Humanoid robots may be capable of impressive kung-fu moves and parkour stunts, but industry experts at a Davos panel on Thursday emphasized that the technology needs to move beyond flashy demonstrations and deliver practical, real-world value at scale. The panel, moderated by Business Insider’s Jamie Heller, featured three leading robotics experts who outlined the critical challenges facing the AI-powered robotics industry.
Jake Loosararian, CEO of infrastructure startup Gecko Robotics, identified deployment as the primary obstacle preventing humanoid robots from making significant real-world impact. The Pittsburgh-based company’s cofounder stressed that building reliable datasets about operational environments is essential for startups seeking to transition from impressive demos to tangible returns. “You have to forward deploy and build your robots as close to the environment as possible,” Loosararian explained, noting that this approach provides crucial information unavailable on the internet or YouTube.
Daniela Rus, director of MIT’s Computer Science and Artificial Intelligence Laboratory, highlighted the substantial gap between laboratory capabilities and real-world deployment. While acknowledging that robots can perform tasks like folding laundry and loading dishwashers, she noted the prohibitive cost—potentially half a billion dollars. Rus emphasized that bridging this gap requires advancements in perception, manipulation capabilities, improved sensors, and new AI models that can handle unprecedented situations.
Shao Tianlan, CEO of Chinese AI and robotics firm Mech-Mind, focused on the importance of intuitive learning methods, particularly demonstration-based training where robots learn directly from human coworkers. Despite delivering over 10,000 intelligent robots in the past year, Tianlan argued that humanoid robots don’t need “Einstein-level” intelligence to perform useful tasks. He predicted robots would assume human roles in controlled environments like logistics and service industries within the next “few hundred days.”
The discussion comes as companies like Tesla and Figure prepare for large-scale manufacturing of humanoid robots. Tesla CEO Elon Musk has made ambitious claims about the company’s Optimus robot, suggesting it could eliminate poverty and become the biggest product ever. However, experts noted that most humanoid robots haven’t been deployed in real-world settings, with many demos occurring in controlled environments or using teleoperation—where human operators remotely pilot the machines. Loosararian called teleoperation the “dirty little secret” of robotics, urging companies to be transparent about whether their robots operate autonomously or with human assistance.
Key Quotes
Deployment is the big problem right now for robotics, in terms of the ability for it to begin to make really large impacts, and for there to be a clear road map
Jake Loosararian, CEO of Gecko Robotics, identified the core challenge facing the humanoid robotics industry—moving from controlled demonstrations to scalable real-world applications that deliver measurable value.
I can give you a robot that will fold your laundry and load your dishwasher, but it might cost you half a billion dollars
Daniela Rus, director of MIT’s Computer Science and Artificial Intelligence Laboratory, highlighted the enormous cost gap between technical capability and commercial viability, illustrating why humanoid robots remain impractical for most applications.
There’s a lot of autonomy for certain tasks, but for the majority of the cases, for humanoids, it’s learning in the environment, and it has to do that with teleoperation
Loosararian revealed what he called the ‘dirty little secret’ of robotics—that most humanoid robots still rely heavily on human operators rather than true AI-powered autonomy, calling for greater transparency from robotics companies.
If we want to deploy a robot, I would say demonstration is the most intuitive way to tell robots what to do. That’s exactly how we humans teach others
Shao Tianlan, CEO of Mech-Mind, emphasized the importance of intuitive learning methods for robots, suggesting that demonstration-based training from human coworkers represents the most practical path to deployment in factory environments.
Our Take
This Davos panel reveals a healthy skepticism emerging within the AI robotics ecosystem—a necessary counterbalance to the hype cycle driven by companies like Tesla. The experts’ focus on deployment challenges, cost barriers, and the prevalence of teleoperation suggests the industry is entering a more realistic phase where practical implementation matters more than viral demonstrations.
The tension between Musk’s grandiose predictions and these experts’ measured assessments highlights a broader pattern in AI development: the gap between what’s technically possible in controlled settings and what’s economically viable at scale. The revelation about teleoperation is particularly significant, as it suggests many “autonomous” robot demonstrations may be misleading consumers and investors about the current state of AI capabilities.
Most importantly, the emphasis on environment-specific data collection and learning suggests that general-purpose humanoid robots may remain elusive, with specialized applications in controlled environments representing the more realistic near-term opportunity for AI-powered automation.
Why This Matters
This discussion represents a critical inflection point for the humanoid robotics industry, which has attracted massive investment and generated enormous hype but has yet to deliver widespread practical applications. The experts’ emphasis on deployment over demonstration signals a maturation of the industry, where stakeholders are demanding real-world value rather than impressive but impractical capabilities.
The AI and robotics convergence faces fundamental challenges in perception, learning, and adaptation that current technology hasn’t fully solved. The revelation that teleoperation remains prevalent—essentially humans controlling robots remotely—suggests the industry may be further from true autonomy than marketing materials suggest. This transparency is crucial for investors, businesses, and policymakers making decisions about robotics adoption.
For businesses and workers, the timeline matters significantly. Tianlan’s prediction of deployment in “few hundred days” for controlled environments suggests certain industries should prepare for robotic integration sooner rather than later. However, the cost barriers identified by Rus indicate that widespread adoption across diverse settings remains years away, providing time for workforce adaptation and policy development around AI-powered automation.