Rhoda AI

Rhoda AI develops robot foundation models intended to bring general intelligence into the physical world. The company, based in Palo Alto, California, was founded by a team of industry veterans and Stanford researchers and recently emerged from stealth mode. It raised $450m in a Series A round at a valuation of $1.7bn.

Its FutureVision platform is a Direct Video Action model. Robots are pre-trained on millions of web-scale videos to acquire priors on motion, physics and dynamics, then fine-tuned on as little as 1–10 hours of trajectory data. The approach combines video understanding, imitation learning and trajectory fine-tuning to transfer general-purpose capability to real hardware.

Alongside the models, Rhoda AI builds general-purpose robotic platforms designed for complex, variable tasks - such as returns processing, bearing decanting and contico breakdown - that previously required human intervention. The company operates across automotive manufacturing, logistics, ecommerce fulfilment and general manufacturing, and is scaling up its model research, infrastructure, hardware development and manufacturing capabilities.