Startup Twelve Labs raises $100M from Amazon and NEA for AI video search

Startup Twelve Labs, developer of AI for video search, received $100 million in investments from Amazon, venture fund NEA, and Korean Naver Ventures. The money will be directed toward developing technology capable of quickly searching and analyzing large volumes of video content.

Who entered the round and why

The composition of investors speaks to strategic, not just financial interest in video search technology.

• $100 million — the size of the round announced on July 1, 2026
• Amazon.com Inc. — the world's largest cloud provider (AWS)
• NEA Management Co. — one of the leading venture funds of Silicon Valley
• Naver Ventures — the investment division of South Korea's leading search engine

Amazon's participation is particularly significant for Twelve Labs. AWS serves thousands of corporate clients around the world, and the investment creates a direct path to integrating video search into Amazon's cloud ecosystem. AWS corporate clients will be able to add AI video search to the tools they already use, without changing infrastructure.

NEA Management — one of the oldest and largest venture funds of Silicon Valley. Its participation shows: video search is perceived by the industry as a priority direction, not a niche development.

Naver — South Korea's leading search engine with a strong in-house AI lab. For it, multimodal search is a direct strategic priority: the ability to search video content by meaning, not just by text, is critical for competing with Google in the Asian market.

Why is video search so hard to solve?

Finding a specific moment in a video archive without specialized AI is virtually impossible. Traditional search engines can only index text: subtitles, descriptions, tags. If there are no subtitles, or if you need to find a visual moment — a person in particular clothing, a logo in a frame, a specific gesture — standard tools won't cope without manual markup of every fragment.

Twelve Labs builds multimodal models that analyze video wholly: visual sequence, audio track and speech simultaneously. This allows making semantic queries and getting precise time stamps without prior markup.

Hundreds of millions of hours of video are created daily: corporate records of meetings and webinars, surveillance footage, media archives, training courses, user-generated content. Video makes up a growing share of corporate data, and the vast majority of these materials remain practically unindexed — this is exactly where AI search creates measurable value for business.

What does this mean?

The $100 million round with Amazon, NEA, and Naver participation cements Twelve Labs as one of the key contenders for leadership in the rapidly emerging AI video search segment. Amazon's strategic participation creates an infrastructural foundation for scaling: in a race where early market entry determines long-term positions, AWS partnership becomes a significant competitive advantage.

What is video search via AI?

This is technology that makes it possible to quickly search and analyze large volumes of video content. According to Twelve Labs developers, traditional technologies still solve this task poorly.

What is AI for video search?

This is a neural network that analyzes and searches content in large volumes of video materials — what traditional technologies solve poorly.

Why is a neural network needed for video?

For quick search of needed content in video databases and analysis of video by meaning, which was previously technically impossible.