InfiniMind raises $5.8M to turn video archives into business data

InfiniMind has secured $5.8 million in seed funding to expand its video infrastructure platform. Along with the funding, the startup plans to relocate its headquarters from Japan to the U.S., while continuing to operate an office in Japan. The seed funding will be used to continue developing its models, expand engineering infrastructure, hire more engineers, and reach additional customers across both Japan and the U.S. The investment round was led by UTEC, with participation from:

• CX2
• Headline Asia
• Chiba Dojo
• An a16z Scout

The startup addresses the problem of “dark data”—the massive, untapped resource of data that companies collect automatically but almost never use—by utilizing vision-language models, which are AI models capable of processing both visual and textual information. These models have seen annual performance gains of roughly 15% to 20% over the last decade. InfiniMind’s first product, TV Pulse, launched in Japan in April 2025 to analyze television content in real time, helping media and retail companies track brand presence and customer sentiment. The platform already has paying customers, including wholesalers and media companies. Meanwhile, its long-form video intelligence platform, DeepFrame, is capable of processing 200 hours of footage to pinpoint specific scenes, speakers, or events. According to CEO Aza Kai, the beta release for DeepFrame is scheduled for March 2026, with a full launch scheduled for April 2026.

InfiniMind was founded by former Google Japan employees Aza Kai, who serves as CEO, and Hiraku Yanagita, the company’s COO. The co-founders started the company after recognizing an inflection point in video technology while still working at Google Japan. Unlike competitors like TwelveLabs, which provide general-purpose video understanding APIs, InfiniMind focuses specifically on enterprise use cases such as monitoring, safety, and security. The startup’s no-code solution allows clients to bring their own data for processing to receive actionable insights. Kai noted that their system integrates audio, sound, and speech understanding alongside visuals, handles unlimited video length, and prioritizes cost efficiency. He views this work as a critical step toward broader AGI (Artificial General Intelligence) capabilities. “This is an exciting space, one of the paths toward AGI,” said InfiniMind CEO Aza Kai.

Why it matters

InfiniMind is building infrastructure to convert petabytes of unviewed video and audio into structured, queryable business data. By doing so, the company aims to solve the “dark data” problem for enterprises, unlocking value from massive archives that currently go unused.