Robotics company Skild AI has raised $1.4 billion in a funding round, showing that robotics is nearing a fundamental shift. Led by SoftBank Group, with participation from NVentures, Macquarie Capital, Jeff Bezos, Disruptive, and 1789 Capital. Lightspeed, Felicis, Coatue and Sequoia Capital doubled their investments.
Several strategic investors also participate, including Samsung, LG, Schneider, CommonSpirit, and Salesforce Ventures. Other investors include TF Capital, Andra Capital, Palo Alto Growth Capital, KIC, Alpha Square, Mirae Asset, and Destiny. The latest funding brings the company’s valuation to more than $14 billion.
The company plans to eventually introduce the robots into homes, with enterprise tasks being the first application. The new capital will be used to continue expanding the company’s model training and scale future deployments of the technology.
Universal brain of physical machine
At the heart of Skild AI’s approach is the Skild Brain, an omnibody foundation model designed to control robots regardless of shape, size, or function. Unlike traditional systems trained for a single machine or use case, this model can manipulate quadrupeds, humanoids, robotic arms, and mobile manipulators without having to be reconfigured for each body type. The same intelligence adapts to a variety of tasks, from cleaning the kitchen and loading the dishwasher to navigating choppy terrain and handling heavy loads.
One of the biggest limitations of robotics is the lack of data. There is no equivalent to the open internet when it comes to physical interaction. Skild AI avoids this problem by training its models on large amounts of human video data and physically-based simulations. By learning how humans move and interact with objects and practicing in simulated environments, the skilled brain builds generalized physical intuition.
This model is designed to be resilient. If a wheel gets stuck, a limb breaks, or the robot’s body changes completely, you can adjust without retraining. This robustness opens the door to deployment in unpredictable and unstructured environments where most robots struggle.
The team behind Skild AI
Skild AI was founded in 2023 by Deepak Pathak and Abhinav Gupta, who left their professorships at Carnegie Mellon University to commercialize years of basic research in robotics. Together, they bring more than 20 years of experience in self-supervised learning and adaptive robotics, which underpin Skild AI’s core technologies.
The company has assembled a team drawn from Meta, Tesla, Nvidia, Amazon, Google, Stanford, and the University of California, Berkeley, blending academic depth with production-grade engineering. This combination has led to important results regarding robot learning. Skild Brain uses live experience to adapt behavior on the fly, rather than retraining robots when they encounter new environments or bodies. This capability has long been considered out of reach in robotics and has been nominated for best paper awards at major research conferences.
From rapid growth to everyday deployment
Skild AI doesn’t position itself as a remote research institute. The company grew its revenue from zero to approximately $30 million within months in 2025, deploying systems in security patrols, facility inspections, warehouses, manufacturing sites, data centers, and construction sites. Its four-legged platform automates inspection and monitoring, and its mobile operation system allows companies to build robotic applications as easily as an API call.
While enterprise use cases are our top priority, consumer environments are also part of our long-term vision. The new funding will be used to scale model training and support broader commercial deployment, with the aim of making advanced robotics available across the industry.
Skild AI’s ambitions are clear. It’s about moving robotics beyond strict automation to a future where machines can think, adapt, and work alongside humans. If Skild Brain continues to expand as planned, it has the potential to redefine how manual labor is done and who does what.
“Skild Brain takes control of robots that have never been trained and adapts in real time to extreme changes in morphology and environment. The model is forced to adapt rather than remember, much like intelligence in nature,” said Deepak Pathak, CEO and co-founder of Skild AI. “We believe that an integrated omnibody brain is the fastest way to establish a continuous data flywheel that improves models with each deployment, regardless of hardware or task.”
“We believe this 360-degree learning is essential to building AGI that works reliably in the physical world and paving the way for robots that can safely assist humans in everyday environments,” said Abhinav Gupta, co-founder and president of Skild AI. “This allows robots to dynamically operate in complex environments without the need for pre-programmed instructions for each scenario.”
“Skild AI is building the foundational technology for physical AI across robots, tasks, and environments,” said Dennis Chan, managing partner at SoftBank Investment Advisors. “We are proud to partner with Deepak, Abhinav, and the Skild AI team to bring their shared vision to real-world applications around the world.”
“Resolving intelligence in the physical world unlocks tremendous commercial value and long-term national strategic importance,” said IQT Partner Rita Waite. “Skild AI is uniquely positioned to do both, and I look forward to helping this team build.”
