en.Wedoany.com Reported - On May 20, Visual Agent startup Chance AI announced the completion of a multi-million dollar angel funding round. This round was led by Meitu, with participation from NYX Ventures and Alibaba-affiliated investment institutions. The funds will be used for model capability iteration, user base growth among North American students, and community and commercialization exploration. This marks Meitu's latest strategic move in the AI visual track and a strategic entry point extending from its imaging tool capabilities upstream into "visual understanding."
Founded in 2025 and headquartered in New York, USA, Chance AI's core product is the world's first AI product using the camera as its primary interaction gateway—the Visual Agent. Unlike traditional AI applications that rely on text input boxes, users open Chance AI and directly capture what is in front of them; the system then understands the visual intent and provides judgments, suggestions, and action plans. Founder and CEO Dr. Zeng Xi elaborated on his product philosophy in an interview: "The input box is not humanity's most natural mode of interaction. The first step in human understanding of the world is often not asking a question, but seeing. We want to enable AI to begin understanding intent from the world the user sees."
Dr. Zeng's academic and industry background provides a unique foundation for this product path. He earned his Ph.D. from the University of Barcelona, with research focusing on cognitive science and contemporary art, specifically how humans understand the world through vision. After graduation, he was responsible for product and design at OnePlus and OPPO, subsequently joining ByteDance's Flow department to participate in the 0-to-1 exploration of AI products like Doubao. After leaving ByteDance in January 2025, Dr. Zeng registered the company in March, officially began operations in July, and launched the first product in September, deeply embedding the camera as a perception gateway into the AI interaction process based on an integrated software-hardware logic.
Upon opening Chance AI, users enter directly into capture mode. Behind the system lies a multi-layered AI reasoning chain: first understanding what the user sees, then understanding why the user captured it—whether seeking styling advice, aesthetic judgment, or knowledge acquisition—and then invoking the corresponding Agent to complete the task. In MMMU-Pro, an authoritative benchmark test measuring multimodal model visual reasoning capabilities, Chance AI's Visual Agent ranked first globally with a score of 86.07. The company simultaneously launched Live Mode—a real-time visual interaction system that integrates knowledge retrieval, content comparison, contextual understanding, and multi-capability orchestration into a complete agent, continuously responding in real-time visual scenarios. In March of this year, Chance AI became the official AI partner of Art Central, marking the first time AI entered the "viewing process" of a major international art fair, with visitors pointing their cameras at artworks and AI participating in real-time viewing and communication.
User data validates the feasibility of the camera-native interaction path. To date, Chance AI's total global downloads have exceeded 200,000, with monthly active users between 40,000 and 60,000. Approximately 40% of users come from the North American market, and the product covers over 35 countries. The company revealed that it has conducted almost no market advertising, with all growth stemming from organic spread. The 30-day return rate reached 49.2%, with core users being young people under 25—a demographic termed "Visual Natives," who grew up in the cultures of Instagram, TikTok, and selfies, inherently understanding the world visually first. In actual use, the product covers diverse scenarios such as interpreting artworks during exhibitions, analyzing outfits while shopping, checking skin conditions, identifying card and designer toy versions, and recognizing plants from menus, with many use cases spontaneously discovered by users rather than pre-planned by the team.
Meitu's Global Investment Head made a clear statement regarding this investment: "We believe the next stage for AI consumer applications is to more naturally enter users' daily decision-making and expression processes. We see Chance AI moving visual AI from image recognition towards aesthetic taste. This is also why we invested—they have productized complex AI capabilities and brought them into the aesthetic lives of young global users." This investment aligns closely with Meitu's recent AI ecosystem strategy. Meitu Chairman Wu Xinhong previously stated publicly that Meitu has opened eight core AI imaging capabilities for integration into the open-source agent ecosystem, while simultaneously promoting internal innovation through a venture capital logic, replacing large-scale decision-making with small-scale trial and error. As a pioneer product in camera-native AI, Chance AI provides a natural vehicle for Meitu to accumulate user scenarios and interaction data at the "visual understanding" level.
Chance AI's direction is converging with the judgments of global tech giants. At the Google I/O conference on May 19, Google launched Gemini Omni, emphasizing AI's seamless multimodal fusion capabilities; OpenAI's GPT-5.5 made visual understanding a core selling point. Vision is becoming the new gateway for AI interaction. Dr. Zeng has a clear judgment on this: "It will be difficult for a single dominant AI company to emerge going forward; everyone will be highly differentiated. We chose to delve deeply into vision because it's not yet a focus today, but it will be mainstream in the future. Our moat is not how strong our model is, but how quickly we can interact with real users."
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
