Skilled AI, a startup that develops AI software for robots, raised $1.4 billion in a funding round that includes a number of big-name investors. The Pittsburgh company is now valued at $14 billion, triple its value since last summer.
Japan’s SoftBank led the Series C funding round with participation from Nvidia’s venture capital arm NVentures, Bezos Expeditions, Macquarie Capital, Disruptive and Florida-based 1789 Capital, of which Donald Trump Jr. is a partner. The new strategic investors include Samsung, LG, Schneider Electric and Salesforce Ventures.
Skilled AI’s stated mission is to build a single, general-purpose artificial brain that can control any robot for any task.
The company said its integrated robotics foundation model, Skilled Brain, is at the center of this effort. It described the model as an industry first.
According to a press release, “Unlike traditional models, which are tailored to specific robot designs, the Skilled Brain is ubiquitous and can control any robot without prior knowledge of their exact body form, including quadrupedal, humanoid, tabletop weapons, and mobile manipulators.”
According to the company, this enables the model to tackle a wide variety of tasks, ranging from basic household chores like loading the dishwasher to more demanding physical challenges like crossing slippery terrain.
Skilled Brain has been trained by watching online videos of humans performing tasks and through physical simulations with thousands of form factors. Unlike robots built for specific applications or restricted environments, this type of in-context learning, as the company calls it, enables robots to develop a diverse skill set and be deployed in different scenarios, collecting data at all times, improving performance and increasing adaptability.
“The Skilled Brain can control robots it has never been trained on, adapting to extreme changes in appearance or environment in real time,” CEO Deepak Pathak said in a press release. “Models are forced to adapt rather than memorize – just like intelligence in nature. We believe an integrated, all-body brain is the fastest way to establish a continuous data flywheel where models get better with every single deployment, no matter what the hardware or task.”
the company has increased rapidly over the past few years, and achieve revenues of $30 million in 2025. It said the latest capital injection will be used to enhance its model training, with the ultimate goal of deploying robotics in the home. However, initially, the focus is likely to be on enterprise use.
