With increasing compute power on the edge, the pendulum is swinging back towards smart clients that can run AI locally, potentially reducing the need for massive data center infrastructure for inference.

Donald Thompson, Distinguished Engineer, LinkedIn

In the Lunar Rover, the AI system needs to make decisions in real-time, within 100 milliseconds, so can’t rely on sending data to the cloud. Our autonomous driving system handles mapping, localization, path planning, obstacle avoidance, and path execution. 

Eliana Stefani, Autonomy & AI Research Engineer, Lockheed Martin (AI Infra Summit speaker 2025)

Our biggest challenge is verifying and validating the safety and reliability of AI/ML systems, especially for certification in commercial aviation.

Bala Chidambaram, Chief Engineer & Senior Director, Autonomy, Office of the CTO, Boeing (AI Infra Summit speaker 2025)