Operator: Good day, and thank you for standing by. Welcome to the Agora, Inc. Fourth Quarter and Fiscal Year 2024 Financial Results Conference Call. After the speakers' presentation, there will be a question and answer session. Please be advised that today's conference is being recorded. The company's earnings results press release, earnings presentation, SEC filings, and a replay of today's call can be found on its IR website at investor.agora.io. Joining me today are Tony Zhao, founder, chairman, and CEO, and Jingbo Wang, the company's CFO. Reconciliations between the company's GAAP and non-GAAP results can be found in its earnings press release. During this call, the company will make forward-looking statements about its future financial performance and other future events and trends. These statements are only predictions that are based on what the company believes today, and actual results may differ materially. These forward-looking statements are subject to risks, uncertainties, assumptions, and other factors that could affect the company's financial results and the performance of its business, which the company details in its filings with the SEC, including today's earnings press release and the risk factors and other information contained in the final press list relating to its initial public offering. Agora, Inc. has no obligation to update any forward-looking statements the company may make on today's call. With that, let me turn the call over to Tony. Tony, you may begin.
Tony Zhao: Alright. Thank you, operator, and welcome everyone to our earnings call. I'll first review our operating results from the past quarter. This quarter, we achieved GAAP profitability, a significant milestone for our company. This success is the result of strong revenue growth from new use cases and disciplined cost management. While we faced challenges over the past few years, from the pandemic to regulatory changes and macroeconomic pressures, our team's hard work, innovation, and focus have brought us back to profitability. As we look ahead to 2025, we remain committed to driving operational efficiency and delivering sustainable profitable growth. Total revenue in the fourth quarter was $34.5 million, up 4% year over year, excluding revenues from certain end-of-sale low-margin products, mainly driven by business expansion in certain use cases, such as live shopping and entertainment. As of the end of 2024, we had over 1,700 active customers for Agora, and close to 2,000 for Xiong Wang, excluding Ismaq, an increase of 2% and 8% respectively, compared to one year ago. Now let's turn to our business, product, and technology updates for the quarter. The generative AI revolution is advancing at an incredible pace. Over the past few months, we've seen groundbreaking developments from OpenAI's reasoning model 01 to Google's model multimodal Gemini 2.0, as well as remarkable open-source innovation like IPSec. This transformation presents extraordinary opportunities for our business. Our real-time engagement technology is uniquely positioned to enable seamless voice-based interactions between humans and AI models. By moving beyond basic text chat to multimodal conversations, we are the key technology to enable and ensure more intuitive and expressive communication experiences. It also introduces entirely new use cases where text input simply is not practical or efficient. However, many large language models today still lack robust voice interaction capabilities. For those that do support voice interaction, the experience often falls short with high latency, awkward pauses, interruptions, and sensitivity to background noises, which can lead to a frustrating user experience. To address this gap, we've supported and contributed to an open-source project in our developer community called the Transformative Expansions Network, or TEN. TEN is an abstraction framework that enables developers to rapidly build multimodal AI agents. It integrates building blocks for speech-to-text, large language models, text-to-speech, retrieval argument generation, video understanding, and many other useful modules, enabling developers to easily create real-time, low-latency AI experiences. Essentially, this framework can enable any text-based large language model to deliver a conversational voice-based experience. For example, a developer can use TEN to quickly build a voice agent combining Microsoft STT, Deep Six large language model, and speech audio text-to-speech, all in a matter of minutes. Since its launch, TEN has become the fastest-growing project of its kind, earning over 4,500 GitHub stars in just a few months. On multiple occasions, TEN topped the trending list as the number one patent project on GitHub and remained on the list for several days. Building on TEN's success, I'm excited to announce the launch of our conversational AI engine, an enterprise-grade solution that empowers developers to create interactive voice experiences with any AI model, whether it's a customer-built model deployed on-premise or a leading large language model accessed via API. Our conversational AI engine can make it talk and interact naturally with students. What sets this solution apart is its ability to integrate easily with new AI models. For instance, we integrated with Google Gemini's multimodal live API within 24 hours of its release last December. The same 24-hour integration was true for Deep Fit in January. Our solution is also compatible with many popular STT and TDS modules. But broad compatibility is not the only advantage. Our conversational AI engine elevates the conversation experience in three key areas. First, conversation dynamics. To create natural human-like interactions, the engine intelligently detects and pauses for user interactions within 350 milliseconds, ensuring conversations feel smooth and intuitive. Second, voice processing. Our selective attention algorithm enables the AI models to focus solely on the primary speaker and filter out distractions from irrelevant speakers in the background. Additionally, our advanced noise suppression and echo cancellation algorithms ensure that the AI model receives crystal-clear human input necessary for generating high-quality responses. Lastly, latency. Leveraging our global infrastructure, the engine responds to queries in as little as 500 milliseconds, dynamically handling packet loss and minimizing delays across different network conditions. We have been working closely with customers to build and test conversational AI applications across various use cases, from virtual companions and live language tutoring to smart IoT devices, virtual shopping assistants, and customer support. We believe this innovation will accelerate the adoption of conversational AI across industries and become a key driver of our future growth. Before I wrap up, I want to take a moment to thank the incredible teams at Agora and Xiong Wang. Your dedication and hard work in this fast-evolving environment have been nothing short of inspiring. Today, we are pushing the boundary of real-time AI interactions and unlocking new possibilities for human-AI communication. I'm excited about what lies ahead, and I look forward to sharing more updates with you soon. With that, let me turn things over to Jingbo, who will review our financial results.
