AI-enhanced digital twins

Full grid visibility from minimal sensors.

A substation-level sensor plus a learned digital-twin AI gives DSOs and data centers real-time observability of every node — voltage, phase, neutral — without rewiring the grid.

Book a pilot → How it works
Sensors / networkSubstation only
Deploy timeDays, not months
Accuracy<0.3 V avg. MAE
GridFactor · AI Digital Twin | Network · 142 nodes live 98.6% conf

Network · live state

Substation 04 — feeder F-04
3-phase voltage · selected bus · last 24h
A
B
C
EN 50160
+10% 253V −10% 207V 230V
The problem

Low-voltage grids are blind.

95%+ of LV distribution nodes are unobserved — difficult and expensive to monitor with the technology available today. Meanwhile EVs, solar, and home batteries are decentralizing every street.

01
Smart metering
Slow rollout. Unclear if smart meters can be used as active grid-management tools.
SCADA-blind
02
Comms bottleneck
Metering data lags. Field-bus and PLC links don't deliver real-time state at scale.
Latency > minutes
03
Power supply quality
Uncontrolled voltage volatility in LV feeders. Under/overvoltage events are rising.
EN 50160 risk
04
LV digital twins
No tools available today to run power flows or simulations on real LV networks.
Tooling gap
05
EV load
EV chargers turning quiet streets into peak-load corridors.
+30% peak demand
06
Solar injection
Rooftop solar pushing voltage above legal limits at midday.
EN 50160 violations
07
Home storage
Home batteries cycling phases unevenly across the neutral conductor.
Phase imbalance ↑
08
EU regulation
Regulators tightening voltage-quality enforcement and digitalization mandates.
EN 50160 audits
Solution

One sensor at the transformer.
Visibility on every node downstream.

Three steps. No customer-side metering. No rewiring. Software does the rest.

Step 01

Install a compact sensor unit at the transformer.

One device per substation. Hours, not weeks. No customer-side metering required.

SENSOR TRANSFORMER
Step 02

Our AI trains on a digital twin of your network.

ANNSE — a learned state estimator — infers what every node sees, from substation data alone.

Step 03

See every node — voltage, phase, neutral — in real time.

Sub-second inference. Anomaly trending and predictive alerts pushed to your DMS.

Quantifiable impact

Validated in four years of PhD research,
with real DSO pilot data.

<0.3V
Average MAE vs. ground-truth meters.
PhD-validated · real DSO data
85–95%
Fewer measurement points required for full LV observability.
vs. traditional state-estimation std.
<1s
Inference latency from substation reading to per-node estimate.
Edge or cloud · production-tested
Comparison

Same observability — at a fraction of the cost,
in days instead of months.

Traditional EPMS / SCADA
GridFactor
Sensors required
at least one per node
substation measurements only
Deployment time
weeks to months
days
CAPEX per substation
€40k – €80k
€5k – €10k
Predictive capability
none
AI-powered anomaly trending
Full observability
impossible
AI-native
Value proposition

Three layers of return,
backed by measurable economics.

01 — Quality of Supply

Avoid penalty costs & capture QoS bonuses.

  • Identify in real-time the duration, customers affected, and severity of under/over-voltages.
  • Steer future QoS interventions to cut penalties (0.5% of total regulated income).
  • Capture continuity and voltage-quality incentive bonuses.
Impact
€0.5–1M / year per €100M of grid assets
02 — OPEX & Maintenance

Cut technical losses and reactive repairs.

  • Poor monitoring adds ~2% in technical losses — €100–200k/year of wasted energy.
  • Reactive repairs run €5–10k each — €100–300k/year in mid-size regions.
  • Predictive alerts replace unplanned interventions.
Impact
€150–400k / year per substation
03 — CAPEX & Future Investments

Software-based retrofit instead of new hardware.

  • Comparable observability at €5–10k per secondary substation.
  • vs. €40–80k for a fully instrumented LV feeder.
  • No smart-metering 3.0 or advanced-sensor rollout required.
Impact
Potential +M€ CAPEX savings per DSO
Two markets · same engine

Sold to whoever owns the low-voltage layer.

$23B → $45B
Smart-grid tech · 2023 → 2031
8.7%
CAGR · smart-grid market
40–60%
addressable at LV layer
$1.5B
EU VC into grid tech · 2023
Vertical 01 — utilities

DSO & utility networks.

Real-time monitoring of LV distribution under tightening EN 50160 voltage-quality regulation. Software retrofit on existing infrastructure — no transformer swaps, no field rewiring.

Buyer
Distribution System Operators
Trigger
EN 50160 audit risk
ACV
€60–250k first-year
Vertical 02 — data centers

Modular data centers.

Reliability-as-a-Service: predictive electrical anomaly detection on rack-row LV distribution — before failures cost uptime.

Buyer
DC operators & modular vendors
Trigger
Uptime SLA · AI-load growth
ACV
€50–150k per site
Team

Two cofounders who built
power systems for a living.

AB
CTO · Founder
Andrea Bragantini, PhD
PhD Electrical Engineering, UPC BarcelonaTECH (2026). Developed the ANNSE methodology end-to-end. Owns all underlying IP.
UPC PhD '26 ANNSE inventor CITCEA
LC
CEO · Cofounder
Lucas Chidiac
MSc Electric Power Systems (UPC), BSc Electronic Engineering. MV/LV electrical design for hyperscale and modular data centers.
GE ABB B-Global Tech
Open seat
Software / product cofounder
Looking for a senior engineer to own the product surface — cloud platform, dashboard, integrations with DMS / EPMS.
Equity Founding
Get in touch

The future LV grid needs real-time intelligence.
We're building the AI engine that makes it possible.

Let's build it together. Looking for pre-seed partners, DSO design partners, and a modular data-center pilot site. Drop a note — we'll reply within 48h.

CEO · Cofounder
Lucas ChidiacUse the contact form
CTO · Founder
Andrea Bragantini, PhDandrea.bragantini@upc.edu · +39 347 9862443
Round
Pre-seed · €50–80k
Based
Barcelona, EU