Agora Beats Q2 Revenue Guidance With 18% Growth and Extends GAAP Profit Run

Agora's Q2 Revenue Tops Guidance as Voice Agents Scale

Agora, Inc. (NASDAQ: API), the parent company of Shengwang, reported second-quarter 2026 revenue of $40.42 million, an 18% year-over-year increase that landed above the top end of its guidance. The company guided third-quarter revenue to between $41 million and $42 million, implying growth of 15.8% to 18.6% from the prior-year period.

CEO Zhao Bin said both the real-time engagement business and the conversational AI business maintained strong momentum, and the company has now been profitable under U.S. GAAP for seven consecutive quarters. He said voice agent use cases are expanding into market research, lead collection and customer service, with the company positioning its AI agent solutions to pull call centers toward AI-driven workflows.

The quarter also brought a set of product and partner moves: a new Agora Agents SDK for Python, Node.js and Go; a self-service copyright music library for its karaoke product; robot-focused integrations with Rockchip's RK3588 and RK3576 platforms and with Axera's AX8850N SoC; and appearances at WAIC 2026 and MWC Shanghai 2026.

How Agora's RTE+AI Products Are Redrawing the Voice-Agent Market

Revenue growth is accelerating, but profitability is the real signal

Agora's Q2 revenue exceeded its guidance high end, and the Q3 range points to continued mid-to-high-teens growth. The company's statement that revenue growth continued to accelerate indicates that the combined real-time engagement and conversational AI segments are compounding faster than the prior-year period. Seven consecutive quarters of GAAP profitability, even though the release does not disclose net income figures, suggests that the company is scaling without sacrificing cost control while it invests in infrastructure.

This should be read as a signal about the commercial traction of voice agents rather than a one-time cost event. Agora is explicitly linking future revenue to AI Agent adoption in market research, lead collection and customer service, which is a recurring, workflow-based use case rather than ad hoc API consumption.

The Agents SDK attacks the developer bottleneck

The newly launched Agora Agents SDK, supporting Python, Node.js and Go, is aimed at reducing the amount of code and time needed to build a voice agent on Agora's conversational AI engine. By packaging low-latency audio over its global edge network with turn-taking and interruption handling, the company is trying to make its infrastructure the default for developers who want AI voice capabilities without building real-time transport themselves.

That matters because developer friction is one of the largest barriers to converting AI agent demos into production systems. The karaoke copyright library's self-service function applies the same logic to a different market: it reduces the operational and compliance work required to launch a licensed music product, potentially shortening the sales cycle for media and social customers.

Robotics partnerships give Agora a second growth axis

The Rockchip and Axera integrations show Agora moving from general-purpose real-time communication into embedded AI hardware. The Rockchip solution built on the RK3588 and RK3576 claims end-to-end latency consistently below 200ms over 4G/5G and Wi-Fi, with median network latency of 76ms. Those figures are relevant for teleoperation, home companion robots, agricultural and industrial machines, where control stability depends on predictable latency.

The Axera AX8850N adaptation combines the chip's AI video processing with Agora's global real-time network for video, remote control and signaling. Together, the two partnerships create reference architectures for robot makers and reduce the cost of building real-time interaction into hardware, extending Agora's developer ecosystem beyond software and cloud services.

What Investors and Developers Should Watch in Agora's Next Phase

• Investors: Use Agora's $41–42 million Q3 revenue guidance as the immediate bar. The company is guiding for continued year-over-year acceleration, but confirmation will depend on whether AI voice-agent usage converts to recurring revenue, not just new SDK adoptions.
• Developers and ISVs: Evaluate the Agora Agents SDK across Python, Node.js and Go for voice-agent builds. The stated value is less code and built-in turn-taking and interruption handling, so compare implementation time and audio latency against your current stack.
• Robotics and smart-hardware teams: Request latency data for the Rockchip RK3588/RK3576 and Axera AX8850N reference designs. The published targets are sub-200ms end-to-end latency and a 76ms median network latency, which matter for teleoperation and remote monitoring.
• Call center and customer-service operators: Treat the shift toward AI-driven workflows as a piloting signal, not an immediate replacement. Agora says its voice agents now handle market research, lead collection and customer service tasks, but production data on error rates and escalation costs are still needed.
• Partners and competitors: Watch whether chip-level integrations with Rockchip and Axera become a template for other SoC vendors, because hardware bundling could shift how real-time AI infrastructure is selected in robotics and consumer devices.