Why Boston Dynamics Lost to Unitree?

Robotics in Everyday Life
3 followers
For over a decade, Boston Dynamics was the undisputed king of robotics, amazing the world with viral videos of backflipping humanoids. So how did companies like Unitree, Tesla and Figure leapfrog the pioneer to dominate the mass-production humanoid race?
#machine #Robotics #robot
AI Generated·May contain errors
Video Summary
AI GeneratedThis video explores why Boston Dynamics, once the undisputed pioneer of humanoid robotics, lost its commercial lead to newer competitors like Unitree, Tesla, and Figure. It analyzes the shift from high-end research to mass production, highlighting the roles of hardware choices, corporate stability, and the integration of AI.
Hi, welcome to Future on Wire. A decade ago, Boston Dynamics showed us what felt like the future of
robotics. Millions of people watched their viral videos of robots doing backflips and running
through the woods, imagining a future straight out of a movie where robots lived alongside us.
Today, that sci-fi future is finally happening, but with a twist. Instead of Boston Dynamics
leading the way, brands like Unitree, Tesla, and Figure are driving innovation,
building robots in factories and selling them to the public.
We've seen humanoids interacting with fans at the NBA Finals and the World Cup.
Meanwhile, Boston Dynamics' famous humanoids seem to have stepped back from the mass production
spotlight. So, with such a massive head start in research, how did Boston Dynamics end up
losing its lead in the commercial humanoid race to companies created less than a decade ago?
Humanoid robotics was pioneered by Waseda University in the 70s. At that time,
Professor Ichiro Kato developed the Wabot-1, which many considered the world's first full-scale
humanoid robot that paved the way for Japan to become a global leader in the industry.
While Western research focused on restationary arms and wheeled rovers for factory floors,
Japanese researchers focused on getting robots to walk on two legs. You might remember presentations
of Honda's Asimo climbing stairs, running, and recognize human gestures. It was a major
showcase for dynamic balance and motion control, but it never entered proper commercial production.
It remained an experimental research project that established Honda and Japan as technology innovators.
In 1992, drawing from these Japanese foundations, MIT researcher Mark Rybert founded Boston Dynamics.
The company initially focused on military R&D for hydraulic-legged robotics, funded by DARPA,
the R&D agency under the U.S. Department of Defense. You've probably seen videos of BigDog,
one of Boston Dynamics' first four-legged robots, being kicked around to demonstrate its balance.
Over time, the company improved its designs, moving from four-legged machines to two-legged
robots that eventually culminated in the hydraulic Atlas. For a long time, Atlas was the main focus
of Boston Dynamics, showcasing whole-body dynamic control and two-legged mechanics.
The company became a world-famous entity through viral videos of Atlas doing parkour,
backflips, and dynamic balance recovery, setting the industry benchmark for humanoid robots.
However, this version of Atlas never entered mass production.
It stayed as a technological showcase rather than a practical tool for the real world.
There are a few key reasons why this happened and why other companies have taken the lead today.
To understand why, we need to look at three key factors.
While Boston Dynamics hasn't lost its engineering prowess or physical capabilities,
it has seeded the leadership it built in the early 2000s. Today, when people think of humanoids,
the first image that often comes to mind is Unitrix G1. These robots are being showcased
all over the world doing acrobatics and going viral, much like Boston Dynamics Atlas used to.
Founded in 2016 by engineer Wang Jingxin in Hangzhou, China, Unitrix focused its initial
efforts on four-legged robots built directly for commercial use, rather than focusing solely on
R&D. Their core advantage was a dramatic cost reduction, enabled by building their own motors
and control systems in-house. This reveals one of the primary reasons Boston Dynamics lost its
dominant lead. For over a decade, Boston Dynamics relied on custom hydraulic motors for the original
Atlas. Hydraulics offered unmatched power density and massive torque, but they were extremely
expensive, prone to leaks, and complex to maintain. Most importantly, hydraulic systems were virtually
impossible to mass-produce at a low cost. But still, hydraulics was the only technology capable
of providing the power density required for the extreme jumps and flips. Industry shifted when
Wang designed an electric motor during his master's degree at Shanghai University. Wang replaced heavy,
expensive hydraulics and complex industrial servos with low-cost, high-torque electric motors.
