Ex-Googlers are building infrastructure to help companies understand their videodata

InfiniMind, co-founded by Aza Kai and Hiraku Yanagita, aims to convert unused video data into actionable business insights. With advancements in AI, particularly vision-language models, and $5.8 million in seed funding, the Tokyo-based startup has launched TV Pulse in Japan and plans to release DeepFrame in 2026. Their solutions cater specifically to enterprises, offering no-code integration and addressing cost challenges.

Key Points

• InfiniMind was founded by Aza Kai and Hiraku Yanagita, former Googlers with extensive experience in data and video technologies.
• The startup addresses the issue of dark data—unused video—by developing infrastructure to convert it into structured information.
• They secured $5.8 million in seed funding to further enhance their products and operations in Japan and the U.S.
• Their first product, TV Pulse, launched in April 2025, analyzes TV content for businesses to track metrics like brand presence and customer sentiment.
• DeepFrame, their long-form video intelligence platform, will be released in 2026, allowing processing of extensive footage to derive specific insights.

Relevance

• The rise in video content generation aligns with the increasing demand for data analysis and business intelligence tools in 2025.
• InfiniMind's advancements mirror trends in AI, such as the evolution of vision-language models and improvements in data processing efficiency.
• The startup's focus on enterprises reflects a broader trend where businesses seek to maximize the value of their collected data through innovative technologies.

InfiniMind is positioned to capitalize on the untapped potential of dark video data, offering innovative solutions to businesses that can enhance decision-making and operational efficiency, marking a significant shift in how enterprises leverage their video assets.