Rhoda AI has emerged from stealth with $450 million in Series A funding and a new robotics intelligence system designed to move autonomous machines beyond controlled laboratory demonstrations and into real-world industrial environments.
The Palo Alto-based company revealed its robotics architecture, FutureVision, built around a system known as Direct Video Action (DVA). The approach relies on video-predictive control that allows robots to continuously observe their surroundings, forecast future physical states as video, convert those predictions into actions, and execute them in a closed loop.
The funding will be used to accelerate research and engineering, expand industrial deployments and customer pilots, and grow Rhoda’s multidisciplinary team spanning generative AI, computer vision, and robotics.
Traditional industrial robots typically operate in structured environments and rely on pre-programmed trajectories. While newer vision-language-action models have demonstrated promising capabilities in laboratory settings, they often struggle when faced with unpredictable layouts, unseen objects, or changing workflows in production environments.
Rhoda’s approach attempts to address this challenge by pretraining its models on internet-scale video datasets containing hundreds of millions of videos. The training helps the system develop a strong understanding of motion, physics, and physical interactions before being refined using smaller datasets of teleoperated robot data that teach the system how to translate predictions into robotic actions.
The resulting system repeatedly observes its environment and predicts future outcomes, allowing it to adjust actions every few hundred milliseconds in response to real-time changes. This continuous feedback loop is intended to provide more reliable control compared with open-loop planning systems that generate plans without ongoing environmental updates.
According to the company, the motion understanding learned during video pretraining significantly reduces the amount of robot-specific data required to train new tasks. In many cases, Rhoda says its system can learn new behaviors with roughly ten hours of teleoperation data.
FutureVision serves as Rhoda’s intelligence layer and foundation model for robotics. The company expects the system to power its own robotics platforms while also being licensed to partners across different hardware and software ecosystems.
Rhoda reports that its technology has already been tested in production manufacturing environments where robots must handle variable materials and constantly changing layouts. In one recent evaluation involving a high-volume manufacturing workflow, a Rhoda robot reportedly completed a component-processing cycle in under two minutes without human intervention while meeting or exceeding customer performance targets.
The company is backed by several major technology investors, including Capricorn Investment Group, Khosla Ventures, Leitmotif, Matter Venture Partners, Mayfield, Premji Invest, Prelude Ventures, Temasek, and Xora, along with Silicon Valley investor John Doerr.
Rhoda was founded by CEO Jagdeep Singh and is supported by a leadership team with backgrounds in generative AI and robotics research. Chief Science Officer Eric Ryan Chan previously worked as a generative model architect at WorldLabs, while Gordon Wetzstein, a Stanford professor, leads the Computational Imaging Lab and contributes expertise in computer vision and imaging systems.
The company’s broader goal is to enable robots to operate autonomously in complex industrial environments such as manufacturing and logistics, where high variability has historically limited automation.
KEY QUOTES:
“We believe the next era of robotics requires models that understand how the world moves — not just what it looks like or how it’s described in language. By learning from internet-scale video and operating in closed loop, our systems are designed to adapt to real-world variability in ways conventional approaches struggle to achieve. The goal is simple: robots that work in the real world, not just controlled lab settings.”
Jagdeep Singh, Cofounder And CEO, Rhoda AI
“In manufacturing, tasks with high variability have historically resisted automation. The real challenge isn’t solving it once, it’s delivering consistent, reliable output under real-world production conditions. What impressed us about Rhoda’s approach is its ability to adapt to conditions that typically require human intervention. Technologies like this can dramatically expand the scope of what can be automated, playing a pivotal role in re-industrializing mature economies.”
Jens Wiese, Managing Partner, Leitmotif And Former Volkswagen Group Executive
“We believe the first company to deploy intelligent, manipulation capable robots at scale in real world environments will kick start a powerful data flywheel, creating a compounding advantage in capturing the long tail of real world edge cases. At Premji Invest, we take a long term view and are highly selective in where we partner. We invest only when we believe a company has the potential to build a truly large, enduring business. We believe Rhoda has assembled the technical foundation, ambition, and execution capability required to achieve that goal, and we are excited to partner with this exceptional team to help bring the next generation of intelligent robots into the world.”
Sandesh Patnam, Managing Partner, Premji Invest
