Skild AI

Skild AI is building a single foundation model, the Skild Brain, that can control many different robot bodies across many different tasks. The launch marks a shift from research demos to paying deployments, with the company reporting that its brain is now running on factory lines, construction sites, commercial kitchens, and inside data centers doing inspection work, picking, welding, and cleaning. One customer cited in the launch video says pick rates roughly tripled from around 50 lines per hour to 130 to 140 after switching to Skild-powered robots, and food services operator AIM is using the system for tray and meal handling at large scale. Skild AI was founded in May 2023 by Abhinav Gupta and Deepak Pathak, two former Carnegie Mellon University professors , and has pursued a hardware-agnostic approach where the same model runs across a wide range of robots, including industrial arms and manipulators, quadrupeds and humanoids . This launch matters because most robotics AI to date has been narrow, tied to a specific arm or a specific task. Skild is arguing that one generalist model, deployed across dozens of customers and embodiments, is finally producing usable work in real facilities rather than staged demos. The pitch to founders, investors, and operators is straightforward. If you run warehouses, plants, kitchens, or data centers and have been evaluating robotics vendors one workflow at a time, Skild is positioning its brain as the layer that sits above the hardware choice, letting the same intelligence take on cleaning, inspection, picking, or cooking as your automation roadmap expands. The company frames this milestone as the start of a feedback loop between fielded robots and the model that improves them, with utility inside customer sites as the forcing function.