In a move that screams the AI boom is far from over, robotics intelligence startup Rhoda AI has burst out of an 18-month stealth period by announcing a colossal $450 million Series A funding round. The investment, led by Premji Invest, catapults the Palo Alto-based company to a hefty $1.7 billion valuation and officially unveils its ambitious plan to give industrial robots a brain trained on internet videos.
Rhoda AI’s platform, dubbed FutureVision, aims to solve a long-standing problem in robotics: creating machines that can adapt to messy, real-world environments instead of being confined to rigidly programmed tasks. The company’s secret sauce is a “Direct Video Action” model. Instead of relying solely on painstaking data gathered by humans remotely operating robots, Rhoda pre-trains its AI on hundreds of millions of public internet videos to build a foundational understanding of physics, motion, and interaction. This knowledge is then fine-tuned with a smaller amount of robot-specific data, allowing its systems to handle unexpected situations in manufacturing and logistics settings.
This strategy of using vast, unstructured video data to create generalist AI models is a significant departure from traditional robotics and mirrors the foundation model approach being championed by giants like Nvidia and Tesla. While Tesla trains its Optimus humanoids and self-driving AI using real-world video fleet data, Nvidia is building the ecosystem for others with its Isaac platform and GR00T foundation model for humanoid robots. Rhoda is positioning itself as the “brains” provider for the broader industrial market, a hardware-agnostic play that could upgrade existing fleets of robots.
Why is this important?
The sheer size of this Series A round for a company focused purely on software signals a massive vote of confidence from investors like Premji Invest, Khosla Ventures, and Temasek. It suggests the market believes the ultimate value in the next wave of automation lies not in the robotic arms and grippers themselves, but in the AI that powers them.
By training robots on the endless chaos of the internet, Rhoda AI is betting it can create a scalable, adaptable intelligence that bypasses the bottleneck of traditional programming. If FutureVision can successfully translate YouTube-level knowledge into reliable, factory-floor actions, it could dramatically lower the barrier to automating complex tasks that have, until now, remained stubbornly human. It’s a bold attempt to build the Android for a world of increasingly capable robotic bodies.
