Energy AI
Know demand before the peak.
Site-level demand forecasting from historical load, tariffs and weather, feeding dispatch and pricing decisions.
- Intelligent dispatch
- Demand forecasting
- Dynamic pricing
HQZN brings demand forecasting, dynamic pricing, edge recognition and digital twins into one intelligence layer behind every charging decision.
Built on the industry's leading stack
Network status
All systems operational
01 / Command Center
From network status to AI recommendations, operations teams monitor, analyse and act inside one light, legible command surface.
Open the command centerAI Command Center
Automations running
08
2 activated today
Connected chargers
1,286
98.7% availability
AI decisions today
1,842
+18.6% vs. baseline
Optimisation value
¥428K
+¥64.2K projected
Agent activity
Autonomous decisions
Live activity
Load rebalanced
14 chargers · 2m ago
Failure predicted
Unit 08 · 8m ago
Price optimised
Zone C · 14m ago
40%+
lower O&M cost (customer-reported)
6
core AI capabilities
5
solutions
4
end-to-end stack layers
02 / Core capabilities
Not a general chatbot, but a set of specialised models trained, deployed and continuously tuned for charging operations.
Explore the platformEnergy AI
Site-level demand forecasting from historical load, tariffs and weather, feeding dispatch and pricing decisions.
Edge AI
Jetson Thor and RK3568 run vision inference locally for millisecond recognition, safety detection and anomaly alerts.
Digital twin platform
NVIDIA Omniverse-based station twins with 3D visualisation, scenario simulation and remote operations.
Edge interaction
RK3568-powered on-site terminals with voice interaction, gesture recognition and guided flows that lower the barrier for drivers.
Charging-scene AI
Plate recognition, charging-behaviour analysis and battery health diagnosis make every session measurable and optimisable.
Intelligent operations
Multi-source time-series anomaly detection and predictive maintenance that open work orders early and cut O&M cost by 40% or more.
03 / Full-stack AI capability matrix
DGX Spark handles training and simulation, Jetson Thor runs real-time inference on site, and Omniverse and OpenUSD keep the physical station and its twin in sync.
Desktop AI supercomputer
Train and fine-tune charging models on site without sending operating data off campus.
Real-time edge inference
Run vision and time-series models at the station with millisecond response to safety events.
Physically based simulation
Build interactive station twins and rehearse dispatch strategies virtually.
Open scene description
Let equipment, BIM and simulation data flow in one format.
04 / Solutions
Five solutions for different stages of charging operations. Deploy them alone or combine them into a complete AI operating system.
View all solutionsOne AI hub for the whole station.
Demand forecasting, dynamic pricing, intelligent dispatch and unattended operation in one platform.
Make charging a flexible grid resource.
V2G, demand response and power-trading strategies that turn charging load into a dispatchable grid asset.
Rehearse first, then execute.
Omniverse-based 3D station twins for simulation and remote operations.
See the health of every pack.
SOH prediction, RUL estimation, safety alerts and lifetime optimisation across the whole battery lifecycle.
A station that speaks for itself.
RK3568-based on-site terminals with voice interaction, gesture recognition and smart guidance.
05 / Digital twin
Map real equipment, live telemetry and AI predictions onto an interactive digital twin of the charging station.
Enter live monitoringCharging station digital twin · live
Live sync · Demo data
HQZN / AI charging operating system
Bring your own station data and see forecasting, dynamic pricing and the digital twin at work.