Agora - Earnings Call - Q2 2025

Transcript

Speaker 3

Hey, thanks, operator. Welcome, everyone, to our earnings call. I'll first review our operating results from the past quarter. We're proud to deliver our third consecutive quarter of GAAP profitability in Q2, with expanding margins driven by solid revenue growth and continued efficiency improvement. Total revenue in Q2 reached $34.3 million, up 11% year over year, excluding revenues from certain in-house sales, low-margin products, with healthy growth momentum from both the Agora and Zhengwang sides. Our GAAP net profit for the quarter was $1.5 million, more than triple the previous quarter, with a GAAP net margin of 4.3%. Considering seasonality and our strong business outlook, we're confident in our ability to continue improving GAAP profitability through the remainder of the year.

At the end of Q2, we had close to 1,900 active customers for Agora and approximately 2,000 for Zhengwang, excluding Eastmount, representing an increase of 12% and 1% respectively compared to one year ago. Now, turning to our business, product, and technology upgrades for the quarter. Since the launch of our Conversational AI Engine product a few months ago, we have rapidly iterated based on our roadmap and customer feedback. At the World Artificial Intelligence Conference last month, we unveiled a major upgrade with three new features. First, we added an advanced attention-locking feature, which allows an AI agent to focus on a specific user's voice to stay on the topic. Even when there are other people talking in the background, this feature is important because without it, those background voices would be fed into the large language model, confusing the model, and sometimes making the conversation impossible to continue.

Second, we added real-time vision-based intelligence to our Conversational AI Engine, enabling the AI to understand visual context along with spoken conversation. For example, a user can ask about buildings or landmarks while traveling abroad or ask for help with a math problem written on a whiteboard. Finally, we introduced natural interactive avatars that can synchronize lifelike facial and body movements with speech, making conversations more engaging for applications in education, customer support, and beyond. These new features are all aimed at one goal: making AI conversations as natural as human interactions. Now, AI agents can lock onto the user's voice, understand what the user sees, and respond with facial and body movements through avatars. We believe these advancements bring us closer to that goal. We are also encouraged by the extremely warm reception from developers and the general audience at the World Artificial Intelligence Conference.

Over the past quarter, developers and customers across various industries continue to build AI applications using our Conversational AI Engine. Many of them have made solid progress, moving from proof of concept to real-world deployment. For example, a major food delivery platform operating in South America leveraged our Conversational AI Engine to power an AI agent that automatically calls restaurants and local businesses. Currently, our solution supports thousands of agent calls per day, just a small fraction of the total potential volume. As we continue to learn from real-world usage, we expect to improve performance and expand into more use cases and higher volume for this customer. We also believe this success will help drive broader adoption of our Conversational AI Engine in the customer service industry. Another inspiring example is Fuzuzu, an adorable AI-companion toy developed by our customer, Robopoint.

Users received the first shipment in mid-July and quickly shared their excitement on social media. Powered by our Conversational AI Engine, Fuzuzu delivers natural and engaging conversations with users. We have observed that a significant portion of sessions between users and Fuzuzu lasts for more than 50 rounds of interaction. Based on our internal assessment, Fuzuzu is a top performer among AI-companion toys on the market in terms of stability, response quality, and human-like emotional expressions. Venturing into uncharted territory means solving complex problems every day. Our solution is still being optimized and requires ongoing improvement to provide the best performance in some corner cases. That said, we've never been more confident in the transformative potential of human-AI conversation.

In addition, given the greater value and higher technical complexity of our Conversational AI Engine, we expect its pricing to be an order of magnitude higher than our standard real-time communication (RTC) audio product, unlocking significant revenue potential ahead. Today, we also announced the resignation of Zheng Zhong from his position as our Director, Chief Technology Officer, and Chief Scientist due to personal reasons. I'll personally take over his operational responsibilities to strengthen our technology leadership and drive Conversational AI innovations. I would like to thank Zheng for his dedicated service and contribution over the past seven years. Tony Wang, a founding member and currently Chief Revenue Officer of the Agora Division, and Jingbo Wang, our Chief Financial Officer, have been appointed as directors of the board. I'm pleased to welcome Tony and Jingbo to the board.

