Next-generation hardware and AI enable faster, more reliable hospital logistics, introducing a new robotic World Model.
Diligent Robotics, a Serve Robotics company responsible for developing Moxi, one of the world's largest deployed fleets of mobile manipulation robots supporting hospital care teams, has begun introducing its next-generation Moxi 2.0 platform to healthcare customers across the United States. The new platform incorporates five years of operational experience across more than 25 hospitals.
The launch also introduces Diligent's proprietary learning flywheel, a robotic World Model that continuously learns from real-world fleet operations and improves with each deployment. As part of Serve Robotics, Diligent also benefits from shared investments in physical AI research, supporting faster innovation across its hospital robotics operations.
Moxi 2.0 is designed to improve delivery speed, operational reliability and robot working hours without requiring hospitals to modify their existing infrastructure. Developed using NVIDIA Isaac Sim's open simulation framework and selected capabilities from NVIDIA Cosmos world models, including its 3D lidar tokenizer, the platform features upgraded NVIDIA-powered A2000 computing capabilities. This enables the robot to perceive and interpret its surroundings 10–15 times faster than the previous generation, allowing it to respond more quickly to changing hospital environments, including crowded corridors, opening elevator doors and moving pedestrian traffic.
The platform also features a modular architecture with specialized packaging that supports the integration of future computing systems as new technologies emerge. Through its learning flywheel, Diligent converts real-world deployment data into continuous software improvements, allowing Moxi to become more capable over time.
Moxi 2.0 is being introduced across several U.S. healthcare systems, including Endeavor Health Edward Hospital, Providence Saint John’s Health Center and Children’s Hospital Los Angeles, with further deployments planned nationwide. Diligent is also accepting orders from healthcare providers interested in implementing the new platform in their hospitals.
Key upgrades in the latest Moxi 2.0 platform include:
- 10x onboard compute, 10–15x faster perception: Upgraded processing enables faster real-time decision-making and more confident navigation through dynamic hospital environments.
- Robotic World Model: A new physical AI architecture purpose-built for the complexity of hospital environments, designed to strengthen navigation, task completion, and adaptation over time as the learning flywheel feeds new fleet experience back into the World Model.
- Up to 18 hours of operating time per day: With a runtime of up to 9 hours at a time and 30% faster charging Moxi is ready for long shifts.
- Improved autonomy and recovery: Enhanced handling of edge cases means Moxi can resolve more situations independently, keeping deliveries moving.
- Upgraded cameras, sensors, and storage: Moxi 2.0 adds an expanded sensor suite and redesigned storage drawers.
- Redesigned handles and ergonomics: Informed by direct feedback from nurses and pharmacy teams, making it easier for care teams to interact with Moxi when they need to.
The latest platform is designed to accelerate Diligent's learning flywheel through upgraded sensors and enhanced edge computing capabilities. Moxi 2.0 can now collect more detailed data from the complex environments of active hospitals. This information is transferred to Diligent's cloud-based training infrastructure, where each World Model iteration improves the robot's ability to understand hospital surroundings, handle unexpected situations and perform tasks more effectively. This continuous learning process creates a cycle of improved data collection, more advanced AI models and increasingly capable robots, delivering greater value to hospitals with every software update.
“We’ve spent years in the field learning what hospitals actually need,” said Thomaz. “This is just the beginning of what a true learning flywheel makes possible.”
The cloud infrastructure supporting this learning flywheel was developed in collaboration with Amazon Web Services (AWS). Diligent used Amazon SageMaker HyperPod to train Moxi’s World Model, enabling continuous improvements through large-scale deployment data. T-Mobile for Business has also partnered with Diligent to support reliable network connectivity across complex hospital environments and prepare for future private 5G deployments in healthcare facilities. Moxi’s safety and autonomous functions operate directly onboard the robot, allowing it to complete tasks independently even without Wi-Fi connectivity. Cellular backup is also available to maintain connectivity in areas with limited hospital network coverage.
“Moxi has become an integral team member at Children’s Hospital Los Angeles,” says Omkar Kulkarni, Chief Innovation and Transformation Officer at Children’s Hospital Los Angeles. “Since Moxi joined our team, it has completed more than 40,000 deliveries, representing over 16,000 hours of work that our staff didn’t have to spend transporting supplies and medications across the hospital. We expanded our Moxi fleet from two to three robots over time, and utilization continues to grow, increasing more than 10% in the second quarter alone. Our nursing leaders and pharmacy team appreciate that we’re investing in technology that allows team members to work at the top of their skill set.”
To learn more about Moxi 2.0 or bring Moxi to your hospital, visit www.diligentrobots.com.