The most telling detail about Devanthro’s humanoid robot is not its arms or its cameras, but the Oculus Quest headset hanging on a wall in Munich. The headset is the bridge, the point where a human operator steps into the machine. For a company whose tagline is “The Robody Company,” the real product is not just a robot, but a connection. It’s a bet that the fastest path to useful robotics in the home isn’t waiting for perfect artificial intelligence, but stitching together the intelligence we already have.
Devanthro, founded in 2018 by Rafael Hostettler and Alona Pammer, builds “Robodies”,humanoid robotic avatars designed for domestic environments, starting with elderly care. The system is a hybrid: onboard AI handles routine navigation and object manipulation, while a remote human, wearing a VR headset and using haptic controllers, can be teleported into the robot to provide complex assistance, empathy, and judgment. Since 2024, the company says its platform has been deployed in real homes, making it, by its own claim, the first with humanoid robots legally used in domestic settings [Munich Startup, 2025].
The wedge is a human in the loop
Most robotics startups racing toward the home are betting everything on autonomy. Devanthro’s contrarian angle is that this is a sucker’s bet, at least for the messy, unpredictable world of personal care. Their wedge is teleoperation. A caregiver,whether a family member, a professional, or a dedicated operator,can put on a headset, see through the robot’s eyes, hear through its microphones, and speak with their own voice. The robot becomes their physical avatar.
The technology stack is telling. They use the Animus SDK from Cyberselves Ltd. and consumer-grade VR hardware, suggesting a focus on intuitive, low-latency control rather than proprietary spectacle [DEVANTHRO]. The implied unit economics are human-centric: one operator could potentially manage several robots across a shift, handling the moments where AI stumbles,picking up a dropped pill bottle, having a comforting conversation, interpreting an ambiguous request.
A founder track record built for hard problems
The founders’ backgrounds read as a direct response to the immense difficulty of the problem. Hostettler is a serial founder who previously launched a CE- and FDA-certified MedTech company, raising €4 million and holding multiple patents [Munich Startup, 2025]. Navigating medical device regulation is a brutal preview for bringing a robot into a private home. His co-founder, Alona Pammer, brings deep technical robotics experience and recognition, having been listed in Forbes’ 30 Under 30 and named one of ‘20 Women in Robotics You Need to Know 2025’ [Munich Startup, 2025].
Their public presence is that of evangelists for a specific, hybrid future of robotics. Hostettler has been a frequent podcast guest, discussing the opportunities and risks of AI in care, while Pammer has spoken at forums like the St. Gallen Symposium on “Robotics and the Human Touch” [Friederike Biffar LinkedIn]. The team remains small, estimated at under ten people, which fits a company likely still deep in R&D and controlled pilot deployments [Prospeo].
Piloting in the wild
Traction, in Devanthro’s case, is measured in homes and research projects, not revenue multiples. The company has developed over three hardware generations and recently completed a dedicated pilot, “Robody Cares,” where its newest robot was secretly tested in three private households [DEVANTHRO]. It also participated in the Teleskoop Research Project, where a prototype remained permanently in households and was controlled via VR and haptic controllers [FZI Forschungszentrum Informatik, 2025].
These are not lab demos. The tasks reported,meal preparation, pill administration, finding lost items, engaging in conversation,are the gritty, daily work of care [interestingengineering.com]. The core premise is being stress-tested where it matters: in an actual living room, with real people.
The scalability question
For all its cleverness, Devanthro’s model introduces a fundamental tension. The very thing that makes it viable today,the human operator,is also its primary scalability bottleneck and cost center. The economics of remote care work are complex and vary wildly by region. The company’s answer appears to be a gradual handoff: as the robots collect more data and the AI improves, the ratio of machines per operator should increase, driving down cost per hour of care.
The competitive landscape is also shifting. While Devanthro claims a first-mover advantage in legal domestic deployment, it is not alone in targeting the care market with robots. The field includes everything from simple companion bots to more advanced mobile manipulators.
- The autonomy race. Companies like Tesla, Figure, and a host of others are pouring billions into general-purpose humanoids that promise full autonomy. If they succeed faster than expected, a teleoperation bridge might seem obsolete.
- Regulatory moat. Devanthro’s early work on legal deployment in homes could become a significant barrier to entry for newcomers, as navigating safety and liability in private spaces is notoriously difficult.
- The human factor. The emotional component of care,companionship, empathy, nuanced communication,remains a formidable challenge for pure AI. Devanthro’s model inherently includes it.
The next twelve months
The immediate roadmap likely involves expanding pilot partnerships, perhaps with care providers or insurance companies, to move from a few homes to a few dozen. Securing institutional funding, which has so far remained undisclosed, would be a major signal of outside belief in the hybrid model. Technically, watch for advancements in the AI’s share of the workload; each percentage point of autonomy gained improves the unit economics.
A back-of-the-envelope calculation illustrates the stakes. Europe faces an acute shortage of 4.3 million caregivers [Munich Startup, 2025]. If a single teleoperator could effectively support four robots, each providing 6 hours of assistive care per day, that’s 24 caregiver-hours displaced per operator shift. The math only works if the total system cost,hardware, software, connectivity, and human labor,comes in below the local cost of in-person care. That’s the equation Devanthro is solving in real time, in real homes.
Ultimately, Devanthro isn’t just racing against other robot makers. Its incumbent to beat is the status quo: an overburdened, understaffed, and astronomically expensive human care system. Its bet is that a helpful robot with a human heart, beamed in through a VR headset, can ease that burden long before a purely artificial one is ready to try.
Sources
- [Munich Startup, 2025] Devanthro profile | https://www.munich-startup.de/en/startups/devanthro/
- [DEVANTHRO] Company website and technology pages | https://www.devanthro.com
- [FZI Forschungszentrum Informatik, 2025] Teleskoop Research Project details | https://www.fzi.de
- [interestingengineering.com] Article on Robody tasks | https://interestingengineering.com
- [Friederike Biffar LinkedIn] Post on Alona Pammer's St. Gallen Symposium talk
- [Prospeo] Company profile with employee estimate
- [FutureTEKnow] Devanthro overview | https://futureteknow.com/devanthro/