Skild AI, a pioneering robotics company advancing the future of AI-powered machines, has unveiled a first-ever glimpse into the capabilities of its groundbreaking foundation model for robotics. Through a newly released video, the company demonstrates the early success of its Physical AI an adaptable, shared brain designed to power a wide range of robotic platforms, environments, and real-world tasks.
“Robotics is marred by Moravec’s paradox: the hard problems are easy and the easy problems are hard. A lot of current robotics models focus on tasks that are hard for humans and easy for robots: dancing, kung-fu, because they are free-space actions and do not require any generalization,” said Deepak Pathak, CEO and Co-Founder of Skild AI. “Skild AI models can not only solve these easy tasks but also solve everyday hard tasks such as climbing stairs even under adversarial conditions, or assembling fine-grained items, which require vision and reasoning about contact dynamics.”
Also Read: Formant F3 Adds Gen AI, Agentic Reasoning to Robot Ops
Unlike traditional vertical approaches that prioritize a single robot or task, Skild AI is pushing the boundaries of generalization in embodied AI.
“Unlike other groups that focus on a vertical approach (focus on a single robot, or single task), Skild AI showcases the general purpose nature of Skild’s Brain: the same brain can be adapted to different tasks and different robot morphologies. From quadrupeds to humanoids, from table-top arms to bimanual AMRs; from home tasks such as dishwashers to physically challenging tasks such as climbing slippery slopes Skild Brain is a one-stop AI Model for all different robot forms,” said Abhinav Gupta, President and Co-Founder of Skild AI.
The video also highlights the robustness of Skild Brain, showcasing its ability to operate safely alongside humans. Trained with end-to-end AI models to apply minimal force, the system demonstrates high adaptability to dynamic environments and unexpected human interaction features often missing in many existing robotics models.
Tackling the Data Challenge
One of the biggest barriers to building a scalable foundation model in robotics is the lack of large-scale, high-quality real-world data. Collecting such data is slow, costly, and complex. Skild AI has taken a unique hybrid approach, combining large-scale simulations and publicly available human video data to pre-train its models. This is further refined using targeted real-world data to ensure performance in practical scenarios.
By incorporating diverse morphologies and training on a wide array of data including human activity Skild AI has created a physical foundation model that works seamlessly across platforms. From quadrupeds and humanoids to tabletop arms and mobile manipulators, the Skild Brain adapts to various use cases and hardware configurations, making it deployment-ready across industries.
A Team Built for Innovation
Skild AI’s team includes world-class researchers and engineers responsible for major robotics advancements such as self-supervised learning, curiosity-driven exploration, sim2real transfer, parkour, and dexterous manipulation. Team members bring experience from leading tech companies including Tesla, Nvidia, Meta, Amazon, Google, and Anduril.
The company is supported by a strong coalition of investors, including Lightspeed Venture Partners, Coatue, SoftBank Group, Jeff Bezos (via Bezos Expeditions), Felicis Ventures, Sequoia, Menlo Ventures, General Catalyst, CRV, Amazon, SV Angel, and Carnegie Mellon University.
Investor Quotes
Stephanie Zhan, Partner at Sequoia: “Skild is the emerging market leader, defining the frontier of general-purpose intelligence capabilities in robotics. Deepak and Abhinav’s life’s work is to bring superintelligence to our physical world, and deliver real impact with customers in real-world applications. The trajectory of our progress is once-in-a-lifetime.”
Raviraj Jain, Partner at Lightspeed: “Skild’s foundation models are truly generalizable across form factors, already showing emergent capabilities and are extremely robust – they represent a new paradigm in embodied AI. Unlike several other robotics demos that are often overfitted for the specific demo environment, Skild robots truly work ‘in-the-wild,’ safely navigating and co-existing with humans.”
Sri Viswanath, Head of Coatue Ventures: “Skild AI has built a brain for every robot, trained on vast simulated and real-world data. I believe its rapid progress shows robots are about to transform every industry and sets a bar the rest of the field is still chasing.”
