According to a 2025 report by the International Federation of Robotics, roughly 542,000 robots were installed in 2024, more than double the amount in 2014. And there’s a good chance a growing number of these robots can be traced back to Boston Dynamics, a robotics leader known for creating two- and four-legged robots powered by artificial intelligence. Some of its earliest prototypes went viral for their unprecedented balance and agility, setting even higher expectations for its industrial robots that are ready to get to work.
What to Know About Boston Dynamics
Boston Dynamics is a robotics lab known for researching and building autonomous mobile robots. The company has experimented with bipeds and quadrupeds over the years, more recently creating robots ready to work in industrial settings. Ultimately, Boston Dynamics envisions legged robots automating repetitive or difficult tasks to make workers’ lives easier — complementing rather than replacing human labor.
Marc Raibert, the company’s founder, envisions bipeds and quadrupeds being commonplace across industries that depend on manual labor — and it seems his vision is becoming a reality. Here’s an overview of Boston Dynamics, including how its robots work, how they’re used and how it plans to shape the future of robotics as it transitions from a research lab into a full-fledged enterprise.
What Is Boston Dynamics?
Boston Dynamics is a robotics company that develops mobile robots intended to automate dangerous or heavy-duty tasks in industry settings. It was founded in 1992 by Marc Raibert, a professor of electrical engineering and computer science at MIT. Raibert previously worked at MIT Leg Lab, which he founded in 1980 as a research hub dedicated to studying and building robots that could move using legs. His pivot to Boston Dynamics was the next step in achieving his dream of autonomous mobile robots.
Over the past three decades, Boston Dynamics has refined its bipedal and quadrupedal robots. The company made headlines in 2004 when it released BigDog, the first robot to step outside the lab. Since then, Boston Dynamics introduced its first biped in 2013 and its first commercialized offering in 2020, laying the groundwork for its slate of intelligent robots to become key members of today’s robot workforce.
Boston Dynamics’ Robot Fleet
Boston Dynamics now provides a lineup of several robots that contribute to the safety and well-being of industrial workforces.
Spot
Often referred to as a “robot dog,” Spot is a four-legged mobile robot designed to sense its surroundings and problem-solve on its own, navigating complex facilities and uneven terrain. Using laser-scanning technology, Spot compiles visual data and builds a digital twin of a worksite that users can reference to gather insights into everyday workflows, identify onsite risks and assess the overall health of their assets. Users can also choose how much control they exercise over Spot, manually operating the robot or granting it full autonomy.
Atlas
Atlas is a humanoid robot that stands upright and is built to handle hard manual labor. It uses a pair of arms to lift up to 110 pounds and carry up to 66 pounds for extended periods. Equipped with advanced AI, Atlas can quickly adjust to new environments, learn new skills and share its knowledge with other members of a fleet. It can even make its way to a charging station and swap out batteries on its own, allowing users to deploy and run Atlas around the clock with limited human supervision.
Stretch
Stretch is a robotic arm designed to perform repetitive tasks in warehouse settings. Wielding a gripper with rows of suction cups, Stretch can carry multiple packages at once and lift 50 pounds, moving hundreds of boxes per hour. It doesn’t need neat stacks of items to function properly, either, using sensors to handle boxes that have been haphazardly arranged or retrieve boxes that it accidentally knocked over. A long battery life enables Stretch to cover multiple shifts as well, sparing workers from sorting items in hot storage containers.
How Do Boston Dynamics’ Robots Work?
Boston Dynamics’ robots run on an interconnected ecosystem of hardware and software. Visual cameras, LiDAR sensors, thermal cameras and acoustic imagers all compile information about a robot’s surroundings, converting physical conditions into concrete data. Large behavior models can then combine this real-world data with simulations to develop data pipelines for training neural networks. This process facilitates reinforcement learning — learning through trial and error — which Atlas relies on to master behaviors and skills.
To scale this approach, the company provides teams with an orchestration platform called Orbit. Not only does Orbit enable users to coordinate actions across an entire robotic fleet, but it also compiles and analyzes visual data with its AI visual inspection (AIVI) models. These models are multimodal, so they can process images to glean insights into assets’ health and alert users of any issues early on. They’re also cloud-based, meaning they share data that can be used to improve them and develop future models.
