Dynamic Creatures: robots with character. Our investment in Dynamic Creatures. Imagine a small moment at a theme park. A five-year-old hangs back from a robotic character, curious and a little afraid. The character waves, hits its mark, and continues the show. Because the performance is fixed, the character cannot notice her hesitation, turn toward her, or wait until she is ready. That nuanced distance between a performance and an encounter is the problem Marc and Farbod started Dynamic Creatures to solve. Robots can be both specialized, intelligent, and humanist, all at the same time. The team is emerging from stealth as Boston Dynamics' official entertainment and hospitality robotics partner and is developing guest-facing character experiences with a large theme park and a retailer. Kindred Ventures invested because their work makes a larger idea concrete: Physical AI in the consumer space will be judged both by what a machine can do and by how clearly its movements and personality communicate intent to the people around it. We at Kindred Ventures are excited to co-lead the seed round for the Dynamic Creatures team alongside our friends at Eniac Ventures, with participation by Heliad, Sunshine Lake, and BlueGrass Ventures, as well as angel investors Rich Wong (Accel)Marc Raibert (Founder of Boston Dynamics), and Lukas Ziegler. Between animatronics and industrial robotics. Most guest-facing robots sit at one of two poles. Task-oriented service robots guide, inform, or deliver, often a function with a chassis. Fixed animatronics bring extraordinary craft and character to carefully authored scenes; Dynamic Creatures is building in the space between them. Think mobile characters behaving with more generalized intelligence and designed for face-to-face engagement in live, variable encounters. At a high level, responsive character robotics is a closed-loop problem. Real-time perception informs behavior; planning and control express the response through motion; new observations begin the cycle again. Guests never see those handoffs, but they feel the result in a pause, a turn, or a gesture that lands a beat too late. In foundational human-robot interaction research, predictability and legibility are distinct, according to Carnegie Mellon's groundbreaking paper. Predictable motion follows the path an observer expects. Legible motion makes the robot's intended goal easier to infer; a character needs both clarity and coherence. Guests should be able to read what the robot is doing without the motion breaking the persona. A robot can complete the right motion and still miss the moment. SnowJay: Character is built all the way down. SnowJay is an embodied-character robotics platform - the operating system for interactive characters - integrating real-time perception, learning-based control, adaptive behavior, and character design and expression for guest-facing environments. As Farbod puts it, "Robust and fluid movement, perception and behavior have to work as one system for a character to engage naturally in the physical world." What is enabled by this unique type of platform? Costume-aware robotics means the exterior of the robot must preserve sensing, actuator clearance, maintenance access, center-of-mass assumptions, thermal margins, and safe clearances around guests. For example, a heavy, plush, childlike character should not move like an exposed industrial quadruped - the workflow combines whole-body motion planning with conventional character animation, while also emphasizing the hardware's physical proportions and motion details as one believable system and user experience. This has been explored initially by Disney and Boston Dynamics, as well as companies based in China. One early modular approach was Boston Dynamic's Spot Sparkles which was an efficient platform costume on a general base robot hardware. A more complex development effort was taken on by Disney's Groot/Kiwi project, creating a native robotic character, with hardware, thermal layout, motion, surface treatment, and costume designed together. They envisioned robots whose silhouette, skin, flight suit, cooling path, locomotion, and animated performance were designed as one character system - not a machine just wearing a costume. Perhaps one of the most difficult distillation of the problem is this: an expressive gesture still has to be physically achievable, repeatable, and appropriate to the contextual physical space and humans interacting around the robot.A graceful motion delivered a beat too late or awkward body language can break the illusion. A face-to-face encounter has no loading spinner. Marc and Farbod: Spinning out to speed up. Marc and Farbod have embarked on this mission to marry animatronics and robotic intelligence by spinning out of Boston Dynamics to start Dynamic Creatures. Together, they bring the two disciplines the product demands. The story for us getting to know Marc and Farbod started at CES earlier this year during the live reveal and demonstration of the Atlas humanoid at the Boston Dynamics booth. After founding two startups and working at Google, Marc served as Chief Strategy Officer at Boston Dynamics and helped turn the research lab into a global commercial robotics business. The grand task at hand there was to build out the infrastructure needed to deploy and service advanced robotic systems at scale. Working closely with engineering teams and customers, Marc figured out the complex structure and process that decides whether a robot remains a demonstration or becomes a true business. His co-founder, Farbod,served as a research scientist leading at team at the Robotics and AI Institute (RAI) whose work has advanced learning-based control. His research spans agile humanoid behavior on Boston Dynamics' Atlas, high-speed quadrupedal locomotion, and the transfer of complex skills from simulation to physical machines. If a character has to move naturally through a lobby full of luggage and toddlers, that research lineage matters. In recent research he co-authored, motion capture, monocular video, and animation became inputs for policies that transferred from simulation to physical robots. Much of these concepts eventually inspired the thinking toward creating SnowJay - in particular,translating expressive motion into embodied skills a physical machine can perform. Marc describes the goal as creating "small moments of wonder that people remember." Though the phrase is simple, the engineering beneath it runs much deeper with new complexities that are unsolved, and need novel solutions. We see the timing in a convergence: established mobile robotics platforms, generalized intelligence models capable of multi-skill reasoning, learning-based control, real-time perception, and new methods for translating human and animated motion into physical behavior. Together, those advances make a long-held creative ambition more commercially plausible: a character that can meet a guest in the moment. Making wonder deployable: The commercial model. Dynamic Creatures is pursuing two product paths. For platform-based characters, the team starts with Boston Dynamics or other existing robotic systems, then integrates control, perception, behavior, character design, and costuming. The company says these characters can be deployed in months, and leasing gives operators a more flexible path to adoption without a large upfront capital commitment. Fully original robots follow a different path, with hardware designed around the character, story, environment, and creative requirements. Together, the two paths give Dynamic Creatures a way to turn deep robotics capability into a repeatable product without reducing every character to the same machine. Theme parks, resorts, museums, and other destination venues are a rigorous proving ground for this work. Close human proximity, irregular traffic, long operating hours, and service requirements leave little room for brittle integration. These venues also understand the value of memory. In entertainment and hospitality, the encounter is often the product. Robotics already creates value by moving goods and completing tasks. Dynamic Creatures is broadening the case toward the experiences machines can create. In spaces built around people, timing, distance, and behavior shape whether someone can understand a robot's intent and feel at ease around it. Character gives that behavior a coherent expressive frame. Dynamic Creatures is competing for two scarce things: attention and emotional connection. As robots begin to look back at us, how they move, wait, and respond will shape whether we welcome them. A wave returned at exactly the right moment may last only seconds. The memory can last much longer. We're excited to support Marc and Farbod on their unique mission to create beloved robots that interact and move with humans in a safe, useful, and delightful way.