Ambi Robotics

Ambi Robotics is a warehouse automation company founded in 2018 as a spinout from UC Berkeley's AUTOLAB that has raised $32M to commercialize deep learning-based bin picking systems. The company's robots use AI trained through simulation with domain randomization to perceive and grasp diverse, unstructured items from bins without requiring pre-programming for each SKU. This capability is critical for e-commerce fulfillment where orders contain an enormous variety of products that change constantly with new SKUs. Ambi's AmbiSort system combines the company's bin picking robots with a software platform that manages order batching, robot coordination, and system performance optimization. The company serves e-commerce retailers, third-party logistics providers, and subscription box operators that handle high mix, variable-volume fulfillment where traditional automation requiring custom tooling for each product is not feasible. Ambi competes with Covariant, Plus One Robotics, and other bin picking startups that are applying deep learning to the historically difficult problem of grasping arbitrary objects from unstructured piles, which is one of the most important remaining challenges in warehouse automation.