI simultaneous interpretation of UDAO dictionary and translation maintain the strong growth in Q1. So also recently we added voice-to-voice live translation feature, so expanding beyond the existing voice-to-text live translation. So sales of AI simultaneous interpretation grew by over 100% year-over-year for the second consecutive quarter in Q1. Another one of our app is Scholar AI or 学术茶. That's also an AI agent for, it's specifically for academic integrity. So colleges, students and researchers can use it to identify potential signs of AI generated content in academic papers and research manuscripts. So with the rapid growth of AI capabilities. So academic integrity in this setting has become increasingly important. So in Q1, Scholar AI achieved a pretty remarkable sales growth of over 200% year over year. So the third and last dimension of how we kind of use AI is making Yudao ourselves AI native. So this is equally important. In the AI area, companies need to become AI native internally, not just launch AI products externally. So this requires continuous iteration across our workflows, systems, and organizational practices. So for example, deploying AI coding internally has recently become a priority for us. We believe... it can significantly improve our engineering productivity as models have really advanced. So this transformation has accelerated meaningfully since the end of last year. So in our education teams, The AI essay grading feature we discussed in our prepared remarks is another example of how we are transforming our team's work, our tutors in this case. So we are also working on multiple projects to enable our internal IT systems for education businesses. So finally, we recently released ThinkFlow. an aggregation platform for AI inference services. So it is an AI infrastructure product based on capabilities we first developed and used internally. So this I think is a good example that reflects our broader approach. So we built AI capabilities for our own operations, validate them in real business scenarios, and then extend them into products and services where uh they make sense uh in uh other other people other companies settings so overall ai is a core to our strategy and our next stage of growth so by advancing specialized models release ai native applications and also transform our work internally with ai we are strengthening our competitive position in education, productivity, and in advertising. So we're also creating new opportunities for sustainable revenue growth, profitability, and cash flow improvements. Yeah, I hope that answers your question. Thank you.

Thanks. That's very clear. Thank you very much.

Thank you. And our next question today comes from Lu Ping Zhao with CICC. Please go ahead.

Good evening, Dr. Zhou and Su Zong. Thanks for taking my questions. I'm curious about the retention for Youdao Lingshi. Could management share some colors on the recent updates? Thank you.

Thank you, Brenda. I'll handle the question first if anyone has more comments. Before we talk about recent retention performance, I want to emphasized from the mid-term to long-term perspective about the top-level policy design has already unlocked an expansive growth runway for the Udao Lingshi first. According to the education powerhouse construction plan, and the 2026 government work report, there is a clear mandate to accelerate the expansion of high school educational resources. Furthermore, during the 15th five-year plan period, it's expected to add over 2 million new high school seats. That has been publicly released recently. And this capacity expansion will trigger the structural growth in high school educational demands. As a pioneer deeply rooted in these sectors, Udall Insurance is uniquely positioned to be a primary benefit of this policy-driven skill dividend. In the first quarter, we launched the English AI SA grading features. In the immediate market, a claim led to an over 20% year-over-year increase from the gross billing, serving as a powerful validation of our product's efficiency and market competitiveness. Then let us talk about recent retention activities. We have seen a very strong momentum with the retention rate exceeding 75%, continuing its upward year-over-year trajectory. This high level of the retention is the testament to the users' recognizations of our AI interactive learning formats and high-quality services. It also solidifies the foundations for the growth in the Q2 and through the full years. Looking ahead, we will continue to leverage our Confucius large-language model to deepen our footprint in the differentiated AI interactive learning format. We are committed to expanding the AI application across the entire learning lifecycle, from diagnostic assessments and personalized learning paths to knowledge expansion and college entrance consultant services. Our goal is to bridge the gap between the technology and accessibility, bring the efficiency of the AI-driven learning to more users nationwide. I hope that answers your question. Thank you, Brenda.

Thanks, Zhuzong. That's helpful.

Thank you. And our next question today comes from Thomas Chong at Jefferies. Please go ahead.

Hi. Good evening. Thanks, management, for taking my questions. Could management provide an outlook for the advertising business in Q2? Thank you.

