This startup secures legal video data for physical AI

This article is a part of Startup Spotlight, a series that features young, up-and-coming startups.

Clairva’s co-founders Sunil Nair and Dushyant Verma / Photo credit: Clairva

AI systems can process vast amounts of online data, but training physical AI requires video showing how people interact with real-world environments. Much of the existing training data also comes from a limited number of markets.

Recognizing this gap, the founders launched Clairva to connect AI developers with scalable, rights-cleared video from underrepresented regions such aslike Asia.

😟 Problem

Companies developing robotics and physical AI systems need large volumes of diverse video showing how people interact with physical environments.

Yet high-quality training data can be difficult to source outside major markets such as the US and Europe. Developers also need to know whether the footage they use can legally be used to train AI models, making unclear copyright provenance a potential barrier to deployment.

The challenge is therefore not just collecting more video, but building datasets with enough geographic diversity and clear permissions for AI training.

💡 Solution

Clairva supplies AI and robotics companies with large volumes of real-world training data. The startup delivers several core capabilities.

Its platform and contributor network provide several capabilities:

• Sources licensed video from existing content libraries across Asia.
• Deploys a contributor network to capture custom egocentric footage from a first-person perspective.
• Structures datasets by geography, environment, and camera perspective.
• Tracks permissions and allowed AI uses at the individual asset level.

Image credit: Clairva

📊 Market size

The global AI training-data market is estimated to reach US$3.9 billion in 2026 and is projected to exceed US$16 billion by 2033.

Clairva targets around 200 high-value buyers globally. The company estimates that securing 50 customers spending between US$500,000 and US$1 million annually could generate up to US$50 million in yearly revenue.

🤝 Team

🚀 Traction

🏆 Competition

💰 Financials

🚩 Risks