Jingbo Wang: Thank you, Tony. Hello, everyone. Let me start by first reviewing the financial results for the fourth quarter of 2024, and then we will discuss the outlook for the first quarter of 2025. Total revenues were $34.5 million in the fourth quarter, excluding revenues from certain end-of-sale low-margin products. Total revenue increased 9.1% quarter over quarter and 3.6% year over year. Our core revenues were $17.4 million in the fourth quarter, an increase of 10.8% quarter over quarter and an increase of 13.7% year over year. The increase was primarily due to business expansion and usage growth in certain verticals such as live shopping and entertainment. Xiong Wang revenues were RMB 122.2 million in the fourth quarter, excluding revenues from certain end-of-sale low-margin products. Xiong Wang revenue increased 8.2% quarter over quarter and decreased 5.4% year over year. The quarter-over-quarter increase was primarily due to an increase in revenues from digital transformation customers. The year-over-year decrease was primarily due to slowing demand from social and entertainment verticals due to regulation and general economic conditions. Our base net retention rate is 95% for Agora and 79% for Xiong Wang, excluding revenues from certain end-of-sale products and discontinued business. Moving on to cost expenses. Gross margin for the fourth quarter was 66.6%. If we exclude the gross profit from certain end-of-sale products, the gross margin of continuing business slightly decreased 0.1% quarter over quarter and 0.8% year over year. As we mentioned in the last earnings call, we restructured and reduced our global workforce in November 2024. Operating expenses were $28.5 million in the fourth quarter, decreased $4.1 million from $32.6 million in Q2 2024. As a baseline, we expect to see additional savings on operating expenses in Q1 2025. R&D expenses were $14.8 million in Q4, decreased 9.3% year over year. R&D expenses represented 42.9% of total revenue in the quarter compared to 45.3% in Q4 last year. Sales and marketing expenses were $7.3 million in Q4, increased 3.1% year over year. Sales and marketing expenses represented 21.1% of total revenue in the quarter, compared to 19.6% in Q4 last year. G&A expenses were $6.4 million in Q4, decreased 18.4% year over year. G&A expenses represented 18.6% of total revenues in the quarter, compared to 21.9% in Q4 last year. Coming to the bottom line, net income was $0.2 million in Q4, translating to a 0.5% net income margin in the fourth quarter, compared to a net loss margin of 7.2% in Q4 last year. As Tony just mentioned, we achieved a significant milestone this quarter: profitability on a GAAP basis for the first time in several years. This not only demonstrates our ability to deliver on our commitments but also reflects our unwavering focus on creating long-term value for our shareholders. Looking ahead, we remain committed to driving innovation, optimizing operations, and maintaining the financial discipline necessary to ensure continued success. Now turning to cash flow. Operating cash flow was $4.5 million in Q4, compared to $3.7 million last year. Free cash flow was $4.3 million compared to $3.4 million last year. On the balance sheet, we ended Q4 with $363.8 million in cash, cash equivalents, bank deposits, and financial products issued by banks. The net cash inflow in the quarter was mainly due to free cash flow of $4.3 million, which was offset in part by a share repurchase of $1.4 million. The board approved our share repurchase program in February 2022, and as of the end of 2024, we have completed 58% of a $200 million share repurchase program. We are pleased to announce that the board has authorized another 12-month extension of our share repurchase program through February 28, 2026, with all other terms unchanged. This decision reflects the board's confidence in the company's long-term growth potential and our commitment to delivering value for our shareholders. Now turning to guidance. For the first quarter of 2025, we currently expect total revenues to be between $31 million and $33 million, compared to $29.7 million in the first quarter of 2024, if revenues from certain end-of-sale low-margin products are excluded. Looking forward, our goal is to maintain sustained profitability throughout 2025, driven by strategic initiatives, operational efficiency, and emerging market opportunities in generative AI. This outlook reflects our current and preliminary views on the market and operational conditions, which are subject to change. In closing, I'd like to express my sincere gratitude to both the Agora and the Xiong Wang teams for your incredible effort. And thank you to everyone for attending the call today. Let's open the floor for questions.
