Manufacturing AUTOMATION

Sanctuary AI, Microsoft partner to enhance AI development for general purpose robots

May 9, 2024
By Manufacturing AUTOMATION

Sanctuary AI is collaborating with Microsoft on the development of AI models for general purpose humanoid robots, including Sanctuary AI’s Phoenix. The companies will reportedly work together to enhance AI research and development. Sanctuary AI will leverage Microsoft’s Azure cloud resources for AI workloads.

Sanctuary AI explains in a press statement that building on the foundation of Large Language Models (LLMs), the company is making progress toward “Large Behaviour Models” (LBMs) that ground AI in the physical world by enabling systems to understand and learn from real-world experience. Sanctuary AI is taking advantage of these models for Carbon, the AI control system for its Phoenix robots. Carbon is designed to enable the robot it controls to do complex tasks, due to its foundation of behavioural data. Sanctuary AI will take advantage of Microsoft Azure infrastructure for training, inference, networking and storage.

“Creating systems that think like, and understand us, is one of the biggest civilization-level technical problems and opportunities that we will ever face,” said Geordie Rose, CEO and co-founder of Sanctuary AI. “A challenge like this requires the best global minds to work together. We’re excited to be working with Microsoft to unlock the next generation of AI models that will power-general purpose robots.”

“We’re excited to be working with Sanctuary AI to accelerate AI model innovation and embodied AI research in areas like reasoning, planning, and human-agent collaboration,” says Ashley Llorens, corporate vice-president and managing director of Microsoft Research. “Through our collaboration, Sanctuary AI will have access to Microsoft Azure infrastructure and services as they explore the future of general-purpose robots that can assist across various use cases and industries.”

Advertisement

Print this page

Advertisement

Story continue below