Hi, this is Junlei. The rapid ascent of advertising business in recent years is at its core, driven by our AI evolution. AI agents like iMagibox have revolutionized the AD creative efficiency, while the AI AD placement optimizer have significantly boosted ROI through provision, profiling, and real-time bidding strategies. This has propelled our AD net revenue from RMB 1.3 billion in 2013 to RMB 2.5 billion in 2025. Consequently, advertising has jumped from 25% to 43% of our total revenue, In the first quarter of this year, the momentum remains unabated, with net revenue reaching RMB $611.1 million, a 20.9% year-over-year increase. Looking ahead, we have confidence in the long-term development prospects of ADS advice-hiding. We are empowered programmatic advertising and marketing through our vertical ADLM, achieving a high efficiency nexus between people and the business content. We will focus our strategic layout on following high potential protocols. The first one is gaming. This remains our cornerstone. By combining native deep gaming DNA with UDA's cutting-edge technology, we continue to consolidate our presence in both domestic and overseas gaming marketing. The second is the AI application. We anticipate this will be the core incremental growth driver. The global explosion of LRM and AI agents has created a surge in demand for positioning user acquisition. Our programmatic capabilities are a perfect fit for those digital products. The third is a globalizing Chinese brand. There is a robust demand for the Chinese manufacturers and the brands going global. For instance, The new energy vehicle industry is shifting from product-centric marketing to the brand-class ecosystem strategy. We intend to capture this global brand opportunity by leveraging our QL marketing paired with the massive reach of problematic AADS. The first is social apps and finance. We will leverage our expertise in data security and compound AD placement to address the high barrier marketing needs of those sectors. In addition, I would like to highlight that the advertising business has to remain the primary contributor to our operating profit. Thank you.

Thank you. Thank you. And our next question today comes from Bo Zhang at Huatai Securities. Please go ahead.

Thanks for taking my question. This is Zhang Bo from Huatai. My question is, could management elaborate on the seasonality of operating profit sets?

Thank you, Zhang Bo, for your question regarding seasonality. Your data financial metrics has Historically, Xevious pronounced seasonality. To provide a clear picture, I will address our business seasonality through three dimensions. Revenue, operating profit, and cash flow. First, seasonality of revenue. Our top line performance typically follows as stronger the second half year H2 pattern, with the third quarter usually being our annual peak. This pattern is primarily attributable to the following factors by segment. In terms of advertising, H2 is sponsored by the Q3 peak for gaming and entertainment marketing during the summer vacation, followed by Q4 Christmas holiday season, which drives both domestic and overseas marketing demand. In respect of learning services, the summer and winter break represents the intensive period for the service delivery. And the Q3 is usually the peak season. As for smart devices, they'll typically peak during the start of a new epidemic year, especially in Q3. The second. operating profit. Typically, higher revenue levels in the second half of the year drive higher operating profit. Meanwhile, quarterly operating profit is also affected by a range of other factors, including business restrictions or strategic investment in key areas. Taking 2025 as an example, 2025 was an anomaly due to our strategic restriction of learning services. We proactively focused on your ,, while scaling back investment in theme and adult classes. The revenue in H1 was largely a lagging effect from H2 2024 customer acquisitions, while sales, marketing, and R&D for H1 2025 were slashed significantly. This results in a typically high operating profit in the first half of last year. Alongside the accelerated application of core AI technology and the steady improvements in health metrics of your . We increased investment in marketing and the R&D resources. Despite robust revenue performance in H2, operating profit is relatively low in the second half of 2025. For 2026 this year, we expect the profit cadence to return to historical norms, with H2 outperforming H1. Given the factors above, we place greater emphasis on the operating profit growth over longer term. which better reflects the overall financial health of our business. Third, seasonality of cash flow. Our operating cash flow generates the net outflow in Q1 and Q3, which are pre-customer acquisition phase, and inflow in Q2 and Q4, which are major retention cycles. In Q1 this year, cash flow position continued to improve rapidly, with the net operating cash outflow narrowed by 54% year-over-year. In summary, on the premise of stable microeconomic environment, we are making good progress on delivering a rapid improvement in both operating profit and operating cash flow for the three years 2020. Thank you. Thank you.

Thank you. And that concludes the question and answer session. I'd like to turn the conference back over to management for any additional or closing comments.

Thank you once again for joining us today. If you have any further questions, please feel free to contact us at UDAW directly or reach out to Pearson Financial Communications in China or the U.S. Have a great day.

Thank you. That concludes today's conference call. We thank you all for attending today's presentation. You may now disconnect your lines and have a wonderful day.