Former Google employees are building infrastructure to help companies understand their video data

Companies are producing more videos than ever before. From years of broadcast archives to thousands of store cameras and countless hours of production footage, most of it unused on servers, unmonitored and unanalyzed. This is dark data: a huge, untapped resource that companies automatically collect but never use in a meaningful way.

To address this problem, Aza Kai (CEO) and Hirako Yanagita (COO), two former Google employees who spent nearly a decade working together at Google Japan, decided to create their own solution. The duo co-founded InfiniMind, a Tokyo-based startup developing infrastructure that turns petabytes of unrendered video and audio into structured, queryable business data.

“My co-founder, who spent a decade leading brands and data solutions at Google Japan, and I saw this turning point when we were still at Google,” Kai said. By 2024, the technology had matured, he added, and market demand had become clear enough that the co-founders felt compelled to build the company themselves.

Existing solutions force a trade-off, explained Kai, who previously worked at Google Japan across cloud, machine learning, ad systems, and video recommendation models and later led data science teams. Previous methods could classify objects in individual frames, but they could not track narrative, understand causality, or answer complex questions about video content. For customers with decades of streaming archives and petabytes of footage, even basic questions about their content often remain unanswered.

What’s really changed is the advances in vision language models between 2021 and 2023. That’s when video AI began to move beyond simple tagging of objects, Kay noted. Lower GPU costs and annual performance gains of about 15% to 20% over the past decade have helped, but the bigger story has been one of power — until recently, models haven’t been able to do the job, he told TechCrunch.

InfiniMind recently secured $5.8 million in seed funding, led by UTEC and joined by CX2, Headline Asia, Chiba Dojo, and AI researcher a16z Scout. The company is moving its headquarters to the United States, while continuing to operate an office in Japan. Japan provided the perfect testing platform: powerful hardware, talented engineers, and a supportive startup ecosystem. Allowing the team to fine-tune its technology with demanding customers before going global.

Its first product, TV Pulse, launched in Japan in April 2025. The AI-powered platform analyzes TV content in real-time, helping media and retail companies “track product exposure, brand presence, customer sentiment, and PR impact” of each startup. After pilot programs with major broadcasters and agencies, it already has paying clients, including wholesalers and media companies.

Now InfiniMind is ready for the international market. Its flagship product, DeepFrame, a long-form video intelligence platform capable of processing 200 hours of footage to identify specific scenes, speakers or events, is scheduled for a beta release in March, followed by a full launch in April 2026, Kay said.

The video analysis space is highly fragmented. Companies like TwelveLabs provide general-purpose APIs for video understanding to a wide range of users, including consumers, consumers and businesses, while InfiniMind focuses specifically on enterprise use cases, including surveillance, safety, security and analyzing video content for deeper insights, Kay said.

“Our solution requires no code; customers bring in their data, and our system processes it, providing actionable insights,” Kay said. “We also integrate audio, audio and speech understanding, not just visuals. Our system can handle unlimited video length, and cost efficiency is a key differentiator. Most existing solutions prioritize accuracy or specific use cases but do not solve cost challenges.”

The seed funding will help the team further develop the DeepFrame model, expand engineering infrastructure, hire more engineers, and reach additional customers across Japan and the United States.

“This is an exciting space, and one of the paths toward artificial general intelligence,” Kay said. “Understanding general video intelligence is about understanding reality. Industrial applications are important, but our ultimate goal is to push the boundaries of technology to better understand reality and help humans make better decisions.”