Robots folding napkins for 24 hours without a break. That’s not a punchline, that’s Dyna Robotics proving that embodied AI is clocking in for work. The Redwood City startup just landed a $120 million Series A, and the backers aren’t chasing science fair demos, they’re betting on robots already earning their keep.
This round was co-led by RoboStrategy, CRV, and First Round Capital, with Salesforce Ventures, NVentures (NVIDIA’s venture capital arm), the Amazon Industrial Innovation Fund, Samsung Next, and LG Technology Ventures all throwing in. It follows a $23.5 million seed round in March, bringing total funding to $143.5 million and lifting valuation above $600 million. A sixfold jump in six months doesn’t happen on hype alone. It happens when investors see a foundation model not just running in the lab, but deployed in hotels, restaurants, laundromats, and gyms.
The founders aren’t rookies. CEO Lindon Gao and CTO York Yang already built and sold Caper AI to Instacart for $350 million, giving them the muscle memory to scale hardware-powered AI into a business. Chief Scientist Jason Ma comes with a PhD from UPenn’s GRASP Lab and research stints at DeepMind, NVIDIA, and Meta AI, bringing the research depth to match the commercial chops. That blend of entrepreneurial exits and frontier science is rare, and it shows in the pace.
Their flagship DYNA-1 doesn’t stumble through backflips or cosplay as a humanoid. It’s a stationary robotic arm powered by a single-weight foundation model designed for dexterous manipulation. In demos it folded 800 napkins in 24 hours without human help. In live deployments it’s pulling 16-hour shifts with a 99 percent success rate at production-grade quality. The secret sauce is zero-shot environment generalization: skills learned in one location transfer to new settings with no retraining. That’s scale built into the code.
The strategy is as disciplined as it is ambitious. Dyna isn’t chasing physical AGI with glossy slides. It’s getting there one mastered task at a time. Each deployment delivers revenue, generates data, and feeds model improvement. That loop creates compounding returns: smarter robots, happier customers, deeper defensibility. With thirty employees and counting, the Series A will expand research and engineering, fuel model development, and accelerate deployment across industries chewing through billions in repetitive labor.
This raise signals more than capital, it marks embodied AI’s move from speculation to commercialization. Congrats to Lindon Gao, York Yang, Jason Ma, and the Dyna Robotics crew. The robots aren’t coming, they’re already on shift.
