Shawn Gong

Founder & CEO @wearluckey · Egocentric data for Physical AI · Prev: Nums (1M+ shipped) · Forbes 30U30 · Tsinghua

[Shawn Gong @gonghuachao]

Luckey is in @nvidia Inception 🚀
We're building the real-world egocentric data layer for Physical AI: wearable cameras capturing first-person video, structured into training data for robotics and world models.
Cosmos generates synthetic worlds. We provide the human side.

[Shawn Gong @gonghuachao]

Commercial kitchens, street markets, garment production, and repair operations.

Different people, objects, workflows, lighting, and environments.

The real-world variation is the point.

[Shawn Gong @gonghuachao]

After multiple generations of hardware and software iteration, Luckey’s egocentric systems have already shipped to customers in the U.S.

We’re now expanding our own in-the-wild collection across real workplaces in Southeast Asia.

[Shawn Gong @gonghuachao]

Capture is only the beginning.

Our pipeline turns raw first-person video into model-ready data with fine-grained hand annotations and structured scene/task descriptions—so robotics teams receive usable training signal, not just hours of footage.

[Shawn Gong @gonghuachao]

Most robotics conversations start with models.

The bottleneck we’re seeing is operational: producing diverse, rights-cleared, synchronized real-world data at scale.

That is why we’re building LUCKEY as both a capture-hardware company and a field data operator.

[Luckey @wearluckey]

Buyer requirements are moving beyond single-view video: stereo ego + IMU, ego + exo, ego + tactile, device calibration, verifiable timestamps and robotics-ready delivery. LUCKEY builds both the capture hardware and the field network that produces the data.

[Shawn Gong @gonghuachao]

Luckey is now an OpenAI Select Partner.

Our thesis: staged data trains staged robots.

We sell the other kind. Real workers, real shifts, our own capture hardware, verified clip by clip before it ships.

Egocentric data infrastructure for physical AI → https://t.co/SjnJRbRvWD

[Shawn Gong @gonghuachao]

Hello the world.