Your Physical AI lab
in a backpack
Four Highlights
Let’s start with FF Nova
Today’s launch is another important step for FFR in making EAI humanoid robots more accessible
FF Nova
For our developers and everyone in our education ecosystem, you can develop with Nova anytime, anywhere. In Phase One, we will offer a limited, invitation-only launch of 100 units for developers and educational institutions. The first 20 will be ready for delivery immediately.
- 01Reinforcement learning
- 02Autonomous intelligence
- 03Model-driven capabilities
- 04Native developer platform
- 05Skills Store
You don’t have to wait until you’ve grown up to shape the world.
Start now. Start with Physical AI. Start with FF Nova.
Let’s take a closer look at Nova
An edge computing platform that’s hard to match in its class, with a multimodal sensor array
SensingEdge computing and a multimodal sensor array
LiDAR, wide-angle vision, depth sensing, and a circular microphone array all work together. Recognition and obstacle avoidance can run locally, right on the robot. Nova can also understand indoor spaces, recognize family members, and locate where a sound is coming from.
MovementA true bipedal humanoid
Nova has 20 degrees of freedom across its body. It can walk, turn, wave—and if it gets knocked over, it can get back up on its own. It also gives you two ways to move. On two feet, it can walk toward you to interact. Seated on its detachable base, it can navigate autonomously.
Voice AI · AI AgentA body in the physical world
Nova connects to a large voice AI model for natural conversation and voiceprint recognition. It also allows developers to connect AI Agents—giving a cloud-based agent a body in the physical world.
Developer platformA fully open SDK
Nova has native support for the developer platform, with a fully open SDK, a ROS 2 development workflow, and ongoing updates over the air. We’ll open the full Sim2Real reinforcement learning pipeline and also provide a Skills Store and developer community, so that, in the future, what you build can be shared, reused, and monetized.
A complete Sim2Real workflow—from simulation training to deployment and validation on a real humanoid robot. An accessible, portable platform that keeps evolving with community contributions and new versions.
Consumer-level pricing
On-device inference and a complete set of developer tools
Carry it with one hand and ship it in a single box
Universities and research institutions
Nova is a Sim2Real experiment platform you can deploy at scale.
Schools and training providers
It brings the EAI Device, curriculum, cloud services, and toolchain together in one package. You can offer Physical AI courses with little R&D investment.
Students
It means learning the real loop of perception, decision, and action.
Your child’s first humanoid robot
America’s first foundational Physical AI education platform built around a true bipedal humanoid.
Others sell a robot. We deliver a learning path that educators can teach and assess
FF Nova Mini
It weighs less than one kilogram, with an aluminum-alloy frame that makes it easy for children to carry and use in class. It can walk on two legs with human-like movement, turn, stand up on its own, maintain dynamic balance, and even do a forward roll. It also supports visual block-based programming and hands-on teaching by demonstration. And it’s ready for a demo right out of the box.
Going forward, FF will work with Montessori organizations, early childhood education providers, and elementary schools to develop teaching kits, work cards, and observation and assessment tools. Our goal is a model for foundational Physical AI education that more schools and institutions can adopt and replicate.
One brain, plus four types of Skills
The brain handles understanding and decision-making. Skills turn those capabilities into services that can be taught, used, and built on
It provides voice AI model conversation, multimodal understanding, task decomposition, and motion control scheduling. One brain powers both Nova and Nova Mini, and it works with the Edu versions of every EAI Device FF has released.
Courses are delivered as Skills: you can install them, update them, and institutions can adapt them on their own.
Recognition and obstacle avoidance run locally, which keeps response times fast and works even when the network is down. Large models and training run in the cloud, so institutions don’t need to build their own GPU infrastructure.
The brain and the EAI Device together form a “prepared environment.” Courses are age-based, staged, and move at each child’s own pace. The robot shows the result of every action in real time, so children can find and correct their own mistakes.
Nova Physical AI Reinforcement Learning Developer Program
It supports reinforcement learning through the NVIDIA Isaac platform. We’re opening our developer community, too—with limited places, by invitation only
Nova opens a complete reinforcement learning pipeline: simulation models, actuator modeling, reward functions, training recipes, and policy deployment for real-robot validation.
At $2,490, Nova gives developers a true bipedal humanoid for training and validating their own policies—from bipedal walking and coordinated whole-body movement to recovery after a fall.
Run parallel training in simulation, tune rewards, apply domain randomization, export policies, rehearse on your computer, and deploy on Nova for real-world validation.
You can apply for an invitation code through the platform, or get one from an existing developer.
If your project stands out, it may receive support through our Developer Support Program. We want product sales and our developer ecosystem to grow together.
Whether you’re an AI algorithm engineer, a robotics developer, an educator, a maker, or a business partner with an innovative robotics application in mind, you’re welcome to join the Nova ecosystem.
- Get early access to Nova’s latest platform capabilities and developer tools
- Receive official technical support, training courses, and developer documentation
- Take part in developer challenges, hackathons, and innovation project incubation
- Explore innovative robotics AI applications with industry experts and partners
- Get opportunities for project showcases, business collaboration, and ecosystem support
Standardized Online. Scenario-Based Offline.
FF’s full line of Physical AI robots is now officially available through online and offline channels. We serve developers, universities and research institutions, schools, educational institutions, and families, with robot products, Skills, courses, and related services
FF’s full robot lineup will be available in stages through the FF website at FF.com, FF Mall, and our official Amazon store. Depending on each product’s policy, you can buy directly, make a reservation, or submit an application.
Through our direct sales team and partners, FF serves schools, school districts, universities, research institutions, vocational education providers, STEM training providers, and enterprise customers. We offer product experiences, demos, education solutions, pilots, and institutional procurement.
Phase One is invitation only for developers and educational institutions, limited to 100 units, with the first 20 available for immediate delivery. Apply at developer.ff.com; applications are reviewed on a rolling basis, and approved applicants can place an order. Phase Two opens to all developers and educational institutions. Phase Three is the consumer launch.
Your child’s first humanoid robot, and a walking programming platform. You can get started with Nova Mini on its own, or pair it with courses, Skills, and teacher tools for use at home, in school classrooms, and at STEM training providers.
It’s a two-month program with 16 live small-group classes. The course is all about making things: each class teaches the robot one new thing, and every assignment becomes a project, building up to a personal Physical AI portfolio.
With our full range of products, Skills, and courses, we’re helping more people go beyond using AI. They can learn, build, and create Physical AI.
Now it’s your turn to build
Today we’ve put a Physical AI lab in your backpack and a first humanoid robot in your child’s hands
Join the Developer Community developer.ff.com