After founding Unitrix in 2016, this design evolved into fully integrated proprietary
electric actuators produced entirely in-house. Boston Dynamics eventually made a pivot to all
electric motors. While competitors engineered electric actuators from day one to achieve
price points between $15,000 and $50,000, Boston Dynamics only retired its hydraulic Atlas in April
2024. This late transition delayed Boston Dynamics' commercial deployment. When it comes to mobile
four-legged robotics, Unitrix holds roughly 55% of all units shipped, while Boston Dynamics holds
approximately 35%. In the emerging humanoid market, Unitrix leads in unit sales with around 60%
of global shipments, while Tesla, Agility, and Figure lead in enterprise commercial logistics pilots.
Another factor that impacted Boston Dynamics' growth was repeated corporate ownership transitions
and shifting strategic directions. Driven by advances in its four- and two-legged platforms,
Boston Dynamics moved away from DARPA defense contracts and was acquired by Google in 2013
alongside seven other robotics startups. The goal was to build next-generation autonomous hardware,
but those plans never fully materialized. Google sold Boston Dynamics to Japan's SoftBank in 2017
due to a lack of near-term commercial revenue, which clashed with the financial discipline
imposed by Alphabet's restructuring. Then, facing financial pressure from losses in other venture
investments, SoftBank sold Boston Dynamics to Hyundai in 2021. Hyundai was looking to diversify
its mobility operations, and acquiring Boston Dynamics gave them world-leading intellectual
property. However, each change in ownership brought changing commercial mandates, which
delayed a unified focus on mass manufacturing. Finally, Boston Dynamics built its reputation
using classical control theory, focusing on precise physics, modeling, and pre-calculated
trajectory optimization. While this allowed for highly repeatable motion, it required manual
engineering effort for every new behavior added to the robot.
Competitors like Tesla, Figure, and Unitry adopted artificial intelligence from the start,
utilizing deep reinforcement learning and visual language action models. Instead of hard-coding
every joint movement, these robots learned tasks directly from human videos or in virtual
simulations. Today, the robotics race is no longer defined strictly by physical mobility or agility,
but by general-purpose intelligence, the ability of a robot to learn new tasks autonomously through
AI. In this new arena, Boston Dynamics is a contender rather than the sole leader. Tesla
and Figure have strong backing from frontier AI initiatives like XAI and Anthropic, while Unitry
is preparing for its IPO on the Shanghai Stock Exchange, with strong backing from leading Chinese
tech firms including DeepSeek. Despite these hurdles and seeding its absolute market lead,
Boston Dynamics remains uniquely positioned to capture the enterprise market.
While Chinese competitors like Unitry lead in low-cost volume and Tesla and Figure lead in
AI software integration, Boston Dynamics combines decades of unmatched mechanical engineering
with Hyundai's massive automotive manufacturing infrastructure, supply chain scale, and capital
investment. Potential policy restrictions or bans on Chinese-made autonomous hardware by
Western governments over national security concerns could create a significant market opportunity for
Boston Dynamics. To compete directly with Tesla and Figure, Hyundai and Boston Dynamics have
established strategic partnerships with Google to integrate state-of-the-art vision language
action foundation models directly into Boston Dynamics hardware. This shift is clearly visible
in the latest version of Atlas Humanoid, officially unveiled at CES 2026. This all-electric Humanoid is
engineered specifically for factory floors and warehouse workflows. Boston Dynamics will rely
heavily on Hyundai's automotive plans for deployment and manufacturing, targeting commercial delivery by
2027. So, what do you think? Will Boston Dynamics return to its former glory as the undisputed
leader or will new players like Unitry, Tesla, and Figure dominate the future of robotics?
Let us know in the comments below. Anyway, thank you for watching until the end.
I'm Rafa and I'll catch you in the next one. Have a good one.
