Marc Raibert (born December 22, 1949) is an American roboticist renowned for pioneering dynamic legged robotics and founding Boston Dynamics, a leading company in advanced mobile robots capable of navigating unstructured environments.[1][2]Raibert earned a B.S. in electrical engineering from Northeastern University in 1973 and a Ph.D. from the Massachusetts Institute of Technology (MIT) in 1977, where his dissertation focused on motor control and learning through state-space models.[3][4] As a graduate student at MIT over 50 years ago, he began his career by developing software for robot manipulator dynamics, laying early foundations in robotic control systems.[2]Early in his academic tenure, Raibert served as an associate professor of computer science and robotics at Carnegie Mellon University (CMU), where he established the Leg Laboratory in 1980 to explore bio-inspired locomotion.[1] He later became a professor of electrical engineering and computer science at MIT from 1986 to 1995, founding the MIT Leg Laboratory and advancing research on self-balancing, one-legged hopping robots that demonstrated principles of dynamic stability.[3][2]In 1992, Raibert spun off from the MIT Leg Lab to co-found Boston Dynamics in Waltham, Massachusetts, initially as a simulation and robotics firm; he served as president and CEO from 1992 until 2019, transitioning to chairman.[5][2] Under his leadership, the company developed groundbreaking robots, including BigDog (2005), a quadruped for rough terrain funded by DARPA; Atlas (2013 onward), a humanoid robot excelling in agility and balance; Spot (2019, commercially launched 2020), a versatile four-legged inspector; Handle (2017), a mobile manipulator; and Stretch (2021), a warehouse automation robot.[5][1] These innovations emphasized robust, animal-inspired mobility, influencing fields from military applications to industrialautomation.[2]In 2022, following Hyundai Motor Group's acquisition of Boston Dynamics, Raibert founded and became executive director of the Boston Dynamics AI Institute, a research organization dedicated to advancing artificial intelligence for physical systems. In February 2025, the AI Institute partnered with Boston Dynamics to advance AI capabilities for the Atlas humanoid robot.[2][6] His work has earned numerous accolades, including election to the National Academy of Engineering in 2008 for contributions to multi-legged robots, the IEEE Pioneer in Robotics and Automation Award in 2022, the Joseph F. Engelberger Robotics Award in 2022, and the IEEE Robotics and Automation Award in 2025.[1][2] He was also named one of TIME's 100 Most Influential People in AI in 2023 for driving the integration of AI with physical robotics.[2] As a founding fellow of the Association for the Advancement of Artificial Intelligence, Raibert's career spans over four decades, fundamentally shaping the evolution of autonomous, dynamic machines.[2]
Early life and education
Birth and family background
Marc Raibert was born on December 22, 1949, in the United States.[1]Public details about Raibert's family background and early life are limited. His father aspired to be an aerospace engineer but pursued other paths; he maintained a home workshop filled with tools and machines, fostering Raibert's early interest in building and engineering.[7] It is known that his wife and daughter are both alumni of Northeastern University, reflecting a family connection to the institution where Raibert began his academic journey.[8]
Academic degrees and early research
Marc Raibert earned a Bachelor of Science in Electrical Engineering from Northeastern University in 1973. He continued his studies at the Massachusetts Institute of Technology (MIT), where he obtained a PhD in Electrical Engineering and Computer Science in 1977.[3][9]Raibert's doctoral dissertation, titled Motor Control and Learning by the State Space Model, was completed at MIT's Artificial Intelligence Laboratory. The work introduced a computational framework for addressing challenges in motor control, learning, and equivalence, utilizing a state space formulation to represent sensorimotor processes. Key concepts included a set of movement primitives that could be combined to generate diverse actions, with learning achieved through a reinforcement procedure that adjusted parameters based on performance feedback.[10][11]This early research, influenced by the interdisciplinary environment of the MIT AI Lab and mentors such as Berthold K. P. Horn, emphasized dynamic modeling of biological movement control, providing foundational ideas for stability and adaptation in complex systems. Raibert's simulations of motor behaviors in the thesis demonstrated how state-based representations could enable robust control strategies, paving the way for his subsequent explorations in robotics.[10][12]