Operator: Certainly. Ladies and gentlemen, if you do have a question, our first question comes from the line of Harry Zwing from Bank of America Securities. Your question, please?
Harry Zwing: Hi, management. Thanks for taking my question. I'm asking on behalf of a few questions. Firstly, on the demand side, what are the trends in overseas and China markets? And what are the key sectors that are driving growth in 2025? Secondly, on the AI side, has management recently seen an increase in AI-related demand? And in what scenarios have we seen the application of our AI-related technologies? And lastly, on the profitability side, what are the gross profit margin and GAAP profit targets for 2025? Thank you.
Tony Zhao: Okay. I'll talk about the demand side. On the China side, we see demand mostly on the IoT side and the digital transformation side. And, of course, the AI comes in with more activities to explore the potentials. On the global market, we see live commerce, retail economy, and Chinese companies going overseas as driving forces for growth. After our big launch of our conversational AI engine, we have received quite a lot of customer interest actively exploring the opportunity to build with that. So that, in general, on the demand side. Specifically, around AI, we see growing interest from developers and customers, particularly with recent model launches such as Big Tech and Gemini, and a lot of open-source model launches. Our conversational AI engine product is a great fit with our plan to build voice AI agents for voice AI-based applications. We see some huge potentials in a couple of use cases, like companionship in apps and in both apps and IoT devices, such as companion robots and smart toys for 2G. Voice AI can also be widely applied in education platforms, especially for long-range learning and practicing. In enterprise-level executions, such as call centers or customer service, there's a clear demand and a lot of experimental activities. In those areas, we actually see many similarities in all major markets. So it happens in parallel in different markets at this moment. In general, we see a lot of interest in exploring all kinds of use cases based on large language models. But our ability to build methodologies and tools to build agents is far from mature. So it will still take a long time for sophisticated agents to be built and launched. In the long term, those agents will be disruptive and create huge value for us. However, given the reality that the whole technology community is still actively improving the foundation models, infrastructures, tools, and frameworks to update agents, we expect current AI adoption to be faster in existing mature use cases. Right. Voice AI agents could be used to only replace a segment of the whole customer experience. That kind of experience has a much higher chance of success, especially those replacements focusing on replacing some of the human functions to improve productivity or increase efficiency, rather than creating a whole new use case or whole new user experience. That's a more successful path we are seeing in the whole experiment of new AI-based use cases. We are working closely with a group of customers at the POC stage and hope we can see some applications launch in the coming months.
Jingbo Wang: Sure. I'll talk about our gross profit and net profit outlook. So internally, gross margin, we expect it to maintain at a cost level around 67%. As we saw in the fourth quarter, we think that's the right balance between profitability and also competitive pressure here. In terms of the bottom line, I want to start with the revenue. Right? So if you look at our revenue guidance for Q1, at midpoint, it implies a year-on-year growth rate of 7.7%. And our plan is that we need to maintain this growth rate throughout the year, if not accelerate. And on the other hand, we don't expect any significant increase in terms of operating expenses. So as a result, we expect to see a stronger growth on the bottom line compared to the top line.
Harry Zwing: Thanks, management.
Operator: Thank you. And our next question comes from the line of Li Peng Zhao from CICC. Your question, please.
