MassRobotics, AWS, and NVIDIA Announce Second Cohort of Physical AI Fellowship

Nine promising startups selected to advance physical AI across industries

Published on Mar. 12, 2026

MassRobotics today announced the second cohort of the Physical AI Fellowship, a virtual program powered by Amazon Web Services (AWS) Startups and NVIDIA Inception that helps high-potential robotics startups build, refine, and scale physical AI solutions. The 2026 cohort brings together nine startups working on physical AI applications in agriculture, construction, industrial automation, retail, logistics, robotics data infrastructure, teleoperation, and wearable/humanoid robotics.

Why it matters

Demand for physical AI is accelerating as industries seek to deploy intelligent machines in real-world environments. The fellowship aims to help startups overcome technical and resource barriers to validate, refine, and scale their physical AI solutions across various applications.

The details

The Physical AI Fellowship provides startups with deep technical guidance, cloud and compute resources, and access to a global robotics ecosystem. Fellows receive support from AWS Generative AI Innovation Center scientists and experts, as well as AWS credits. The fellowship also provides benefits via NVIDIA Inception, including mentorship, technical training, and the latest developer tools, alongside access to MassRobotics' facilities and robotics community.

  • The 2026 cohort was announced on March 12, 2026.
  • The startups will showcase their work at the Robotics Summit & Expo in Boston on May 27-28, 2026.

The players

MassRobotics

The world's largest independent robotics hub dedicated to accelerating robotics innovation, commercialization and adoption.

Amazon Web Services (AWS) Startups

A division of AWS that supports startups through cloud resources, technical guidance, and access to a global ecosystem.

NVIDIA Inception

NVIDIA's startup program that provides mentorship, technical training, and access to the latest developer tools.

Tom Ryden

Executive Director at MassRobotics.

Jason Bennett

VP and Global Head of Startups at AWS.

Howard Wright

Vice President of Startup Ecosystem at NVIDIA.

Got photos? Submit your photos here. ›

What they’re saying

“We're excited to launch our second Physical AI Fellowship cohort and continue working alongside industry leaders like AWS and NVIDIA to help startups scale real-world physical AI solutions.”

— Tom Ryden, Executive Director at MassRobotics (MassRobotics)

“Physical AI represents one of the most transformative opportunities in robotics, and startups are leading the way in bringing these intelligent systems from lab to market. This fellowship is designed to remove technical and resource barriers that slow innovation by embedding our scientists and engineers directly with cohort companies, providing substantial cloud credits and AI services, and connecting them to enterprise customers who are eager to deploy these solutions. This model has proven to accelerate how quickly startups can validate, refine, and scale physical AI across real-world applications.”

— Jason Bennett, VP and Global Head of Startups at AWS (MassRobotics)

“Startups are an incredibly powerful catalyst for innovation. By taking early bets on emerging technologies, they transform bold ideas into scalable businesses. As physical AI brings intelligence to machines, factories, and infrastructure, these innovators will drive greater efficiency and productivity across the global economy. In collaboration with AWS, we're showing that our technologies are better together – helping startups accelerate innovation and bring transformative AI solutions to market.”

— Howard Wright, Vice President of Startup Ecosystem at NVIDIA (MassRobotics)

What’s next

The startups in the 2026 cohort will showcase their work at the Robotics Summit & Expo in Boston on May 27-28, 2026.

The takeaway

The Physical AI Fellowship is helping high-potential robotics startups build, refine, and scale physical AI solutions across a range of industries, demonstrating the transformative potential of intelligent machines in real-world environments.