Building the Renewable, Self-Driving Grid
summary
Company builds grid-scale management software that uses machine learning to create pseudo real-time monitoring from slow smart‑meter feeds (often 2–24 hour delays), models capacity at every node, and orchestrates storage and flexible loads to avoid costly upgrades. Claims existing networks can do ~3–4x more work if software enforces hierarchies of autonomy and local caching (batteries as buffers) and if operators adopt canaryed automation. Demonstrations so far are pilot deployments and simulated large‑scale tests; full rollouts need trust, ops integration and about a decade of work.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 3:39Climate urgency and decarbonization
- 10:04Modern versus traditional grids
- 17:51Google’s cloud lessons
- 23:33Reliable systems from unreliable parts
- 30:13Software for grid efficiency
- 37:56Autonomy hierarchies in energy systems
- 46:01Future of self-driving grids
- 55:54Closing thoughts
related talks
16mBuilding Resilient Smart City Platforms
14mFrom Dashboard to Defense: Automating Resilience at
05mBuilding Self-Adaptive Autonomous Infrastructure
31mMonitoring Communication Networks of Energy Grids With the Open-source Framework...
33mWhen Observability Meets Sustainability: A Real World Experience
49mPanel Discussion: Moving Towards...
21mAdaptive Power Optimization: Battery-Operated Chip Design
12mSRE Principles for High-Speed Infrastructure Monitoring