After dialing down its aggressive push into the Metaverse, Meta is now turning its attention toward artificial intelligence and humanoid robotics. Led by Mark Zuckerberg, the company is accelerating efforts to build advanced AI systems that can power robots capable of understanding and interacting with humans.
In its latest move, Meta has acquired Assured Robot Intelligence (ARI), a startup focused on developing AI models for robotic systems. While financial details of the deal remain undisclosed, the acquisition signals a clear shift in Meta’s long-term strategy toward robotics and AI-driven platforms.
A Meta spokesperson told TechCrunch, “We acquired Assured Robot Intelligence, a company at the frontier of robotic intelligence designed to enable robots to understand, predict, and adapt to human behaviours in complex and dynamic environments.”
This acquisition comes shortly after reports that Meta has slowed down development of its Metaverse platform, including limiting updates to Horizon Worlds. Instead, the company appears to be channeling resources into emerging AI technologies such as large language models, intelligent agents, and robotics.
ARI’s team, including co-founders Xiaolong Wang, Lerrel Pinto, and Xuxin Cheng, is expected to join Meta’s Superintelligence Labs (MSL). This division is considered the company’s core AI research unit and is currently led by Alexandr Wang following Meta’s earlier $14.3 billion acquisition of Scale AI.
The incoming ARI researchers bring strong academic and industry backgrounds. Xiaolong Wang has worked with Nvidia and the University of California, San Diego, while Lerrel Pinto has held teaching positions at New York University and previously co-founded Fauna Robotics. Their combined expertise is expected to strengthen Meta’s robotics capabilities significantly.
The new team will also collaborate with Meta Robotics Studio, an initiative launched to develop foundational technologies for humanoid robots. Reports suggest that Meta has been quietly investing in robotics for several years, exploring both hardware and software solutions aimed at consumer and enterprise use.
One of Meta’s long-term ambitions is to build a universal platform for humanoid robots—something akin to an operating system that manufacturers can adopt. The idea mirrors how Android transformed the smartphone ecosystem by offering a common software foundation.
Meta CTO Andrew Bosworth hinted at this vision in 2025, noting that the company aims to create software that other robotics firms can license. Such a platform could standardize how robots perceive, learn, and interact with their environment, potentially accelerating adoption across industries.
However, Meta is not alone in this race. Tesla is advancing its Optimus humanoid robot, while Google is developing its Gemini Robotics models. Meanwhile, Chinese firms like Unitree are rapidly innovating, with some robots reportedly approaching human-level speed and agility benchmarks.
As competition intensifies, Meta’s pivot highlights a broader industry trend: the convergence of AI and robotics as the next frontier in technology. While the Metaverse may have taken a backseat, Meta’s renewed focus suggests that intelligent machines could define its future.
