The two bets founders are making in physical AI | Sifted

The physical AI robotics sector is splitting between two strategies: scaling data and models first, or designing for real-world deployment from the start.

Companies like Physical Intelligence, 1X and Generalist favor data-first scaling in controlled settings, while FieldAI, Flexion and Waymo lean toward architecture-first systems built for messy environments.

That choice has major implications for safety, reliability, deployment scope and how quickly robots can become commercial products.

The key competitive question is who can move from impressive demos to dependable real-world systems fastest.