Rotaku is trying to make humanoid robots usable for development teams that need more than a desktop demo but cannot justify enterprise pricing. Its Domo platform starts at $2,999 and is aimed squarely at embodied AI research, placing it in a rare middle ground between expensive lab hardware and low-cost robots that cannot truly walk.
A humanoid built around developer workflows
Rather than positioning Domo as a display piece or a remote-controlled show robot, Rotaku is centering the product on research workflows. The company wants developers to run AI models directly on the robot, which is why the platform includes SSH access and onboard deployment for policies that have already been trained.
That approach also explains why Domo is not presented as a drag-and-drop programming tool. Rotaku appears to be targeting teams already comfortable with command-line tools and machine learning pipelines.
Two sizes, two research targets
Domo comes in two versions: Domo Developer and Domo Plus. Both are built for whole-body policy learning, meaning the robot learns coordinated tasks from demonstrations instead of explicit programming for each action.
Domo Developer measures 90 cm in height, weighs 20 kg, and uses 23 degrees of freedom with a peak torque of 70 Nm per actuator. Domo Plus is larger at 130 cm tall and 35 kg, with 25 DoF and 110 Nm torque, putting it closer to a half-scale humanoid.
Both models use an integrated aluminum chassis and a hot-swappable battery. Rotaku says the battery supports two hours of continuous use.
Why the platform is aimed at embodied AI
One of Domo’s core capabilities is VR teleoperation, which supports natural motion capture from two arms. That matters for imitation learning, where realistic human demonstrations are important for training models.
The platform also includes gesture recognition and an LLM-based voice assistant. These features sit alongside the development tools, reinforcing the idea that Domo is meant for practical experimentation rather than simple presentation.
Pricing that changes the equation
The strongest part of Rotaku’s pitch is price. At under $3,000, Domo sits well below several humanoids that target similar users.
Unitree R1 is priced at around $5,000, while Unitree G1 costs $13,500. Noetix Bumi is cheaper at around $1,370, but it is aimed at home users and relies on simpler programming.
| Model | Price | Positioning |
|---|---|---|
| Domo | < $3,000 | Entry for developer-focused humanoid research |
| Unitree R1 | ~ $5,000 | Higher-priced alternative |
| Unitree G1 | $13,500 | More expensive humanoid platform |
| Noetix Bumi | ~ $1,370 | Home users, simpler programming |
That puts Rotaku in a narrow gap that many robotics labs have struggled with for years. Enterprise humanoids can be far too expensive, while cheaper desktop robots often fall short of full-body mobility and research usefulness.
A direct bid for lab adoption
Rotaku’s strategy is to make advanced humanoid research more accessible without stripping away the features that labs need. The company is betting that teams working in embodied AI want a robot they can train, deploy to, and control through the same tools they already use.
The startup, based in the Bay Area, still has to prove it can scale manufacturing and product support against more established robotics players. Reservations for the first production batch are already open, giving early adopters a chance to secure a platform that tries to make full-body humanoid research more affordable.
