AI-native is to network planning what cloud was to compute.
A category shift, not a feature. AgenticGrid was built with AI from first principles and is operated by governed AI agents — every step inspectable, every number citable. The competitive moat is the architecture, not the model.
Why now
The planning bottleneck is real and structural.
Distribution networks built for one-way flow now manage two-way flow, solar export, EV import and clustered controllable load. Top-down models are slow, expensive and opaque — networks can rigorously model only a handful of the proposals they receive and triage the rest by judgement. The bottleneck the industry named to us is scenario-throughput, and that is the bottleneck AgenticGrid removes.
Traction
Four networks, two regulators, one academic institute.
Within the RACE for 2030 GridGuru Phase 2 program (Project 24.NT6.R.0928, ten months, July 2025 to May 2026), the platform has reached:
- Transgrid (TNSP) — industry funder; official program demonstration January 2026.
- Essential Energy (DNSP) — demonstration → scoped proof-of-concept → live microgrid-sizing pilots for two remote NSW communities.
- Ausgrid (DNSP) — deep engagement on per-substation load allocation and DSO strategy.
- Energy Queensland (DNSP) — demonstrated to the planning team; live interest in a regional pilot.
- AER + AEMC — engaged via an independent AER-style RIT-D re-run and the AEMC ISP-framework synthesis demo.
- Monash University (Monash AI Institute + Monash Energy Institute) — academic backing for the community-battery optimisation method, customer-archetype synthetic load generation, and EV charging-planning.
Unit economics
Per-tenant SaaS, deployed in minutes.
One Docker image. One isolated Azure Files share per tenant. One audience-specific system prompt. The marginal cost of a new tenant is the Azure compute + storage — the platform itself ships once.
The next tenant takes an afternoon. The work on Broken Hill makes the next town faster, which makes the next network faster again. Ruflo's structural learning loop compounds what the platform knows across customers without leaking data between them.
Roadmap
Pilots now, AEA next.
Live Essential Energy pilots are running. Active interest from Energy Queensland and Ausgrid is in the next-pilot conversation. Follow-on funding is being pursued through the Australian Economic Accelerator in partnership with Monash to carry the research through to commercialisation. The architecture points at a grid in which every substation has its own agent — the same governed, transparent, defensible architecture proven on Broken Hill, applied at national scale.