Academic career
Role at Carnegie Mellon University
After earning his PhD from MIT in 1977, Marc Raibert joined Carnegie Mellon University (CMU) as an associate professor in the Computer Science Department and the Robotics Institute in 1980.[13] In this role, he focused on advancing the emerging field of mobile robotics, particularly emphasizing dynamic control systems for legged machines, which built on his dissertation work in motor control and manipulation.[14] His appointment came shortly after the establishment of CMU's Robotics Institute in 1979, where he contributed to its early development as a hub for interdisciplinary robotics research.[15]In 1980, Raibert founded the CMU Leg Laboratory, a pioneering research group dedicated to exploring the principles of legged locomotion through dynamic stability rather than static balance.[16] The lab's initial projects centered on one-legged hopping robots, such as the planar Hopper prototype, which demonstrated controlled hopping and landing to maintain upright posture without relying on multiple legs for support.[16] A key advancement was the 3D one-legged hopper (1983-1984), which achieved autonomous hopping and running at speeds up to approximately 2.2 meters per second, along with precise turning maneuvers in three-dimensional space without external support.[17] These experiments highlighted the challenges of dynamic balance, where robots had to actively manage momentum and external forces during motion, laying groundwork for more complex multi-legged systems.[18]Key prototypes from this period included early dynamically stable machines capable of demonstrating running gaits, such as planar biped models that executed bounding or alternating leg patterns while regulating speed and height, and the initial development of the quadruped with trotting gaits in 1984.[18] Raibert's seminal publication, "Dynamically Stable Legged Locomotion" (1983), detailed these innovations, including control strategies for hopping height, forward velocity, and body attitude, which became foundational for subsequent robotics research.[18] Through the Leg Lab, Raibert collaborated closely with graduate students, mentoring them on hands-on prototyping and control algorithms, which helped shape CMU's robotics curriculum by integrating practical dynamic locomotion topics into coursework and seminars.[19] This emphasis on experimental validation influenced the training of a generation of roboticists at the institution.[20]
Establishment of MIT Leg Laboratory
In 1986, Marc Raibert joined the Massachusetts Institute of Technology (MIT) as a professor of electrical engineering and computer science and a member of the Artificial Intelligence Laboratory.[3][13] This appointment marked a pivotal transition in his academic career, building on his prior work at Carnegie Mellon University (CMU), where he had founded the Leg Laboratory in 1980 as an early precursor to advanced dynamic robotics research.[16]The Leg Laboratory was relocated and expanded at MIT in 1987, enabling Raibert to scale up investigations into legged locomotion with greater resources and interdisciplinary collaboration.[16] This expansion continued and advanced focus on multi-legged configurations, including the quadruped—a four-legged running machine designed to mimic animal-like gaits. The quadruped demonstrated trotting, pacing, and bounding motions, achieving stable running speeds and seamless transitions between gaits by 1987, which highlighted the lab's progress in scaling control algorithms across multiple limbs.[16][21]Central to the lab's research during this period were control strategies for maintaining stability in dynamic environments, emphasizing decoupled subsystems for locomotion. Vertical force control was employed to regulate hopping height and ensure consistent ground reaction forces during stance phases, while body attitude adjustment utilized differential leg thrusts and hip torques to stabilize orientation and counteract disturbances.[18] These strategies, generalized from one-legged to multi-legged platforms, drew inspiration from biomechanical data on animals like cats and humans to achieve symmetric, energy-efficient motion.[16]
Professional career
Founding and leading Boston Dynamics
