Zhou Zhifeng, managing partner of Qiming Venture Partners, said: The universal video model is expected to emerge in 12-24 months.

"Qiming Venture CapitalEntrepreneurship and Investment Forum - Entrepreneurial Investment Opens the Resonance Cycle of AI Technology and Applications" was held on July 28 during the 2025 World Artificial Intelligence Conference. Zhou Zhifeng, Managing Partner of Qiming Venture Capital, announced the top ten AI prospects for Qiming Venture Capital in 2025 at the forum. Zhou Zhifeng stated that a general video model is expected to appear within 12-24 months, which can handle generation, inference, and task understanding in the video modality, promoting innovation in video content generation and interaction. In addition, embodied intelligent robots will be the first to achieve large-scale deployment in scenarios such as sorting, handling, and assembly, accumulating a large amount of robot first-person perspective data and tactile operation data, building a closed-loop flywheel of "model-ontology-scene data." This flywheel will drive model capability iteration, ultimately pushing universal robots towards large-scale implementation.