Physical AI is Getting Cheaper and Cheaper, Fueling Rapid Evolution
Table of Contents
A few short months ago, almost every robot made by the hundreds of companies working on humanoid robots could charitably be described as slow, topping out at around three mph. Walking was on the edge of plodding, and arm movements bordered on sluggish and awkward. But just this weekend, we saw a robot running quickly, gracefully, and smoothly.
That’s one clear signal that robots – and other machines that are getting eyesight and intelligence – are evolving quicker now than ever before.
“All of a sudden giving AI a body is becoming economically viable at scale,” says Mat Gilbert, director of AI and data at Synapse in a recent TechFirst interview.
The costs to embed AI in physical hardware are coming down at exactly the same time AI is getting orders of magnitude better. Hardware is no longer a blocker, and physical AI – AI embedded in robots and other smart machines – is already delivering important ROI, says Gilbert.
Humanoid Robotics Accelerates: A New Era of Progress and the Challenges Ahead
The field of humanoid robotics is experiencing a period of unprecedented acceleration, driven by advancements in artificial intelligence, more affordable hardware, and innovative hybrid computing architectures. Companies like Foundation, Apptronik, and Figure are rapidly developing and deploying increasingly elegant humanoid robots, signaling a potential shift towards practical, real-world applications. However, significant hurdles remain, especially concerning safety and the complexities of operating in unstructured environments like the home.
The convergence of Factors Driving Progress
Recent breakthroughs aren’t stemming from a single innovation, but rather a confluence of improvements across multiple areas. AI advancements, particularly in areas like computer vision and reinforcement learning, are enabling robots to perceive and interact with their surroundings more effectively. Simultaneously,the cost of essential hardware components – sensors,actuators,and processors – has decreased,making robot development more accessible.
A key architectural trend is the adoption of a hybrid approach to computing. Critical, real-time tasks requiring immediate responses are processed on the robot itself (on-device). More complex reasoning and learning tasks, which can tolerate higher latency, are offloaded to powerful cloud-based AI models. This allows for both quick reflexes and continuous improvement through data analysis and model optimization.
Rapid Prototyping and Company Progress
The speed of development is remarkable. Sankaet Pathak, CEO of Foundation robotics, reported having a functional humanoid robot prototype just 18 months after the company’s founding. https://www.foundationrobotics.com/ Apptronik is also making significant strides, and Figure, known for its agile and rapidly improving robots, began operations just over two years ago. https://figure.ai/ Figure’s robots are demonstrating increasingly fluid and human-like movement, as highlighted in recent demonstrations. https://www.forbes.com/sites/johnkoetsier/2024/12/04/figure-ceo-teases-running-humanoid-robot-it-looks-fast-agile-smooth/
The Safety Imperative and the “Last Frontier” of the Home
Despite the rapid progress, deploying humanoid robots safely alongside humans remains a paramount concern. Unlike errors in digital AI systems (like Large Language models), mistakes made by physical robots can have tangible and possibly dangerous consequences.As Chris Gilbert,Chief Scientist at Figure,explains,physical actions are frequently enough irreversible.A miscalculation isn’t a “wrong sentence,” but a potentially “catastrophic physical movement.”
This reality makes the home the most challenging – and therefore the ultimate – testing ground for general-purpose humanoid robots. Homes are inherently unstructured and unpredictable environments, filled with dynamic elements like children, pets, and the unpredictable behavior of the people living within them.
Gilbert posits that achieving reliable performance in a home environment signifies a major breakthrough: “We know we’ll have solved general humanoid robotics when you can take a humanoid robot and it can walk into any American home and make a cup of coffee.”
Affordability: The Final Piece of the Puzzle
While technical hurdles are being overcome, the widespread adoption of humanoid robots hinges on affordability. Currently, the cost of these advanced machines remains prohibitive for most families. Continued innovation and economies of scale will be crucial to bringing the price point down and making humanoid robots a practical reality for everyday life.
Related reading