In 1992, Marc Raibert founded Boston Dynamics as a spin-off from the MIT Leg Laboratory, where he had previously directed research on dynamic legged systems; the new company initially concentrated on advanced simulation tools and robotics research and development to bridge academic innovations with practical applications.[5][22] Raibert served as CEO from 1992 until 2019, guiding the organization through its evolution from a research-focused entity to a leader in mobile robotics, securing substantial funding from the Defense Advanced Research Projects Agency (DARPA) and other government and private sources to support projects spanning military logistics and civilian automation.[23] He became chairman in 2020 and held that role until 2022.[2] Under his leadership, Boston Dynamics emphasized robust, terrain-adaptive machines, leveraging early DARPA contracts to pioneer technologies that could operate in unstructured environments beyond the capabilities of wheeled or tracked vehicles.[24]Key milestones during Raibert's tenure highlighted the company's progress in legged robotics. In 2005, Boston Dynamics unveiled BigDog, a quadruped robot capable of carrying heavy loads across rough terrain at speeds up to 4 miles per hour while maintaining balance on slopes and uneven ground, funded primarily by DARPA for potential military resupply roles.[25] This was followed in 2013 by Atlas, an advanced humanoid robot developed specifically for the DARPARobotics Challenge, designed to perform complex manipulation tasks in disaster-response scenarios, such as navigating debris and using tools.[26] In 2017, the company introduced Handle, a hybrid wheeled-and-legged robot optimized for warehouse environments, enabling efficient palletizing and material handling with a payload capacity of up to 15 kilograms per arm.[27] These developments underscored Raibert's vision of integrating dynamic control systems—rooted in his Leg Lab work—for real-world deployment.[28]Raibert's strategic oversight also facilitated Boston Dynamics' shift toward commercialization amid multiple ownership changes. Acquired by Google in 2013 to bolster its robotics portfolio, the company retained operational independence under Raibert, allowing continued innovation in advanced mobility.[29][30] In 2017, SoftBank purchased the firm from Alphabet (Google's parent), providing resources to accelerate product maturation without military constraints.[31][32] By 2019, this culminated in the commercial launch of Spot, a versatile quadruped robot for industrial inspection and data collection, marking Boston Dynamics' entry into the consumer market with over 400 units deployed initially.[33][34] The 2021 acquisition by Hyundai Motor Group for $1.1 billion further aligned the company with automotive and logistics applications, where Raibert actively championed the transition from prototypes to scalable, revenue-generating products before transitioning from chairman in 2022.[35][36]
Creation of Boston Dynamics AI Institute
In 2022, following Hyundai Motor Group's acquisition of Boston Dynamics in 2021, Marc Raibert transitioned from his role as chairman of the company to establish the Boston Dynamics AI Institute as an independent research organization dedicated to advancing artificial intelligence in robotics.[37] The institute, headquartered in Cambridge, Massachusetts, received an initial investment exceeding $400 million from Hyundai and Boston Dynamics to support long-term fundamental research.[38] Raibert serves as executive director, leveraging his decades of leadership at Boston Dynamics to guide the institute's focus on integrating AI with robotic systems for real-world applications.[39]The institute's mission centers on tackling core challenges in AI-robotics integration, including perception, manipulation, and human-robot interaction, to develop smarter autonomous systems capable of operating in complex, unstructured environments.[40] Key research areas encompass cognitive intelligence for generalization and planning, athletic intelligence for dynamic locomotion and balance, and organic hardware design mimicking human capabilities, all aimed at creating agile, perceptive, and safer machines.[38] Initial projects have emphasized reinforcement learning techniques; for instance, in January 2024, the institute demonstrated advancements in whole-body manipulation using its Spot robot to handle heavy tires through sampling-based optimization and learning algorithms.[41]To accelerate progress, the Boston Dynamics AI Institute—now operating as the Robotics and AI Institute (RAI)—has pursued collaborations with Boston Dynamics and academic institutions.[42] A notable partnership announced on February 5, 2025, involves applying reinforcement learning to enhance the electric Atlas humanoid robot, targeting sim-to-real transfer for agile locomotion, loco-manipulation tasks like operating doors and levers, and full-body contact strategies for dynamic behaviors.[42] Raibert has emphasized that this work will expand humanoid capabilities by broadening skillsets and optimizing skill acquisition processes.[42] The institute continues to seek partnerships with universities and corporate labs to address ethical considerations and societal impacts of advanced robotics.[38]In 2025, the RAI Institute expanded its initiatives, including a collaboration with ETH Zurich to attract top talent in AI-robotics research (announced September 2025) and exploration of generative AI for solving complex robotics problems, as highlighted in Raibert's July 2025 interviews. Additionally, the institute launched a public pop-up robot lab at CambridgeSide mall in summer 2025 to demonstrate robotics advancements and engage the community.[43][28][44]
