Meta Humanoid Robot Prototype: Home Research, Not a Product

Meta has reportedly built a humanoid robot prototype with the Chinese robotics company Fourier, aimed at exploring how robots might work in homes. The details come from a report by The Information, as summarized by Interesting Engineering. Meta says this is research only and not a product.

What was reported

According to the report, the robot is internally called Max. It was built in about three months, using customized hardware based on Fourier's existing robot platforms. It stands roughly 4.11 feet (1.5 meters) tall and weighs 105 pounds (48 kilograms). It reportedly runs on NVIDIA's Jetson Thor computing platform.

The article says Max is not just a relabeled version of Fourier's GR-3 humanoid. The Information describes it as a custom platform drawing on engineering from three generations of Fourier robots. Reported changes include different body proportions, changes to the head, a display, and a different camera layout. The robot has five-fingered hands meant to handle small objects such as bottles and toys, chest-mounted speakers, and a face that can show expressions.

What Meta wants it to do

The reported design requirements focus on how the robot deals with people, not on chores. They include:

• Recognizing spoken commands from about five meters away.
• Interpreting gestures such as waving from about three meters away.
• Keeping eye contact during an interaction.
• Signaling its state, such as listening, processing, speaking, or unsure about a task.

The article is explicit that these are development targets, not capabilities Max has been independently shown to have. There is no mention of folding laundry, loading a dishwasher, or cleaning anything. Handling a bottle or a toy is the closest the report gets to a household task.

Max 2 and the hiring

Work on a second version, Max 2, reportedly began in September. Planned changes include a redesigned torso, improved hands, a projector, and extra sensors to help the robot navigate stairs and understand its surroundings.

The Information also reported that Meta has been recruiting robotics engineers in Shanghai and Shenzhen, in roles covering prototype development and preparation for potential large-scale manufacturing. Separately, the article says Meta acquired Assured Robot Intelligence, a company focused on robot learning and AI that helps robots understand environments and perform physical tasks. Meta CTO Andrew Bosworth has reportedly discussed humanoids internally as a possible physical form for Meta's Muse AI agent.

What Meta says, and what is unknown

A Meta spokesperson told The Information that the Fourier partnership is limited to research and experimentation, and that Meta has no plans to sell the robots developed through it. The report therefore describes an internal effort, not a confirmed commercial program.

That sits a little awkwardly next to the reported manufacturing-related hiring, and the article does not explain the gap. It may simply reflect early planning, but that is a reading, not something the report states.

Several things a buyer would want to know are not in the article:

• Any price.
• Any release date or sales plan.
• Whether Max has performed any tasks in a real home.
• How it would be kept safe around children, pets, or older people.
• What the cameras and microphones would mean for household privacy, or where that data would go.

The article itself notes that questions remain about whether such robots can handle complex household tasks and operate safely around people.

What it means for your home

For now, this is a signal about where a large consumer technology company is looking, not a product you can plan around. Meta says it will not sell these robots, and nothing in the report suggests a consumer version, a price, or a timeline.

The emphasis on voice, gestures, eye contact, and status signals is still worth noting. A robot sharing a kitchen with people needs to show what it is doing and respond to a wave or a spoken request from across the room. Those are reasonable priorities, but they are goals on a requirements list, and a robot that follows them well is not the same as one that can reliably tidy a living room.

If you are weighing when a robot might help with chores, treat this as one more early-stage research effort. Look for evidence that matters: unedited demos in ordinary homes, clear statements on whether a human is remotely operating the robot, published safety testing, and plain answers on data collection.

More coverage of humanoid and home-robot developments is on our blog.