Walden Robotics is presenting a somewhat different vision for humanoid robots: rather than starting with a robot capable of imitating humans in every detail, the company starts with tasks that can deliver clear commercial value and then designs a platform suited to them. The company emerged from stealth on July 15 with $300 million in funding and a $1.1 billion valuation, after nearly ten years of working on difficult robotics problems inside the Toyota Research Institute, from which it emerged.
This timing reflects a shift in the humanoid robotics industry. After a wave of enthusiasm for building human-like robots, even when their near-term commercial goals were unclear, companies have begun facing sharper questions about practical viability and the value they can provide to customers. In this context, the article argues that founding a humanoid robotics company today is very different from founding one several years ago, because success is no longer tied merely to the ability to build a human form, but to proving that this form is suited to a real problem.
Starting with the Use Case, Not the Form
Ross Tedrake, co-founder and CEO of Walden Robotics, says the company chose to forgo legs in its current design. Although there are many reasons to develop legged robots, the practical question for him is how large the market that actually needs legs is, and what proportion can be covered using a wheeled base.
Walden has not yet specified its target applications in detail, citing confidentiality related to its current commercial partners. But manufacturing and logistics appear to be among the likely areas, particularly simple, repetitive tasks that do not readily suit conveyor belts or preprogrammed robotic arms. However, the existence of these tasks does not mean deploying robots will be automatic; the robot competes with human workers who have greater flexibility and whose hiring costs may be lower.
For this reason, Tedrake focuses on high-utilization applications, where the robot operates 24 hours a day, 7 days a week. In his view, the economics of robotics require this level of utilization to make sense in manufacturing. But he explains that economic viability is a necessary condition, not the only one, within Walden's approach or its partnership with Toyota.
The Toyota Partnership and the Priority of Working Conditions
Walden's partnership with Toyota is a natural extension of the company's origins in the Toyota Research Institute, Toyota's research and development arm in Silicon Valley. The article quotes Tedrake as saying that Toyota was proud of the work completed at the institute and willing to expand in this area.
Tedrake also presents the partnership from a people-focused perspective, noting that Toyota's leadership did not focus, in conversations with him, on how much money the project could generate, but on how to improve people's quality of life. In the near term, Walden's vision for this improvement is to assign repetitive manufacturing tasks to robots, allowing skilled craftspeople to make greater use of their expertise and potentially increasing their efficiency, productivity, and job satisfaction. But the article points out that Walden's ability to achieve this outcome is limited, and that customers will not necessarily share Toyota's priorities.
Why a Wheeled Base?
From the company's perspective, a large wheeled base gives the robot several practical advantages. It allows more batteries to be placed close to the ground, while a lower center of gravity helps improve stability and address the risk of running out of power during the workday. Most importantly, a statically stable robot does not face some of the safety challenges associated with legged robots, chief among them the possibility of falling.
Tedrake notes that factories already use autonomous mobile robots with wheels and have safety frameworks designed for these systems. Therefore, under this logic, existing safety procedures can be leveraged instead of creating an entirely new operating case for a legged robot working near people. By contrast, legs remain useful in environments containing stairs, and legged robots may also occupy less space than robots mounted on wheeled bases.
Durable Hands Before Movement Dexterity
Walden applies the same practical logic to its robot-hand design. Rather than the highly dexterous and complex five-fingered hands used by other companies, the company focuses on simple, durable grippers suited to commercial deployment. Tedrake says the real test in a factory is whether the hand can withstand heavy use throughout the week, not how theoretically complex it is.
He notes that Walden's systems have been tested inside a Toyota factory, where the hands are subjected to significant strain. Accordingly, the company designed hands capable of withstanding this use, with Tedrake believing that more dexterous hands were not necessarily capable of performing the same work with the required durability.
From General-Purpose Robots to Versatile Robots
The long-term plan is to build “general-purpose robots,” but the article questions how clear this term is. The idea of a robot capable of doing everything seems undefined, and the author suggests that the practically closer expression is a versatile robot: a platform that can be taught a useful number of different skills, without raising expectations to the level of being able to perform any task.
Walden's optimism about expanding the range of skills is based on previous research at the Toyota Research Institute into diffusion policy, a method that helps robots learn new skills more quickly by using previously learned skills as a foundation. Tedrake believes that multitasking can lead to a general-purpose robot if a single platform succeeds in carrying out a large number of tasks valuable to real customers. The experience accumulated from actual deployments could, under this vision, form the basis for building broader capabilities in the future.