Disclosure: Some links on this page are affiliate links, meaning we may receive a commission if you click
The volume of video content created by businesses is at an all-time high. Spanning from extensive broadcast archives to numerous surveillance systems and production footage, much of this valuable material remains underutilized, stored away on servers without scrutiny or analysis. This is what we refer to as dark data—a significant, yet largely neglected asset that organizations gather without leveraging its potential.
To address this opportunity, Aza Kai (CEO) and Hiraku Yanagita (COO)—veterans from Google Japan who collaborated for nearly a decade—embarked on an innovative venture. They co-founded InfiniMind, a startup based in Tokyo that aims to revolutionize how businesses convert vast amounts of unviewed audio and video data into organized, query-friendly information.
“While still at Google, my co-founder, who led brand and data strategies there, and I recognized a shifting landscape,” Kai stated. As 2024 approached, the co-founders saw a convergence of technology maturity and market demand compelling enough to launch their own enterprise, he noted.
With a background at Google Japan that encompassed cloud technologies, machine learning, advertisement systems, and video recommendation models, Kai underscored a critical shortfall of existing systems. Previous methodologies could only detect objects in single frames without tracking larger narratives or discerning causality, leaving many questions from clients—especially those with decades of content—unanswered.
The breakthrough came with advancements in vision-language models from 2021 to 2023, as video AI evolved past basic tagging functionalities, according to Kai. While the decline in GPU prices and annual performance improvements of approximately 15% to 20% were beneficial, the significant evolution was in the model capabilities, which previously lacked the necessary sophistication for comprehensive analysis.
Recently, InfiniMind secured $5.8 million in seed funding led by UTEC, with participation from CX2, Headline Asia, Chiba Dojo, and an AI researcher affiliated with a16z Scout. The company is shifting its headquarters to the United States, though it maintains operations in Japan. This strategic move allows them to leverage a robust technological infrastructure and a talented workforce available in Japan, honing their technology with demanding clients before entering broader markets.
InfiniMind launched its inaugural product, TV Pulse, in Japan in April 2025. This AI-enhanced platform provides real-time analysis of television content, equipping media and retail sectors with the tools to assess factors such as product visibility, brand engagement, consumer sentiment, and public relations effects. Following successful pilot programs with key broadcasters and marketing firms, the platform has secured initial paying customers, including wholesalers and media companies.
|
June 23, 2026
Now poised for global engagement, InfiniMind is set to unveil its flagship product, DeepFrame, an extensive video intelligence platform capable of analyzing 200 hours of footage to identify specific scenes, individuals, or events. The beta version is expected to release in March, with a full-scale launch planned for April 2026, according to Kai.
The landscape of video analysis remains diverse, with many companies like TwelveLabs offering generalized video understanding APIs for a wide array of users. However, InfiniMind distinguishes itself by focusing on enterprise-level applications, including monitoring and security, while also deriving deeper insights from video content.
“Our platform is designed to be user-friendly; clients simply upload their data, and we handle the processing, delivering actionable insights,” Kai remarked. “Furthermore, we integrate audio, sound, and language comprehension alongside visuals, ensuring comprehensive analysis. Our system supports unlimited video lengths and excels in cost-effectiveness—setting us apart from many existing solutions that prioritize either accuracy or functionality over cost.”
The newly acquired funding will bolster the team’s development of the DeepFrame model, enhance their engineering framework, expand personnel, and grow their customer base across both Japan and the U.S.
“This sector is thrilling, representing a crucial pathway toward general artificial intelligence,” Kai commented. “Gaining insight into video intelligence equates to understanding reality itself. While industrial applications are vital, our ultimate aspiration is to push technological boundaries to enhance human decision-making.”
### Hustle Verdict
Our take is that InfiniMind’s innovative approach to video analysis is setting a new standard in the industry. By effectively transforming dark data into actionable insights, they not only improve operational efficiency for businesses but also pave the way for a future where data-driven decisions are the norm. We believe this technology will be crucial as organizations increasingly rely on multimedia assets to gain competitive advantages.
