Why we invested in Dyna, building the foundation for general-purpose robots

When Lindon Gao and Jason Ma started Dyna in 2024, they weren’t thinking small. They set out to solve one of the hardest problems in technology: how to make robots that can handle the messiness of the real world.

In less than a year, the team released DYNA-1, a robotics foundation model that achieved over 99% success in nonstop 24-hour operation. Those results came not from controlled lab demos, but from robots folding napkins, packing cups, portioning food, and running shifts in restaurants, laundromats, gyms, and hotels. The model didn’t just work, it generalized: when deployed in new environments, it operated out of the box with no extra training.

That ability to generalize is what sets Dyna apart. Their vertically integrated stack combines a reward-model control system with real-time uncertainty forecasting and continual self-play learning. Each robot streams new data back to the cloud, improving the model overnight, so the whole fleet gets better every day. The system is hardware-agnostic today and margin-accretive tomorrow, with a roadmap toward affordable in-house arms that can drive software-class economics.

We’ve been impressed not only by the technology but by the people behind it. Lindon previously built and sold Caper AI to Instacart for $350 million, proving he knows how to turn cutting-edge AI into commercial products. Jason spent years at DeepMind and NVIDIA advancing robotics research, and together with co-founder York Yang and a world-class team of engineers, they’re executing at remarkable speed.

At Samsung Next, we see a future where in-home, multi-purpose robots become as common as smartphones. For that to happen, robots need to succeed in cluttered, unpredictable environments with dynamic human behavior—the same challenges Dyna’s systems are solving today in commercial kitchens and laundries.

That’s why we’re excited to join Dyna’s $120M Series A, alongside Robostrategy, CRV, and First Round Capital, with participation from Salesforce Ventures, NVentures (NVIDIA's venture capital arm), the Amazon Industrial Innovation Fund, and LG Technology Ventures. It’s rare to see a robotics company prove both technical breakthrough and commercial traction this early. Dyna is already building toward physical AGI, and their progress shows it’s not just a vision—it’s happening now.

Niko Ciminelli is an investor at Samsung Next. Samsung Next’s investment strategy is limited to its own views and does not reflect the vision or strategy of any other Samsung business unit, including, but not limited to, Samsung Electronics.

Next
Next

Why we invested in Kite, building the foundational infrastructure for the agentic economy