Closing the infrastructure gap in physical AI

As the intelligence for robots, humanoids, autonomous vehicles, and drones moves from pilot to production, the infrastructure choices that companies make now determine how fast, and how safely, they can close the gap to full automation. IDC surveyed 603 industrial decision-makers to find out where their organizations stand in IoT-plus-AI maturity. In this Info Snapshot sponsored by Nebius and NVIDIA, IDC’s Dave McCarthy, Sarah Lee, and Carlos Gonzalez map the five stages of physical AI readiness and what it takes for organizations to advance toward trusted, autonomous execution.

Key findings:

• 60.2% of organizations have adopted AI that can either recommend next-best actions reliably (41.8%) or operate machines autonomously (18.4%)
• Adoption peaks at the prescriptive stage, where AI guides decisions but execution still depends largely on human validation

Inside the report:

• What separates organizations still focused on analysis and prediction from those putting AI-driven insights into real-time action
• Why cloud infrastructure for compute, storage, orchestration, and managed services make the leap from prescriptive AI to trusted autonomous execution possible