Faraday Future Just Launched a Humanoid Robot for $2,490. The Price Changes Everything.

On October 15, Faraday Future (NASDAQ: FFAI) and its robotics affiliate FFR (NASDAQ: FFR) held a joint launch event in Los Angeles and introduced FF Nova, an entry-level humanoid robot priced at $2,490 for the device, with an optional $390 Developer Skills Package. The companies call it America's first humanoid robot priced below $2,500, and whether or not that superlative survives scrutiny, it pushes embodied AI hardware into laptop price territory for the first time.

The same event introduced the FF Nova Mini, a $299 bipedal humanoid aimed at children ages eight and up, a $499 "Physical AI Spark Program" curriculum, and a reinforcement learning developer platform built on NVIDIA Isaac. Alongside the hardware, the companies disclosed that FFR will acquire Faraday Future's robotics business in an all-stock transaction valuing it at roughly $200 million.

What the FF Nova actually is

Strip away the marketing language and the Nova is a developer-grade humanoid: 20 degrees of freedom across its body, true bipedal walking, the ability to turn, wave, and get back on its feet after being knocked over. It moves in two modes: walking on two legs to interact with people, or riding on a detachable base for autonomous navigation when interaction is not the point.

The sensor stack is serious for the price class: LiDAR, wide-angle vision, depth sensing, and a circular microphone array, feeding an edge computing platform the company says is hard to match in this price band. The robot connects to a large voice AI model for natural conversation and voiceprint recognition, and it supports AI agent integration, letting a cloud-based agent take a physical body in the real world.

For developers, FF Nova ships with an open SDK, a ROS 2 development workflow, and over-the-air updates. It connects to an end-to-end Sim2Real reinforcement learning pipeline (train in simulation, validate on hardware) and a Skills Store where developers can share their creations and, eventually, monetize them.

At $2,490, the robot costs less than the workstation used to program it. That is the price breakpoint embodied AI has been waiting for.

The rollout: 100 units first, consumers later

FF is launching Nova in three phases, and phase one is deliberately scarce. Invitation only, limited to 100 units, with the first 20 ready for immediate delivery. Applications run through developer.ff.com on a seven-day window, reviewed on a rolling basis, with universities, research labs, independent developers, and education innovators as the target cohort.

Pioneer Edition users get priority delivery, early access to the SDK and developer documentation, sample projects, official technical support, and a shot at having their creations featured in the FF Skills ecosystem. Phase two opens the robot to all developers and educational institutions. Phase three is the full consumer launch, which is where the sub-$2,500 price has to prove itself outside friendly lab conditions.

Why the price matters more than the specs

Robots do not change the world through specs. They change it through who gets to build for them. At $2,490, a humanoid robot is no longer a capital expenditure that needs a procurement committee. It is a line item a university lab, a high school robotics club, a startup, or an indie hacker can justify. That is what "a Physical AI lab in a backpack" means in practice: the addressable developer population goes from hundreds of well-funded labs to tens of thousands of curious people.

There is a catch, and it is worth stating plainly. Everything here comes from the companies' own announcement; independent testing, teardowns, and reviews do not exist yet. Faraday Future is best known as an electric vehicle company that has had a turbulent decade, and FFR was only renamed from AIxCrypto Holdings on September 30. The "first humanoid under $2,500 in the U.S." claim is FF's own billing, not an independently verified fact. And price alone does not make a robot useful: the software ecosystem, the reliability of the hardware, and the quality of that Sim2Real pipeline will decide whether Nova is a landmark or a curiosity.

The education bet may be the real story

The Nova Mini deserves as much attention as its big sibling. It is a true bipedal humanoid weighing under one kilogram, with an aluminum-alloy frame light enough for children to carry. It walks on two legs with human-like movement, turns, stands up on its own, maintains dynamic balance, and can even perform a forward roll. Programming is visual and block-based, with teach-by-demonstration: physically guide the robot and it learns.

The curriculum around it is the full package, not an afterthought: 16 live small-group classes over two months, lesson plans, experiment manuals, classroom kits, and teacher training. The $499 curriculum price includes an exclusive voucher toward the $299 device. FFR says it will work with Montessori organizations, early childhood providers, and elementary schools to build teaching kits, work cards, and assessment tools.

If physical AI follows anything like the trajectory of personal computing, the generation that grows up programming small robots is the generation that invents the field. A $299 bipedal robot on a classroom desk is how you find out.

What to watch next

Three things will tell us whether this launch is real or theater. First, independent reviews of the phase-one units: can the robot actually do what the announcement claims, day after day? Second, whether the developer ecosystem materializes: an open SDK means nothing without an active community, and the invitation-only Nova Developer Community is just recruiting. Third, whether the corporate machinery works: the FFR acquisition of FF's robotics business (FFAI to hold about 65.7% of FFR on a fully diluted basis, with an 18-month lockup, plus up to $10 million in PIPE financing) has to close cleanly.

Price breakpoints do not guarantee success. But they open doors that were closed before, and a door at $2,490 is open to a lot of people. The Physical AI era will not be built by the labs that could afford $20,000 robots. It will be built by whoever gets their hands on one cheap enough to break it, fix it, and teach it something new.