Skild AI has closed a $1.4 billion funding round led by SoftBank Group, with participation from NVIDIA Ventures, Macquarie Capital, and Jeff Bezos through Bezos Expeditions. The investment brings the company's valuation to over $14 billion.
The round also includes returning investors Lightspeed, Felicis, Coatue, and Sequoia Capital, who doubled down on their positions. Strategic partners joining the cap table include Samsung, LG, Schneider Electric, CommonSpirit, and Salesforce Ventures. Additional backers include Disruptive, 1789 Capital, IQT, TF Capital, Andra Capital, Palo Alto Growth Capital, KIC, Alpha Square, Mirae Asset, and Destiny.
The Pittsburgh-based company is developing what it calls the Skild Brain, a unified robotics foundation model designed to operate any robot without needing prior knowledge of its physical form. Unlike traditional robotics systems built for specific machines or tasks, the Skild Brain works across different robot types—quadrupeds, humanoids, tabletop arms, and mobile manipulators.
The model enables robots to perform tasks ranging from household chores like loading dishwashers to more demanding activities such as navigating difficult terrain. Skild AI's approach addresses a fundamental challenge in robotics: the absence of large-scale training data comparable to what exists for language or video models.
To overcome this, the company pre-trains the Skild Brain using human videos from the internet and physics-based simulations. This method allows the model to learn from vastly larger datasets than would be possible using robot-specific data alone. The omni-bodied design means the model can transfer knowledge across different robot forms, expanding its training foundation exponentially.
"The Skild Brain can control robots it has never trained on, adapting in real time to extreme changes in form or environments. The model is forced to adapt rather than memorize – much like intelligence in nature," said Deepak Pathak, CEO and Co-Founder of Skild AI. "We believe that a unified, omni-bodied brain is the fastest way to establish a continuous data flywheel where the model gets better with every single deployment, no matter what the hardware or task."
One distinguishing feature of the Skild Brain is its ability to adapt through what the company calls in-context learning. When introduced to a new robot body or environment where initial actions fail, the model adjusts behavior based on live experience rather than requiring retraining or fine-tuning.
This capability allows robots to handle unpredictable scenarios such as loss of limbs, jammed wheels, increased payload, or entirely new body configurations. The achievement has earned the company's research team Best Paper nominations at top robotics conferences.
"We believe this omni-bodied learning is essential for building AGI that works reliably in the physical world, paving the way for robots that can safely help humans in everyday environments," said Abhinav Gupta, Co-Founder and President of Skild AI. "This enables robots to operate dynamically in complex environments, without requiring preprogrammed instructions for each scenario."
Over the past year, Skild AI has published research demonstrating the model's ability to adjust robot movements based on what they perceive in their surroundings. The company says this represents a research breakthrough not previously seen in robotics.
Skild AI has emerged as one of the fastest-growing robotics companies. The startup grew from zero to approximately $30 million in revenue within just a few months in 2025, according to the company.
Current deployments span multiple sectors including security and facility inspection, last-mile and point-to-point delivery, warehouses, manufacturing, data centers, and construction. The company positions itself as a potential force in American manufacturing automation.
While enterprise applications come first, Skild AI plans to ultimately deploy its robotics technology in consumer homes. The new capital will fund continued scaling of model training and expansion of deployment operations. The company operates offices in Pittsburgh, the San Francisco Bay Area, and Bengaluru, India.
SoftBank's investment reflects strong conviction in the company's approach to physical AI. "Skild AI is building foundational technology for Physical AI across robots, tasks, and environments," said Dennis Chang, Managing Partner at SoftBank Investment Advisers. "We're proud to partner with Deepak, Abhinav, and the Skild AI team to bring that shared vision into real-world applications worldwide."
IQT, another investor in the round, emphasized both commercial and strategic importance. "Solving intelligence for the physical world unlocks enormous commercial value and long-term strategic national importance," said Rita Waite, Partner at IQT. "Skild AI is uniquely positioned to do both, and we're excited to be working with this team as they build."
Skild AI was founded in 2023 by Deepak Pathak and Abhinav Gupta, both recognized as pioneers in self-supervised and adaptive robotics. The company is building what it describes as a scalable foundation model for robotics that serves as a shared, general-purpose brain across diverse robot embodiments and real-world tasks.
The company's data infrastructure pulls from four sources: large-scale simulation generating synthetic experiences, internet videos showing human actions, teleoperation data mapping images to joint movements, and real-world deployment data from robots operating across various sectors. Each deployment feeds back into the model, creating what the company calls a data flywheel that continuously improves performance.
