You've hit on a really important point! When people think about lights-out factories, they usually focus on the flashy robotics and AI systems, but the infrastructure challenges are what truly determine success or failure. Based on what I've found, here are the most overlooked technical hurdles:

1. **Power Infrastructure Resilience**: This goes way beyond basic backup generators. You need multi-layered power redundancy with seamless switching between grid power, UPS systems, and backup generators. Even micro-second power fluctuations can crash sensitive automation systems. The real challenge is maintaining power quality - clean, stable electricity without spikes or sags that can damage equipment.

2. **Environmental Control Precision**: Temperature and humidity control isn't just about comfort - it's about equipment performance. Many automated systems have tight environmental tolerances. HVAC failures can cause thermal expansion issues, condensation on electronics, or lubrication breakdowns. You need redundant climate control systems with real-time monitoring and automated failover.

3. **Remote Diagnostics Depth**: Basic monitoring isn't enough. You need predictive analytics that can detect subtle changes in vibration patterns, thermal signatures, or power consumption that signal impending failures. The challenge is creating diagnostic systems that can differentiate between normal operational variations and actual problems.

4. **Network Security & Reliability**: In a lights-out factory, the network IS the factory. Any disruption means complete shutdown. You need redundant, physically separate network paths with automatic failover. Security becomes critical - a cyber attack could literally stop production with no humans on-site to intervene.

5. **Automated Material Handling**: How do raw materials get in and finished products get out without human intervention? This requires sophisticated automated guided vehicles, robotic loading systems, and smart inventory management that can handle unexpected variations or jams.

6. **Predictive Maintenance Integration**: This is where many implementations fail. You need maintenance systems that can not only predict failures but also schedule and execute repairs autonomously - ordering parts, dispatching service robots, and validating repairs before restarting production.

The companies that succeed with lights-out manufacturing are the ones that treat infrastructure as mission-critical, not just supporting systems. They invest in redundancy at every level and build systems that can self-diagnose and self-recover from failures. It's less about the robots and more about creating an ecosystem that can survive and thrive without human intervention!