Li Peng Zhao: Good morning, Tony and Jingbo. Thanks for taking my questions and congratulations on the GAAP profitable quarter. I have two questions here. First, I noticed that Agora's growth rate in Q4 reached 13.7% year over year. Could you give more color on the strong recovery of the growth rate? And should we expect the growth momentum of Agora to continue in 2025? Second, regarding the impact of DeepSeq, as we see that it has significantly boosted adoption on both the enterprise side and the consumer side, how do you view the demand for voice and multimodal interaction in the next two years, and which scenarios will be prioritized for Agora and Xiong Wang to focus on in the next two years? Thank you.
Jingbo Wang: I'll take the first question. So, yes, Agora delivered a very nice growth rate in the fourth quarter. And as we guided multiple times in the past, it was driven really by the expansion of new usage in some novel use cases. For example, live commerce, we talked about this many quarters. Two years ago, live commerce was virtually zero, was not a revenue contributor. And in less than two years, it grew to about more than 10% revenue globally and close to 20% of revenue for Agora. And it is still growing. We believe it's still in the early stage of video-based shopping experience. Video-based live shopping experience. In the US and other western markets, video-based live shopping is already everywhere. You know, all the big e-commerce platforms have this functionality. But in Western markets, there's still you can only see that in some new generation apps, not in the mainstream e-commerce apps. So I think there's still a long way to go. In addition to that, we have a creator economy, time-based social networks, which we don't really see that many in China, but it's a promising use case. So that has been growing for a while. We have some customers with Chinese backgrounds while operating completely in Western or Middle East, Southeast Asian markets. They have also delivered growth in the past quarter. So overall, we are still optimistic about the Agora business, and we certainly hope to see further growth.
Tony Zhao: Yeah. I can add more on the growth areas for both sides. As I said, the use case of AI-enabled conversational AI can be divided into two categories: service agent efficiency and saving time, and companionship, which is basically for entertainment and maybe spending more time on having better emotional support or other social demands. We see companionship apps are very active, especially with IoT devices. Education use cases are also trends. We already see real volume start to grow on that. But we see a huge potential on the road in the next two years to continue to grow. IoT devices, as I said, on communication, have a clear chance to grow into volume in the next two years. Customer service and call centers would transform a lot of the initial volume, maybe all one calls first, to agent-based use cases. And some other use cases, including for efficiency side, helping people to run all types of copilot, and those copilots all have a chance to turn into conversational user interactions or interfaces. So those are the areas we see might have a good chance. And, of course, AI community development would also give some strong push, like the open-source community gaining stronger power, and the price of such large language models keeps dropping quite quickly, helping us to enable those growths on the overall trend side.
Li Peng Zhao: Oh, the second question is about DeepSeq. Because the company has significantly boosted the adoption of AI on both the enterprise side and the consumer side, I want to know whether Agora and Xiong Wang want to find some trend in these areas and which kind of key business focus areas for the next two years?
Tony Zhao: Yes. For DeepSeq, as I mentioned just now a little bit, it's definitely a strong push to the whole community and agent builders, where it supplies a high talent intelligence and low-cost model and open-source model on the demand side. So not just on the cost side and the open-source side, but also on the intelligence side, it enhances the quality of the language model capabilities. So that will enable agent building a lot. We are seeing agent building to be the new passport and growing very fast. Innovations around building agents are quite active and start to gain more developer attention, and we are part of it, especially in enabling voice conversational experiences for our agents. In the next two years, multimodal or voice conversational AI experiences will start to grow into more practical use cases. Multimodal understanding will also help enrich the power and reach of such use cases. Voice interactions will be much more efficient compared to currently text-based interactions and will likely become a critical capability of AI agents. As model capabilities of AI agents continue to advance, there are more new use cases to be unlocked. On the framework side, I think there will be more evolution around multi-agent collaboration, agent operation, to help developers' agent building activities. This will enable multi-agent coordination and exchange, paving the way for more complex interactive use cases among human users and AI agents. So those are all factors that are going to accelerate the growth of agents in the next two years.
Li Peng Zhao: Thanks, Tony and Jingbo. That's very helpful.
Operator: Thank you. And once again, if you have a question, thank you, everyone, for attending the company's call today. As a reminder, the recording and the earnings release will be available on the company's website at investor.agora.io. If there are any questions, please feel free to email the company. Thank you.