Scientific contributions
Advancements in dynamic legged locomotion
Marc Raibert's research pioneered dynamic legged locomotion, a paradigm where robots achieve balance and mobility through active control mechanisms rather than relying on static stability, drawing inspiration from the efficient, energy-conserving gaits observed in animals. Unlike static walkers that maintain multiple points of contact to prevent tipping, dynamic systems operate near the edge of instability, using rapid sensorfeedback and actuator adjustments to sustain periodic motions such as hopping or running. This approach emphasizes the separation of control into independent components—height, speed, and orientation—allowing legged machines to traverse uneven terrain with agility and robustness.[45]Central to Raibert's framework are three key control principles applied across one-legged, bipedal, and quadrupedal systems. Vertical hopping control regulates body height by modulating legthrust during the stance phase to compensate for energy losses, ensuring consistent flight trajectories; for instance, in one-legged hoppers, this involves adjusting the impulse delivered by the leg to maintain a desired apex height. Forward speed control is achieved by positioning the foot relative to the body's center of mass during the flight phase, with the leg angle at touchdown determining propulsion—earlier placement accelerates forward motion, while later placement decelerates it. Attitude stability corrects for pitch and roll disturbances using hip torques or differential thrust in multi-legged configurations, enabling the body to remain upright even under external perturbations, as demonstrated in early bipedal prototypes that bounded at speeds up to 1.8 m/s. These principles generalize from the simplest one-legged hopper to more complex quadrupeds, where leg pairs coordinate to mimic trotting or galloping gaits.[18][45]Raibert's work evolved from theoretical models developed starting in 1980 at Carnegie Mellon University, where he analyzed ballistic hopping in two dimensions as part of his research in the newly established Leg Laboratory, to physical prototypes in the 1980s at MIT's Leg Laboratory. Early simulations explored resonant oscillations in spring-mass systems, leading to the construction of a one-legged planar machine in 1980 that achieved stable hopping at 1.2 m/s, followed by a three-dimensional version in 1983 capable of running at 2.2 m/s while balancing on one leg. By the mid-1980s, this progressed to bipedal and quadrupedal designs, including a 35 kg four-legged robot that executed dynamic gaits like pronking. A hallmark of stability in these systems is the approximate equality of stance and flight times, which simplifies control by aligning the periods of ground contact and aerial phases; mathematically, for a symmetric gait, stance duration t_s \approx t_f, where t_f = 2\sqrt{2h/g} (with h as hop height and g as gravity), ensuring the center of mass returns to the same horizontal position relative to the foot at each touchdown. This relation, derived from energy conservation in the flight phase, facilitated robust locomotion without precise timing synchronization.[46][45][18]Raibert's principles have profoundly influenced modern robotics, providing the foundational control strategies for agile maneuvers in advanced legged platforms, such as backflips and terrainadaptation in humanoid and quadruped robots. These concepts underpin the dynamic stability seen in systems like the Atlas humanoid and Spot quadruped, enabling real-world applications in inspection and exploration.[47][45]
Key patents in robotics
