The Next Storage Frontier: Physical AI Takes Center Stage

Shifting the focus beyond traditional data centers, Micron Technology's CEO Sanjay Mehrotra has highlighted "Physical AI" as the emerging powerhouse for future memory and storage demand. This signals a strategic pivot towards intelligent systems that interact directly with the physical world.

Defining Physical AI: From Theory to Tangible Impact

Physical AI refers to artificial intelligence embedded within physical entities, enabling them to perceive, learn from, and act autonomously in real-world environments. It represents a move from purely digital computation to integrated, real-world intelligence.

Autonomous vehicles are leading the initial charge. Level 4 and above self-driving cars typically require over 200GB of memory and multiple terabytes of storage—an order of magnitude greater than the needs of today's common L2+ or L3 semi-autonomous vehicles.

The Expanding Horizon: Beyond Cars to a Robotic Future

The deployment of Physical AI is expected to broaden significantly. Mehrotra anticipates its expansion into humanoid robots and other intelligent autonomous systems.

  • Humanoid Robots: Projected to have storage demands comparable to advanced autonomous vehicles, driven by needs for complex sensor data processing, environmental mapping, and continuous learning.
  • Diverse Autonomous Systems: This includes smart industrial equipment and advanced drones, all contributing to the growing ecosystem.

Growth will be fueled by a dual dynamic: a rapid increase in the number of these intelligent devices coupled with a substantial rise in the storage capacity each device requires. This combination is set to unlock a major new market segment.

Implications for the Storage Industry

Micron's analysis suggests that by the end of this decade, Physical AI could become a critical driver for storage demand. This outlook not only guides chipmakers' product development but also hints at a broader transformation across the smart hardware supply chain.

The evolution of future devices will depend on parallel advances in both processing power (the "brain") and storage capacity (the "memory") to support the real-time decision-making and adaptive learning essential for Physical AI. A hardware-powered intelligent revolution is on the horizon.