Meta Expands into Physical AI with Acquisition of Robotics AI Startup

• By John K. Waters
• 05/06/26

Meta Platforms has acquired Assured Robot Intelligence (ARI), a robotics artificial intelligence startup focused on humanoid systems, as the company expands its AI work beyond software and into models that could help robots operate in physical environments. Financial terms were not disclosed.

Meta confirmed the acquisition to The Wall Street Journal, saying ARI works on robotic intelligence intended to help robots understand, predict, and adapt to human behavior in complex settings.

ARI co-founder Xiaolong Wang announced the acquisition in a post on X, which read in part:

"When we started ARI one year ago, our mission was clear: achieve physical AGI. Through deep customer engagements and real-world deployments, it became clear to us that serving the massive opportunity ahead requires training a truly general-purpose physical agent.

"We believe this agent will be humanoid — and that scaling will come from learning directly from human experience, not teleoperation alone. Meta's ecosystem brings together the key components needed to make this vision possible. We will be joining Meta Superintelligence Labs (MSL) to help bring personal superintelligence into the physical world."

ARI's team, including co-founders Wang and Lerrel Pinto, will join Meta Superintelligence Labs, Meta's AI research organization. The group is expected to work on model capabilities for robot control, self-learning, and whole-body humanoid control.

The acquisition appears to be aimed less at buying a finished robot product than at adding robotics AI talent and technical expertise. ARI was developing foundation models for humanoid robots, including systems capable of supporting physical tasks such as household work.

The founders bring academic and industry experience in robotics and machine learning. Wang previously worked as a researcher at Nvidia and has been an associate professor at the University of California, San Diego, while Pinto previously taught at New York University and co-founded Fauna Robotics, a small humanoid robotics startup later acquired by Amazon.

The deal comes as Meta continues to increase spending on AI infrastructure and related research. The Wall Street Journal reported that Meta recently raised its projected 2026 capital expenditures by $10 billion, to a range of $125 billion to $145 billion, citing higher component costs and additional AI data-center spending.

Meta has also been shifting resources toward AI after years of heavy investment in augmented and virtual reality. The ARI acquisition suggests that Meta sees robotics as one possible extension of its AI strategy, although the company has not announced a consumer humanoid robot product or a timeline for one.

The acquisition fits a wider industry trend toward what is often called "embodied AI" or "physical AI." Large technology companies and startups are increasingly focused on models that can interact with the real world through robots, rather than only generate text, images, code, or video. In that model, the difficult problems include perception, dexterity, manipulation, navigation, safety, and adapting to unpredictable human environments.

The deal also reflects growing interest in the software layer of robotics. Instead of focusing solely on building robot bodies, companies are developing the "brains" that could enable robots to learn tasks, manipulate objects, and operate across different hardware platforms. That is still an early market, and it remains unclear whether general-purpose humanoid robots will become commercially practical in homes, warehouses, factories, or healthcare settings.

For Meta, ARI gives its AI organization a small, specialized robotics team at a time when rivals are also investing in physical systems. Amazon acquired Fauna Robotics earlier this year, while other startups are raising capital to build robot learning systems and humanoid platforms.

The commercial case remains uncertain. Humanoid robots face high costs, reliability problems, safety requirements, regulatory questions, battery constraints, and the difficulty of performing ordinary physical tasks consistently. Even advanced AI models can struggle when moved from digital benchmarks into homes, offices, or industrial sites.