Three years ago, two Carnegie Mellon University professors walked away from their tenured positions with a simple, staggeringly ambitious goal: build one AI brain that could control any robot on earth. This spring, that brain is on the factory floor.
Skild AI, the East Liberty startup co-founded by Abhinav Gupta and Deepak Pathak in 2023, announced in March a set of commercial partnerships that represent the first large-scale, real-world deployment of its flagship product, the Skild Brain. Working with NVIDIA, ABB Robotics, and Teradyne Robotics' Universal Robots, the Pittsburgh company is integrating its omni-bodied robotics foundation model into some of the most sophisticated manufacturing operations on the planet, including the Foxconn assembly lines that build NVIDIA's Blackwell GPU server systems in Houston, Texas.
The Skild Brain is best understood as the GPT moment for physical robots. Where traditional robotics software must be programmed for each specific machine and each specific task, Skild's model is trained across a vast range of robot bodies and environments at once, learning to generalize intelligence the way a large language model generalizes language. The result is a single AI system that can control quadrupeds, humanoid robots, tabletop arms, and mobile industrial manipulators with no prior knowledge of their exact form factor. One brain, any body, any task.
"Our mission is to bring AI into the physical world. The Skild Brain is the foundation model that makes that possible at scale."
Deepak Pathak, Co-Founder & CEO, Skild AI
From East Liberty to the Factory Floor
The March 17 announcement formalized what had been building quietly since Skild's record-breaking $1.4 billion Series C in January. NVIDIA invested in Skild's last two funding rounds and has been working to connect the Pittsburgh startup with two of the world's most widely deployed industrial robot platforms: ABB Robotics and Universal Robots. By embedding the Skild Brain's shared intelligence layer into those manufacturers' existing robot portfolios, the partnership gives factories the ability to extend automation into dynamic, highly variable production environments without needing to write task-specific code for every new workflow.
The Foxconn deployment is the proof of concept the industry has been waiting for. Assembling NVIDIA's Blackwell GPU systems requires precision work at a scale and complexity that has historically resisted full automation. Skild's model, trained on data from thousands of robot deployments across industries ranging from security inspection to package delivery to construction, is now running on those lines in Houston. The Pittsburgh-designed brain, in other words, is helping build the chips that will power the next generation of AI.
A City That Saw It Coming
Gupta and Pathak's decision to stay in Pittsburgh after leaving CMU was deliberate. Both had spent years building research labs inside the university, and both believed the city's concentration of robotics talent and its culture of practical engineering gave Skild advantages it could not replicate anywhere else. That bet has paid off. Skild is actively hiring across its Pittsburgh engineering, AI, and operations teams as it scales to meet demand from its new manufacturing partners, and the company's growth is widely credited with anchoring Pittsburgh's claim to be the most serious robotics city in North America outside of a major coastal hub.
The broader Pittsburgh technology community has watched Skild's rise with something close to civic pride. The startup's trajectory, from a university research project to a $14 billion company in less than three years, is the kind of story that changes how investors, engineers, and global corporations think about a city. When the world's largest semiconductor manufacturer needs a robot brain for its most advanced production line, it calls a company headquartered on Penn Avenue in East Liberty. For Pittsburgh, that is not a footnote. It is the point.