Marc Raibert has been a prolific inventor in robotics, contributing to numerous patents that underpin advancements in legged locomotion and actuator technologies, with listings on platforms like Justia Patents showing at least 18 granted under his name as inventor or co-inventor.[48] These inventions focus on enabling dynamic, stable motion in robotic systems through innovative control and mechanical designs.One of Raibert's influential patents is US8126592B2, titled "Actuator system," issued in 2012 and assigned to Boston Dynamics.[49] This patent describes an actuator subsystem for robotic or bionic linkages, featuring at least two actuators per joint—one for high-power output and another for compliance—allowing robots to achieve both forceful movements and adaptive responses to terrain. The design incorporates hydraulic or pneumatic elements that provide on-demand power while enabling energy storage and release, crucial for dynamic legged devices like quadrupeds to maintain balance during irregular locomotion. This compliant actuation approach has been integral to technologies enabling robots to navigate rough environments with stability.Another key patent is US6484068B1, titled "Robot apparatus and method for controlling jumping of robot device," issued in 2002 and jointly assigned to Sony Corporation and Boston Dynamics.[50] Co-invented by Raibert along with Robert Playter and others, it outlines a leg structure using a four-point link mechanism with a coil spring as an elastic member to ensure a linear trajectory for the foot during jumps. The innovation emphasizes independent control of vertical and horizontal forces: vertical control adjusts hopping height and stability via spring extension, while horizontal control manages forward momentum and directional changes, preventing tipping in multi-legged robots. This method supports hopping stability in devices simulating animal-like gaits, building on Raibert's foundational research in dynamic balance.Raibert's early work in the 1980s at Carnegie Mellon and MIT, including filings related to running machines, laid the groundwork for later patents on legged systems, though specific grants from that era are less documented in public databases. Overall, his portfolio, exceeding dozens of filings through Boston Dynamics, has protected core technologies such as hydraulic actuators in products like the BigDog quadruped robot, which relies on high-force, compliant systems for load-carrying over uneven terrain.[49] These patents have facilitated the commercialization of robust, agile robotics, influencing defense and exploration applications.
Awards and honors
Professional recognitions
Marc Raibert is a Founding Fellow of the Association for the Advancement of Artificial Intelligence (AAAI), an honor bestowed for his pioneering work at the intersection of artificial intelligence and robotics.In 2008, he was elected to the National Academy of Engineering for his biomechanically motivated analysis, synthesis, control, and application of multi-legged robots, highlighting his foundational research during his academic tenure at institutions like Carnegie Mellon University and MIT.[51]Raibert received the IEEE Pioneer in Robotics and Automation Award in 2022, recognizing his lifelong contributions to the development of dynamic legged locomotion and mobile robotics systems.Furthermore, Raibert has served in advisory capacities for academic and industry robotics initiatives, including as a mentor and expert for the MassRobotics Accelerator program, supporting emerging talent in the field.[52]
Recent accolades
In 2022, Marc Raibert received the Joseph F. Engelberger Robotics Award in the Technology category from the Association for Advancing Automation (formerly the Robotics Industries Association), recognizing his foundational contributions to dynamic legged robots and leadership in advancing robotics innovation through Boston Dynamics.[53][54]Raibert was named one of TIME magazine's 100 Most Influential People in AI in 2023, highlighted for his role in developing agile, animal-inspired robots that integrate artificial intelligence to perform complex tasks in real-world environments, underscoring his influence on the convergence of robotics and AI.[55]In October 2024, the IEEE Robotics and Automation Society announced that Raibert would receive the 2025 IEEE Robotics and Automation Award for pioneering and leading the field of dynamic legged locomotion, an honor he accepted at the 2025 International Conference on Robotics and Automation in Atlanta, reflecting his ongoing impact on mobile robotics amid his work at the Boston Dynamics AI Institute.[56][57]
Media and cultural impact
Public appearances and talks
