Meta plans major humanoid robotics push, tapping ex-Cruise CEO and targeting Figure AI partnership

Summary

Meta is making a significant bet on humanoid robotics by forming a new team within Reality Labs and hiring Mark Whitten, the former CEO of Cruise, to lead the group. According to Bloomberg's Mark Gurman, the company is positioning itself as "the engine powering the humanoid market" rather than a robot manufacturer itself—at least initially.

Meta is in discussions with Figure AI and Unitree Robotics about potential partnerships. The company does not plan to build a Meta-branded robot that would directly compete with Tesla's Optimus, though executives have not ruled out doing so in the future. Meta's strategy centers on software and AI rather than hardware development.

The data advantage

Andrew Bosworth, Meta's CTO, argues that the company's existing investments across Reality Labs and AI position it well for robotics. Specifically, Meta believes its advances in hand tracking, low-bandwidth computing, and always-on sensors—paired with data collected from Quest augmented and virtual reality devices—could accelerate progress in humanoid robotics. The logic is straightforward: training effective humanoid robots requires massive amounts of sensor data on how humans move, grab objects, and perform dexterous tasks. That data almost doesn't exist in the broader training corpus because consumer devices have never been instrumented to capture it at scale. If Meta can leverage Quest's hand and body tracking across millions of devices, it gains a potentially unique dataset for training algorithms.

The valuation story

Figure AI is currently raising at a $33.9 billion valuation, a 15x jump from its last funding round. This creates a striking comparison: Figure, which has delivered robots to customers but at unclear scale and with uncertain real-world deployment, is being valued higher than Ford Motor Company. Ford carries $185 billion in annualized revenue and $11 billion in EBITDA but trades below Figure's valuation in public markets. The divergence reflects investor conviction that whoever wins humanoid robotics will command sustained premium valuations before the market commoditizes—a bet on manufacturing difficulty and first-mover advantage rather than near-term revenue.

The depreciation problem

A meaningful friction in consumer humanoid adoption is depreciation risk. If a homeowner buys a $30,000 Optimus robot and advances in the field accelerate quickly, the robot could depreciate $80,000 or more within a year as newer models ship. At that economics, a consumer faces a rational trade-off: pay for a teleop-controlled robot to fold laundry part-time and absorb massive depreciation, or simply hire a full-time employee. This suggests humanoids will likely see adoption first in military, industrial, and nuclear facilities—where a human operator cannot go—rather than in homes.

Market competition

Unitree Robotics, a Chinese manufacturer, is already selling humanoid robots directly to consumers via its website. This positions China to repeat its playbook in drones: flooding the market with low-cost hardware, potentially at losses, to establish dominance before competitors scale. The competitive intensity is real, even if Figure and other startups are attracting venture capital at valuations that imply winner-take-most outcomes.