Dyna Robotics builds full-stack hardware to overcome supply chain bottlenecks
Dyna Robotics is developing its own actuators and inference stack to minimize latency and address US manufacturing gaps. Jason Ma argues that hardware constraints, not model performance, are the primary bottleneck for scaling complex robotic tasks.
Dyna Robotics operates as a full-stack company, owning its hardware, models, and data. Jason Ma states that the company builds its own inference stack to minimize latency, which is critical for real-time physical interaction. The company is developing its own actuators because off-the-shelf components are not modelable enough for simulation and control. Ma highlights that the US lacks domestic manufacturing for critical components like rare earth magnets, creating supply chain vulnerabilities. He argues that hardware constraints, such as dexterous hands overheating, are currently the primary bottleneck for scaling complex tasks, more so than model performance. The company is exploring a hybrid approach to manufacturing, balancing US-based assembly with global supply chains to mitigate geopolitical and shipping risks. Ma notes that previous experience at NVIDIA and DeepMind showed the limitations of not controlling the hardware layer, reinforcing the need for full-stack ownership to optimize the entire system for real-world deployment.