Their strategic vision and operational expertise will be invaluable to the board as we continue to drive innovation and deliver long-term value for our shareholders. Before I continue, I want to express my appreciation to our leadership team for driving transformation and to every member of the Agora and Zhengwang teams for your dedication. I also want to thank our shareholders for your continued trust and support and for believing in our long-term vision and the significant market opportunity ahead. Together, we are shaping the future of human-AI interaction and engagement. With that, let me turn things over to Jingbo, who will review our financial results.

Speaker 1

Thank you, Tony. Hello, everyone. Let me start by first reviewing financial results for the second quarter of 2025, and then I will discuss outlook for the third quarter. Total revenues for the second quarter reached $34.3 million, excluding revenues from certain end-of-sale low-margin products. Revenues grew 11% year over year, representing our second consecutive quarter of double-digit organic growth. Agora revenues reached $18.2 million in Q2, representing 16.7% year-over-year growth and a 2.2% quarter-over-quarter decline. The strong year-over-year growth reflects our successful market penetration and growing adoption, particularly in high-growth verticals such as live shopping and entertainment. The marginal sequential decrease is mainly due to the earliest fluctuation of customers. Zhengwang revenues reached RMB 115.5 million in Q2. Excluding certain end-of-sale low-margin products, Zhengwang revenues grew 6.7% year over year and 9.5% sequentially, driven by continued business expansion and adoption in key verticals such as digital transformation and IoT.

Dollar-based net retention rate is 97% for Agora and 87% for Zhengwang, marking the second consecutive quarter of improvement for both businesses. Gross margin for the second quarter was 66.8%. If we exclude gross profits from certain end-of-sale low-margin products, gross margin of continuing business slightly decreased 1.2% year over year and sequentially. Moving on to expenses, R&D expenses were $14 million in Q2, decreased 23% year over year. R&D expenses represented 14.8% of total revenues in the quarter, compared to 53% in Q2 last year. Sales and marketing expenses were $6.5 million in Q2, increased 4% year over year. Sales and marketing expenses represented 19% of total revenues in the quarter, compared to 18.3% in Q2 last year. G&A expenses were $6 million in Q2, decreased 26.6% year over year. G&A expenses represented 17.6% of total revenues in the quarter, compared to 24% in Q2 last year.

Moving on to the bottom line, we delivered net income of $1.5 million in Q2, more than triple the previous quarter, representing a 4.3% net income margin. This result represents a significant improvement from the 23% net loss margin in Q2 last year and marks our third consecutive quarter of GAAP profitability. Based on our current business momentum and visibility into the second half, we expect net income to grow sequentially throughout 2025. Now, turning to cash flow, operating cash flow was netted $0.4 million in Q2, compared to netted $7.6 million last year. Moving on to the balance sheet, we ended Q1 with $377 million in cash, cash equivalents, bank deposits, and financial products issued by banks. Net cash flow in the quarter was mainly due to share repurchase of $10.9 million.

In the second quarter, we repurchased 13.1 million ordinary shares or 3.3 million ETFs, representing 3.5% of our entire outstanding shares at the beginning of the quarter. Since our board approved the share repurchase program in February 2025, we have repurchased $127.2 million worth of shares through June 30, 2025. We are fully committed to creating shareholder value through this program. While keeping the financial flexibility, we invest in growth opportunities such as conversational AI. Now, turning to guidance for the third quarter of 2025, we currently expect total revenues to be between $34 and $36 million, compared to $31.6 million in the third quarter last year, representing a year-over-year growth rate of 7.6% to 13.9%. This outlook reflects our current and preliminary views on the market and operational conditions, which are subject to change.

