Problem

Robotics remains difficult and expensive for people who want to build useful physical applications. The open-source hobbyist community lacks robust, accessible platforms, while conventional research-grade systems can require substantial hardware, compute, integration, and robotics expertise. Innate’s core thesis is that robots should be open, programmable, and teachable rather than vertically controlled products that users can only operate in predefined ways.

The primary use case is therefore a developer, maker, or educator who wants to prototype physical AI without assembling a mobile base, arm, sensors, GPU, and software stack from scratch. The initial target is builders who can teach a robot new behaviors for applications such as home security, playing chess, handling objects, or other customized physical tasks; the economic promise is to make experimentation with general-purpose robotics practical at a consumer-accessible price rather than a laboratory-scale investment.

Product / Service

Innate’s product is MARS, an open-source general-purpose mobile robot with an arm, onboard compute, cameras, LiDAR, extension ports, and an agentic operating system. The current website describes it as a robot that works out of the box, with a programmable foundation agent called BASIC that can reason over goals, call skills, adapt in real time, and be extended with new behaviors, tools, sensors, and hardware. MARS is offered through a direct hardware model, with the current site listing a $1,999 preorder for a second batch and new units coming soon.

The software is the differentiator: users can prompt the robot, interact with it conversationally, teleoperate the arm to collect training data, and program or share skills. Innate says new physical skills can be trained in under 30 minutes of data collection, while its underlying system combines specialized models for manipulation, navigation, reasoning, planning, memory, and real-time interaction. The benefit is a relatively complete, modifiable platform for embodied-AI experimentation rather than a single-purpose appliance.

Market

Innate competes at the intersection of consumer robotics, physical AI, developer tools, and open-source robotics platforms. Its closest alternatives are not necessarily identical products: hobbyists can combine platforms such as the SO-101 or LeKiwi with their own computers and software, while mobile robots such as Unitree’s Go2 offer another lower-cost route to robotic experimentation. It also sits in the broader competitive field of general-purpose robotics companies such as Figure and General Robotics, although those companies pursue different hardware forms, customers, and levels of industrial scale.

The company appears to be in early commercial beta rather than pre-product. Innate reported shipping MARS to select early supporters, and the current site is accepting paid second-batch preorders; its YC profile lists the company as active, founded in 2024, with a team of seven. The available materials do not disclose customer counts, revenue, or a scaled deployment base, so the strongest evidence of traction is product availability, early shipments, an open-source developer/community strategy, and ongoing demand testing through preorders rather than demonstrated revenue scale.