Humanoid Robots Face a Technology Dead End

Tehran - BORNA - For decades, humans have dreamed of creating beings that are neither fully animals nor machines, but something in between human-like robots capable of walking, thinking, and even showing emotions. From the first mechanical arms of the 1960s to today, this dream has been repeatedly discussed by scientists and tech executives, yet it has never felt as close to reality as it does now.

In the mid-2020s, companies like Tesla, Figure, Agility, Unitree, and even startups in China and South Korea are racing to develop the first generation of “robotic workers.” Promotional videos showing robots lifting boxes in warehouses or waving with artificial smiles have gone viral on social media, attracting millions of views. Venture capitalists are pouring billions into startups, many of which still lack a tangible product. Behind this excitement lies a quiet reality that many engineers and robotics researchers view with skepticism.

At the heart of the matter is a simple question with a complex answer: Is the world truly ready for humanoid robots?

Billion-Dollar Investments in “Human-Like” Form

Rodney Brooks, founder of iRobot and one of the world’s leading robotics experts, says the answer is still “no.” He warns that the recent surge of investment in humanoid robots could follow the same fate as the autonomous vehicle industry: big promises, impressive demonstrations, but slower-than-expected real-world progress.

Brooks explains in an analytical essay: “Dexterity the fine motor skills required to handle objects is not something that can be solved with a few deep learning algorithms. We still don’t know how a robot can gently hold an object or place a knife on a table with precision.”

According to PitchBook, more than $5 billion in venture capital has flowed into humanoid robotics companies since 2022. Companies like Figure AI, valued at over $39 billion, lead the pack. Yet these impressive numbers contrast sharply with actual technological progress; many projects remain in prototype stages or only exist as digital renderings.

Fady Saad, a partner at robotics-focused venture fund Cybernetix Ventures, says: “The real market for humanoid robots doesn’t exist yet. They might be useful in space missions or hazardous operations, but in ordinary homes or workplaces, there is no serious demand.”

He also emphasizes safety and psychological concerns: “Imagine a robot falling in a home and injuring a child or pet or simply shutting down and failing to restart. People are still wary of such technology.”

Lessons from Autonomous Vehicles

To predict the trajectory of humanoid robots, one can look at autonomous vehicles. In 2016 - 2018, companies like Waymo, Cruise, and Uber promised streets full of self-driving cars by 2020. Now, in 2025, the technology is still limited to city trials under human supervision.

Sanja Fidler, VP of AI Research at Nvidia, says: “People thought self-driving cars were almost solved, but scaling and interacting with unpredictable real-world environments slowed everything down. Humanoid robots are in the same trap.”

Bill Dally, Nvidia’s chief scientist, also notes that building the infrastructure for robots to operate like humans requires enormous data and computing power, which is still not fully available.

Tesla and the Unfinished Optimus

In 2021, Elon Musk unveiled Tesla’s humanoid robot, Optimus, promising factory deployment by 2023. Two years later, what audiences have seen are largely controlled demonstrations rather than true autonomous operations.

Videos show Optimus folding clothes or walking, but many engineers confirm much of the action is pre-programmed or remotely controlled. Tesla still promises a commercially available version by 2026. Meanwhile, Figure claims its robots will soon operate in BMW factories, though no technical evidence or images have been released. Experts see this level of ambiguity as selling a dream rather than a product.

Mimicking the Human Body Is Hard

Moving a humanoid robot naturally and safely is extremely difficult. Each robot must control over 60 degrees of freedom dozens of motors, sensors, and algorithms must work in perfect coordination.

Seth Winterroth, a partner at Eclipse Ventures, says: “Updating software on a system with six degrees of freedom is already a challenge; imagine one with sixty. And you also need to ensure it’s economically viable to manufacture.”

Touch sensors and “artificial skin” are still immature. Startups like Loomia and Proception are developing these technologies, but commercial application is still years away.

The Future Market: Wheels over Humanoids

Many experts believe future robots won’t necessarily resemble humans. Brooks predicts: “The world will be filled with robots with wheels, arms, or other unusual forms. Functionality matters more than appearance.”

Indeed, robots like Boston Dynamics’ Spot and Stretch show that non-human forms can be far more efficient. Agility Robotics’ Digit operates in logistics warehouses, where social interaction is unnecessary and only work efficiency matters.

Yet the biggest hurdle may not be engineering but human psychology. Studies of the “uncanny valley” phenomenon show that the closer a robot looks like a human, the more discomfort and distrust it generates. A famous example is Hanson Robotics’ Sophia, introduced in 2017 to much fanfare but later criticized for creating more hype than real intelligence.

Realism over Fantasy

Humanoid robots are undoubtedly part of the technological future, but not at the speed or in the form promised. They may one day play critical roles in factories and hazardous missions, but for homes, streets, and everyday life, more time is needed.

As Rodney Brooks concludes: “The world is not ready for humanoid robots but thinking about them is.” While humanity may not yet be ready for “iron people,” it has certainly entered an era where imagining them is unavoidable.

End article