Meta has officially thrown its hat into the robotics ring, announcing an ambitious plan to develop a universal software platform that other companies can license to power their humanoid robots. The strategy, which mirrors Google’s incredibly successful Android model for smartphones, was detailed by Chief Technology Officer Andrew Bosworth and represents Meta’s next major “AR-size bet” on the future of technology.
“I don’t think the hardware is the hard part,” Bosworth stated, emphasizing that the real challenge lies in creating the intelligence to make robots useful. “The bottleneck is the software.”
To tackle this, Meta’s robotics division is developing its own prototype humanoid, dubbed “Metabot,” but its primary goal isn’t to sell hardware. Instead, the focus is on creating the underlying operating system and AI that any robotics company could adopt, provided their hardware meets certain specifications. This platform-first approach allows Meta to leverage its significant advancements in artificial intelligence, particularly its work on creating a “world model”—an AI capable of simulating the complex physics required for delicate tasks like grasping a glass of water without breaking it.
Meta enters a competitive landscape where Tesla’s Optimus robot and Google’s DeepMind-powered robotics are already making headlines, and Apple is reportedly exploring its own concepts. However, Meta’s software-licensing model could be a game-changer, allowing it to benefit from the growth of the entire industry, regardless of which hardware manufacturer ultimately succeeds. It’s a strategic pivot that seems to apply lessons learned from its costly investments in the virtual reality market.
With a recent real-world robot deployment in Germany providing valuable data and a commitment to invest billions in AI infrastructure, Meta is signaling its serious long-term vision: to be the intelligence that brings the next generation of robots to life in the physical world.