DYNA-1 is the first robot foundation model that achieves levels of performance and autonomy suitable for widespread commercial use

Dyna Robotics today announced the launch of Dynamism v1 (DYNA-1), a groundbreaking robot foundation model engineered for sustained, high-performance autonomous operation in real-world environments. DYNA-1 is the first dexterous robot foundation model deployed in commercial settings, which marks a major milestone for embodied AI. Demonstrating high robustness and efficiency, DYNA-1 autonomously executes complex, high-dexterity tasks such as napkin folding using a pair of stationary robotic arms for over 24 hours, with zero human intervention.

DYNA-1 executes complex, high-dexterity tasks using a pair of stationary robotic arms, with zero human intervention.

"We've met with hundreds of customers across industries, and the number one thing they want—unequivocally—is performance, measured by speed and quality," said York Yang, co-founder of Dyna Robotics. "We tried traditional ML, but it struggles to adapt to new environments and can't handle complex, long-horizon tasks like folding. Foundation models are more adaptive, and DYNA-1 is the first embodied AI model to deliver high-quality results at speeds that enable commercial viability."

DYNA-1 achieves an industry-leading 99.4% success rate without human intervention, operating at 60% of human throughput, and can run on its own for more than 24 hours. This is in contrast to other models, which often struggle with sustained operation, encountering unrecoverable errors after an hour or two of operation.

Robot models often fail when they encounter edge cases - a challenge that Dyna Robotics addresses with its proprietary engineered reward model (RM). This RM provides precise feedback on robot interactions and allows DYNA-1 to not only perform tasks with high accuracy but also to autonomously explore effective strategies to recover from errors and generate high-quality training data.

"Most models available today can only operate successfully in very defined, fine-tuned environments, but DYNA-1 demonstrates significant zero-shot environment generalization - meaning our robots can learn tasks in one environment, and perform them successfully in other environments," said Yang.

In a demonstration of its capabilities, DYNA-1 autonomously folded over 800+ napkins in a continuous 24-hour period, meeting production-grade quality standards set by clients.

The success of DYNA-1 extends beyond the lab, and is already being used in real-world environments. The advancements in DYNA-1 also show promising positive transfer to other complex manipulation tasks, including laundry folding and food packaging, highlighting the generalizability of the underlying technology.

"We are at the beginning of hockey stick growth for robotic capabilities," continued Yang. "DYNA-1 has been highly successful at meeting critical benchmarks for commercial viability and early results show that mastery of one skill is highly transferable to new skills, so we expect our models' capabilities will advance rapidly over time."

Customers can contact Dyna Robotics to inquire about commercial uses. Dyna Robotics is actively hiring across AI research and engineering for full-time and internship roles.

About Dyna Robotics

Dyna Robotics makes AI powered dexterous manipulation robots for companies of all sizes. Their robots master one task at a time — from folding to food preparation, allowing their embodied AI foundation models to cost-effectively learn in production environments as the company works toward the ultimate goal of general-purpose embodied AI. Dyna Robotics is founded by repeat founders Lindon Gao and York Yang, who sold Caper AI for $350 million, and former DeepMind research scientist Jason Ma. The company is backed by top investors including CRV and First Round. Learn more at dyna.co.