Marc Raibert delivered a notable discussion at TEDxMIT in April 2022 alongside Daniela Rus, director of MIT's Computer Science and Artificial Intelligence Laboratory, where they explored the evolution of robotics and showcased capabilities of Boston Dynamics' humanoid robot Atlas, including dynamic movements akin to parkour demonstrated in the company's public videos from the 2010s.[58][1]Raibert has been a prominent keynote speaker at IEEE conferences, including the 2013 International Conference on Intelligent Robots and Systems (IROS) in Tokyo, where he presented updates on legged robots since the development of BigDog, and the 2023 International Conference on Robotics and Automation (ICRA) in London, focusing on the integration of AI to enable more autonomous robotic systems.[59][47]At EmTech MIT 2023, organized by MIT Technology Review, Raibert spoke on the future of AI and robotics, drawing from his experience founding the Boston Dynamics AI Institute to discuss challenges in creating intelligent, athletic machines capable of real-world tasks.[60]In September 2016, Raibert demonstrated the Spot quadruped robot at TechCrunch Disrupt SF, highlighting its agility in navigating obstacles and potential for commercial applications such as inspection and delivery in confined spaces.[61]Raibert provided insights into his career and the trajectory of robotics in a 2023 IEEE Spectrum interview, reflecting on transitions from dynamic locomotion research to AI-driven autonomy while leading the Boston Dynamics AI Institute.[47]In the inaugural episode of the Automated podcast, released in September 2025 by the Association for Advancing Automation, Raibert discussed his half-century of contributions to robotics, from early MIT work to current efforts in physical AI, emphasizing persistent innovation amid technological shifts.[62]In October 2024, at the AI + Robotics Summit in Washington, D.C., Raibert delivered a keynote titled "The Next 100 Years of Robotics," envisioning long-term advancements in AI-integrated mobile robots.[63]At DeepFest 2025 in Riyadh, Saudi Arabia, in February 2025, Raibert participated in a fireside chat on the future of robotics and AI, discussing ethical implications and breakthroughs in physical intelligence.[64]
Depictions in fiction and media
Marc Raibert and the robots developed under his leadership at Boston Dynamics have appeared in several documentaries highlighting advancements in robotics. In the 2021 60 Minutes segment "Boston Dynamics: Inside the workshop where robots of the future are being built," Raibert discusses the company's humanoid and quadruped robots, including Atlas, emphasizing their potential to mimic animal and human movement.[65] Similarly, Raibert featured in the 2013 PBSNOVA episode "Making Stuff Wilder," where he showcased dynamic legged robots and their engineering challenges.[66] A 2017 virtual reality documentary series by Within, titled "Boston Dynamics," explored the company's lab environment, with Raibert defending the realistic portrayal of robots' capabilities against public misconceptions of them as overly aggressive.[67]Boston Dynamics' demonstration videos have frequently gone viral, amplifying Raibert's influence in popular media. The 2017 video of the Atlas robot performing a backflip garnered millions of views on YouTube and was widely shared across platforms, with coverage often attributing the breakthrough to Raibert's foundational work in dynamic locomotion.[68] These clips, including earlier 2016 footage of Atlas navigating rough terrain, have shaped public perceptions of robotics, blending awe with occasional unease about agile machines.[69]Raibert's creations have inspired fictional portrayals of robots in media, particularly emphasizing quadruped and humanoid agility. The 2017 Black Mirror episode "Metalhead" features relentless robotic "dogs" hunting humans in a post-apocalyptic setting, directly inspired by Boston Dynamics' BigDog and Spot prototypes, as confirmed by series creator Charlie Brooker, who cited the company's viral videos as a key influence.[70] In the 2021 Netflix film Outside the Wire, director Mikael Håfström drew from Boston Dynamics' dancing robots to design the agile android protagonists, noting how the real demonstrations blurred lines between science fiction and emerging technology.[71] While Raibert himself has limited direct roles in fiction, his robots' feats have contributed to sci-fi tropes of versatile, terrain-adapting machines in films and books exploring human-robot coexistence.Raibert is frequently depicted in news media as a "robotics visionary," underscoring his cultural impact through Boston Dynamics' innovations. A 2023 TIME profile included him in the 100 Most Influential People in AI, praising his role in advancing practical, dynamic robots that bridge research and real-world applications.[55] Similarly, a 2024 IEEE announcement for his Robotics and Automation Technical Field Award described his "visionary approach" to legged locomotion as transformative, influencing global discussions on AI and automation.[72] These portrayals highlight limited direct fictional cameos but emphasize his indirect shaping of narratives around agile, autonomous machines.