Skild AI is an American technology company focused on developing robotic systems integrated with general-purpose artificial intelligence (AI) models for operation in the physical world. Founded in 2023 by robotics and AI experts Abhinav Gupta and Deepak Pathak, both of whom have academic backgrounds at Carnegie Mellon University, the company operates from Pittsburgh, Pennsylvania, and the San Francisco Bay Area. Skild AI is building a scalable foundation model intended to enable robots to perform across a wide range of tasks and scenarios. The company’s long-term goal is to develop Artificial General Intelligence (AGI) capable of functioning in real-world environments.
Founding and Funding
Skild AI was founded in 2023 by Deepak Pathak and Abhinav Gupta, who have made significant academic contributions to fields such as robotic learning and self-supervised learning. The company's central model is designed to serve as a unified "brain" for various types of robots. In 2024, Skild AI raised $300 million in a Series A funding round led by Lightspeed Venture Partners, Coatue Management, SoftBank Group, and Bezos Expeditions, reaching a valuation of $1.5 billion. Additional investors included Sequoia Capital, Felicis Ventures, Menlo Ventures, General Catalyst, CRV, SV Angel, Amazon’s Industrial Innovation Fund, and Alexa Fund, as well as Carnegie Mellon University.
Technological Approach
Rather than developing vertically integrated solutions for individual robotic systems, Skild AI is building a generalizable model—named Skild Brain—that can operate across diverse robot hardware. The model demonstrates adaptability in tasks such as manipulation, locomotion, and navigation. It is trained on a dataset claimed to be 1,000 times larger than those used by comparable competitors. The training data includes human teleoperation sessions and publicly available video content, enabling the model to learn from a wide range of interactions and scenarios.
Applications
Skild Brain is designed for deployment across various sectors. Robots equipped with visual perception systems and powered by Skild Brain can perform tasks in domestic and industrial environments, while quadrupedal robots can operate in rugged conditions. The company envisions its platform enabling low-cost robots to address labor shortages in sectors such as healthcare, manufacturing, construction, and security. According to the National Association of Manufacturers, the U.S. may face a shortfall of 2.1 million manufacturing positions by 2030—a gap Skild AI believes its robotics technology can help bridge.
Infrastructure and Partnerships
Skild AI's models run on a specialized AI infrastructure developed in partnership with Hewlett-Packard Enterprise (HPE), STN, and NVIDIA. The system utilizes HPE Cray XD670 and ProLiant DL380a Gen12 servers equipped with NVIDIA HGX H200 and L40S GPUs. These resources support model training, multimodal processing, and visualization tasks. Additionally, Skild Brain is supported by NVIDIA’s Cosmos world foundation models (WFMs) and the Isaac Lab simulation environment, which help the model learn and perform tasks under real-world conditions through simulation-based training.
Future Outlook
Skild AI aims to develop AGI grounded in physical-world interaction. The company challenges the notion that AGI can be achieved solely through digital data and instead emphasizes the importance of systems that can perceive, learn from, and act within their physical environments. To support this goal, Skild AI continues to expand its teams in engineering, safety, operations, and algorithm development. The company’s strategy reflects a belief in building autonomous systems capable of adaptive decision-making in complex, real-world settings.
