Robotics investments based on flawed, unsustainable assumptions, renowned researcher

Rodney Brooks, a renowned roboticist who co-founded iRobot and former MIT researcher, has warned that current investments in humanoid robots are based on flawed assumptions and may not be sustainable in the long term.

Brooks argues that many companies, including Tesla and Figure, rely on optimistic assumptions about robots learning that do not conform to the laws of physics, sensing, and manipulation in the real world.

According to Brooks, approaches that rely on video-based learning to teach human-like dexterity won’t be sustainable. He characterized the approach as pure fantasy thinking, saying that human hands have multiple touch receptors that no robot can match. He highlighted the things a robot may fail to replicate, including human tactile feedback to pressure, vibration, and slip detection, which are central to dexterity.

Future robots won’t look human, Rodney predicts 

In an essay published by Rodney Brooks, he highlights the lack of established infrastructure in the robotics industry to capture, store, and use touch data. Unlike vision and speech, where there are plenty of datasets and signal processing frameworks, touch remains underdeveloped in hardware and algorithm design. 

According to Brooks, many robot systems today lack adequate force feedback, fine finger control, and robust tactile sensing, which are essential for general-purpose manipulation. He also highlighted safety concerns when robots continue to scale. Brooks believes that as size increases, so does the size and body of robots, meaning a fall from a larger robot could be far more dangerous than from a smaller one. 

Scaling up in size for robots means its mass grows with the cube of linear scaling, while also structural strength grows with the square, creating safety challenges. Brooks revealed that walking upright is another major hurdle for the existing humanoid robots, pointing out that they use energy-intensive control and are not yet structurally safe to be near humans when in motion.

He argues that to operate in human environments, robots must be certified as safe when people stand within mere centimeters, pointing out that the current designs have not achieved that. 

Brooks predicts that in the next fifteen years, the most successful robots marketed as humanoid will not look like humans; instead, they will rely on wheels, multiple arms, and specialized sensors rather than walking upright on two legs. He added that the current funding rounds to support humanoid experiments are unlikely to result in the mass production of humanoid machines.

Mass humanoid robot adoption is much further away than expected

Chris Camillo, a renowned investor who turned $20K to $60 million by investing in robotics, has also sounded his warnings, revealing that after spending more than 400 hours researching the space, he remains optimistic about long-term potential.

Still, the challenges ahead are being misinterpreted by many investors. At the Independent Investor summit, Camillo points out that the perception that manufacturing capacity is the key bottleneck for humanoid robots is misguided. 

“Most people I speak to that are deeply engaged with this space, we all agree on one thing: at the point where we actually have a scalable humanoid robotics platform, fully scalable that has met all the KPI thresholds… we will start a 10–15 year supply-demand imbalance, where it doesn’t matter how much you can manufacture,” 

–Chris Camillo, Renowned Robotic Investor

Camilllo explained that the difficulty lies in deployment, not production, and pointed out the integration challenges companies face when adopting generalized robotics. He cited that it’s hard to deploy generalized robots at Walmart, FedEx, or Coca-Cola, and that it will take years of strategic work, integration, preparing a workforce, and digitizing inventory before incorporating humanoid robots into the manufacturing floor or attempting to do anything.

Both Camillio and Brooks point out that much of the investor community currently fails to appreciate how long and complex the process will be. As a result, while humanoid robots may eventually play a role in global industries, the timeline for real-world value may be far longer than they predict. 

Join Bybit now and claim a $50 bonus in minutes