Polaris AI Secures $150M to Build Foundation Models for Industrial Robotics

AI · 5 min read

Polaris trains multimodal foundation models using simulation and real-world robotic behavior to enable few-shot task specification through natural language and demonstrations. The company claims improved generalization across grippers, payloads, and kinematic chains.

The funding, led by Industrial Growth Partners, will support large-scale synthetic data generation, expanded simulator fidelity, and partnerships with robot OEMs. Polaris plans to release a developer SDK enabling rapid task composition and transfer learning for factory use-cases.

Operational teams will need to validate safety, repeatability, and compliance before replacing hand-coded controllers. Polaris intends to provide certified validation suites and explainability tools to help integrators adopt model-driven robot controllers responsibly.