In closing, I want to take a moment to recognize the exceptional work of the Agora and Zhengwang teams. Our double-digit revenue growth and the improving profits are direct results of your effort and focus. To our shareholders, thank you for your trust and support. Together, we are building a lasting company at the forefront of AI innovation. Thank you all for joining the call today. Let's open it up for questions.

Speaker 2

Thank you. As a reminder, to ask a question, please press star 1 1 on your telephone and wait for your name to be announced. To withdraw your question, please press star 1 1 again. Please stand by while we compile our Q&A roster. Our first question will come from the line of Dailey Le with B of A. Your line is open. Please go ahead.

Speaker 0

Thanks, management, for taking my question. Firstly, congrats on the strong results and the solid growth. I have two questions. Number one is regarding the AI-related business. You mentioned, like, for the downstream demand from the IoT and also the AI-companion toys, do we expect which application area will see the fast application and the potential revenue contribution to us going forward? My second question is about the gross profit margin. In Q2, we achieved 4.8% points year-over-year expansion. We see some slightly drop. Is this a signality or any other reasons? How do you see the future trend for gross profit margin and net profit margin? Thank you.

Speaker 3

Thank you for the question. For AI use cases in Conversational AI applications, currently, three use cases have progressed to a more advanced stage, namely call centers, education, and companionship toys. For those use cases, we already see some customers have moved from proof of concept phase to real-world production. Given the vast scale and potential usage of those verticals, we expect the success of those customers will drive broader adoption.

Speaker 1

Okay. I'll review in terms of the margins. The gross margin of the current real-time communication (RTC) RTE business has been quite stable in the past year, fluctuating between 66% and 68%. I expect this to continue in the coming quarters. The sequential and year-over-year decline of about 1% in this quarter is more due to normal fluctuation. As to the new products, the Conversational AI products, I think it's still too early to say about its gross margin. We might have a lower margin initially due to ramp-up and testing and prototyping cost. Once we reach a more steady state, I expect the gross margin of this product to be at least the same as RTC, if not higher, because of the higher value and the higher technical certification of this new product. In terms of net margin, we expect the expenses to be radically stable.

As we ramp up on the revenues, we do expect the net margin to improve in the coming quarters this year. In the longer term, as we guided before, we do think we should be able to achieve an operating margin of between 15% and 20%. That's for medium to long term.

Speaker 3

Thank you.

Speaker 2

Thank you. If you would like to ask a question, please press star 11 on your telephone. Our next question comes from the line of Chen and Lei Wang with CICC. Your line is open. Please go ahead.

Speaker 0

Good morning. I'm Chen Lei Wang from CICC, and thank you for taking my question. My question concerns the market reception of the newly launched advanced attention-locking and the avatar features. Could you please share the early adoption trends and also the client feedback on these innovations?

Speaker 3

Okay. Thank you for the question. When we initially planned and designed these features, we had already taken real-world applications and customer needs into consideration. This actually comes from some of the corner cases we received from early adoption of our Conversational AI Engine. Advanced attention-locking is the essential feature in a noisy environment, especially with other people talking in the background. For example, our customers' AI-companion toy, Fuzuzu, will naturally be carried around by users. When users want to talk to Fuzuzu in a place like a subway station or shopping mall, attention-locking is crucial to ensure a seamless user experience. Natural interactive avatars, on the other hand, are a must-have for users for use cases like education and customer service as they enable more lifelike, engaging interactions. Those are all coming from the demand of solving certain corner cases or providing a better experience for certain use cases.

Speaker 0

Appreciate the highlight. Thanks.

Speaker 2

Thank you. If you have a question at this time, please press star 11 on your telephone. Thank you. I am showing no further questions. This concludes today's question and answer session. Thank you, everybody, for attending the company's call today. As a reminder, the recording is in the earnings release and will be available on the company's website at investor.agora.io. If there are any questions, feel free to email the company. Thank you. You may now disconnect. Everyone, have a great day.