Aishi Tech Secures $300M AI Video Mega-Funding

🔑 Enhanced Key Takeaways

• •Aishi Technology, founded in April 2023 by ex-ByteDance executive Wang Changhu, has grown its user base to over 100 million across apps like PixVerse (overseas) and Paiwo AI (domestic), with 16 million MAU and ARR exceeding $40 million by end-2025.[1][2][5]
• •The company previously raised over $60M in Series B funding led by Alibaba (426M yuan), marking the prior record for largest single-round in China's AIGC video sector, plus earlier rounds including seed (tens of millions RMB) and B+ (¥100M).[6][7][8]
• •PixVerse V5, launched August 27, 2025, tops image-to-video benchmarks, with over 100 partners on its open platform generating 10M+ videos via API; participated in 2025 Global AI for Good Summit.[2]
• •Latest model PixVerse R1 (January 2026) introduces real-time generation and infinite canvas via autoregressive architecture, shifting focus to C-end consumer apps since Q4 2024.[5]

🛠️ Technical Deep Dive

• •Core technologies include multimodal AI alignment, large-scale training systems, efficient data mining, and DiT (Diffusion Transformer) architecture benchmarking Sora.[1][3]
• •Key innovations: World's first Character2Video model (March 2024) for ID consistency in diffusion models; V1 (Jan 2024), V2 (July 2024), V5 (Aug 2025) iterations improving clarity, motion dynamics, physical laws, instruction following, and Agent assistant for prompt-free generation.[2][3]
• •PixVerse R1 (Jan 2026) uses autoregressive architectures for real-time video generation and infinite visual scene extension.[5]

🔮 Future ImplicationsAI analysis grounded in cited sources

⏳ Timeline