At the same time, Boston Dynamics’ progress wouldn’t be possible without key partnerships. The company tapped Google to enhance its AIVI models with Gemini models, and it’s turned to FieldAI’s foundation models to improve its robots’ real-world performance. That said, Boston Dynamics has invested plenty of its own resources in an AI and robotics center near its headquarters in Waltham, Massachusetts, to solidify its leadership in mobile robots for years to come.
Boston Dynamics’ Primary Use Cases Across Industries
Boston Dynamics’ robots are taking on difficult tasks and working to make conditions safer across various industries, including the following:
- Manufacturing: Spot can compile data from facilities, including areas unsafe for workers. Teams can then monitor their assets, spot potential workplace dangers and conduct predictive maintenance as needed.
- Logistics: Stretch can take on the heavy lifting in warehouses, helping unload trailers and sorting boxes for specific orders. Meanwhile, Spot can patrol the area to inspect equipment and alert teams of any issues.
- Energy: Spot is designed to access challenging environments, from nuclear facilities to underground mines. It can provide visual data and analysis for teams to check the health of their facilities without putting workers in danger.
- Construction: Spot compiles images and videos of construction sites for workers to track their progress. Site leads can also send Spot to more hazardous zones to assess conditions before workers enter the area.
- Government: Government and law enforcement personnel can deploy Spot in high-risk situations like standoffs, package checks and burning buildings. It can then assess or de-escalate to ensure authorities don’t take unnecessary risks.
- Education: Spot is easy to set up and comes with an open platform, making it ideal for academic research. Teachers and professors can use the robot to discover new applications or train future robotics engineers.
- Security: As more AI companies erect data centers, they’re turning to Spot robots for 24/7 security amid fierce public backlash. Spot can monitor any unusual activity within data centers and detect perimeter breaches.
Ethical Policies and Governance
The promise of artificial intelligence and robotics is often overshadowed by the ethical questions they raise. After all, AI is often blamed for undermining the college pipeline that jumpstarts a white-collar career, and integrating it into robots may threaten blue-collar jobs as well. To make matters worse, companies have agreed to grant government access to powerful models as part of dubious deals that could accelerate the development of autonomous weapon systems.
It’s no surprise then that these technologies have struggled with negative perceptions. That’s why Boston Dynamics has attempted to distinguish itself from worst-case scenarios and define its robotics approach with a list of principles that can be summed up as:
- Automation should be used to “improve our lives” by taking on the more physically demanding and dangerous aspects of today’s jobs.
- Only people can navigate the “complexity and richness of real-world conditions,” so robots should be designed to complement workers, not replace them.
- To earn the trust of workers and the general public, robots must be “predictable, understandable, transparent and in service of human needs.”
- Robots should not be converted into autonomous weapons, and Boston Dynamics will never partner with organizations that have this goal.
- Lines around personal privacy and civil rights must not be crossed by robotics, given the risks computer vision and other technologies pose in public spaces.
- Partnerships with government agencies for “preserving safety and security” can be fruitful, as long as they agree to the previous two principles.
- Regulations focused on robotics safety benefit all parties involved, especially when various public and private stakeholders get to provide input.
Frequently Asked Questions
What does Boston Dynamics actually do?
Boston Dynamics is known for its research on autonomous mobile robots, developing bipeds and quadrupeds that demonstrate impressive balance, agility and coordination. The company has more recently designed robots that can automate challenging or repetitive work in industrial settings, such as sorting packages, lifting heavy objects and navigating hazardous sites. That said, its robots are meant to support human workers rather than replace them outright.
Who owns Boston Dynamics?
Hyundai Motor Group owns Boston Dynamics. It acquired the robotics company in 2021, securing an 80 percent stake. SoftBank Group owns the remaining 20 percent.
Can a normal person buy a Boston Dynamics robot?
No, Boston Dynamics’ robots are not intended for individual use. The company mainly sells its products to businesses in industrial settings and research institutions. Its robots also cost at least hundreds of thousands of dollars, preventing most people from even affording one.
