Walden Robotics launches at $1.1B valuation for general-purpose robots
Amazon — walden is designed to be a general-purpose robot that continuously learns while working. Walden Robotics today emerged from stealth with $300 million in funding.
The company said it is building and deploying robots that continuously learn and improve while performing real work. “Core advances in physical AI, and all of the excitement and attention surrounding it, has made disruptive change possible,” stated Dr. Russ Tedrake, co-founder and CEO of Walden. “But providing real value to customers and building a robust and scalable business requires a deep understanding and respect for how manufacturing is done today. The best way to make fast and positive progress is by working closely together with the real experts.” Tedrake is a professor at MIT and former senior vice president of large behavior models at Toyota Research Institute (TRI), which spun out Walden Robotics in January. His co-founders at the startup include experts in robotics and AI from Stanford and Amazon. “This team actually invented several of the core innovations that are being used by nearly every physical AI and humanoid startup. You won’t find a stronger team,” Tedrake told The Robot Report. “This team has taken the time to build strong relationships with Toyota; we are very well aligned both technically, and with the core mission about amplifying, not replacing people.” Semi-humanoids designed to let people craft with purpose Cambridge, Mass.-based Walden Robotics said it combines large behavior models (LBMs) with real-world operations. Its expert teams delegate difficult-to-automate tasks so they can focus on problem-solving and judgment. Tedrake said the company’s name, an allusion to Henry David Thoreau’s book, “reflects the deep-seated principle that this transformative technology should be used to expand human potential and enable everyone to work, craft, and live with greater purpose.” The technology builds on a decade of research into diffusion policy, the Universal Manipulation Interface (UMI), LBMs, OpenVLA, and the Drake open-source simulator, explained Walden. The company said this enables its robots to quickly learn new tasks such as machine tending, tool setting, parts kitting, and assembly. The robot has a humanoid torso with two arms and a wheeled base, which Tedrake said makes it easier to certify for safety in factories. Walden also uses imitation learning, simulation, and teleoperation for training dexterous manipulation. The company plans to develop its systems in-house, he added.
