Vancouver-based Sanctuary AI, a company focused on developing human-like intelligence, is collaborating with Microsoft to accelerate the development of artificial intelligence models for humanoid robots. Specifically, this collaboration includes the use of Microsoft’s Azure cloud resources to evaluate model performance, efficiency and planning capabilities.
Sanctuary AI is built on a framework that aims to create human-like intelligence in general-purpose robots and is based on Large Language Models (LLMs). What the company calls “Large Behavioral Models” (LBMs) aim to adapt AI to the physical world by learning from real-world experiences. These models will be integrated into Carbon, Sanctuary AI’s AI control system for Phoenix robots that will enable the execution of complex tasks.
Geordie Rose, CEO and Co-Founder of Sanctuary AI, emphasized the importance of the collaboration: “Creating systems that think and understand like us is one of the biggest technical challenges and opportunities we will face as a civilization. This kind of challenge requires the best global minds to work together.”
Microsoft and Sanctuary AI Collaborate for AI Research
Working with a leadership team that includes founders D-Wave, Kindred AI and Softbank Robotics, Sanctuary AI offers a wide range of expertise. Sanctuary AI has proven its capabilities in commercial environments, having tested its robots in over 400 tasks across a variety of industries.
“We are excited to work with Sanctuary AI to accelerate AI model innovation and tangible AI research in areas such as reasoning, planning, and human-machine collaboration,” said Ashley Llorens, Corporate Vice President and General Manager, Microsoft Research.

