Trade the move.
Before the market sees it.

Our agent Athena watches your positions, queries every model, and surfaces what's about to shift on the assets you own and the spreads you trade. Powered by EPT-2, the most accurate forecast on the planet.

In production across utilities and trading desks worldwide
Axpo
Switzerland
TotalEnergies
France
Shell
United Kingdom
Enel
Italy
Statkraft
Norway
EnBW
Germany
EDF
France
Hydro-Québec
Canada
Adani Energy
India
Vitol
Netherlands
Origin Energy
Australia
ESB
Ireland
What Athena does on the desk

Pick a role, pick a question. Watch Athena reason through it.

These are the same simulations Athena runs in production. Reasoning, data acquisition, code, analysis, conclusion. Click a question to start it again.

Example queries
Example queries
Example queries
Athena
10:28:57
Scroll into view to start.

Working10:29:53
Creating 2 data sources10:29:53
Created ENTSO-E market data: DE Today – Prices & Generation
Created ENTSO-E market data: FR Today – Prices & Generation
✓ Completed: 2 created, 0 failed
Creating 3 widgets10:30:25
✓ Created chart: DE vs FR Day-Ahead Prices – Today
✓ Created chart: DE Generation Mix – Today
✓ Created chart: FR Generation Mix – Today
✓ Completed: 3 created, 0 failed
Running 1 Python script10:32:01

Running 1 Python script to analyze your data.

DE vs FR Spread Analysis

Running analysis...

DE/FR Spread — April 9

09:00–14:00 UTC: Both zones briefly converged at near-zero / negative prices (solar peak drove both to -13 EUR/MWh). Spread collapsed to zero.

After 14:00 UTC: As solar fades in Germany, DE prices spiked hard (gas/lignite back on the margin), while FR stayed suppressed by nuclear. Spread exploded to 85–122 EUR/MWh.

Structural driver: Nuclear

France's ~42.5 GW nuclear fleet acts as cheap, always-on supply, putting a floor on FR prices. Germany has no nuclear, its price tracks marginal gas/coal much more directly.

Generation comparison

DE Solar: 22,800 MW · DE Wind: 16,300 MW · DE Nuclear: 0 MW | FR Solar: 7,825 MW · FR Wind: 1,390 MW · FR Nuclear: 42,500 MW

Bottom line

What Athena reads from

Your edge model. The market's models. The market itself.

A trading desk needs more than the best forecast. It needs the forecasts every other desk is looking at, and the markets they all react to. Athena puts all three in the same context window, plus your own data when you connect it.

01Your edge model
EPT-2
Flagship global AI forecast.
EPT-2 Early
Runs sooner after observations land.
EPT-2 Ensemble
Probabilistic forecasts and uncertainty.
EPT-2 Rapid Refresh
Frequent updates through the day.
EPT-2 High Resolution
High-resolution forecasts for European assets.
EPT-1.5
Previous generation, still available.
02The market's models
ECMWF
IFS, ENS, EC46, AIFS.
ICON
Global, EU, D2.
GFS
Deterministic, ensemble, GraphCast.
Microsoft Aurora
AI foundation model.
Météo-France
AROME.
KNMI
HARMONIE.
UK Met Office
UKMO.
03The market itself
ENTSO-E
European TSO transparency data.
EEX
Power and gas exchanges.
EPEX
European spot power markets.
BMRS / ELEXON
UK power data.
Netztransparenz
German grid transparency.
Prediction markets
Polymarket and others.
What you do with it

Forecasts. Positions. Workflows. One place.

Pick a capability to see what it covers.

Power forecasts

Power output as a separate model class on top of every weather model in the catalogue, plus Jua's power actuals nowcast.

01
MW, not m/s

Every weather model is also available as a first-class power output forecast. Deterministic and ensemble.

02
Power actuals

Jua's nowcast of current power output. The real-time MW state of the assets you care about, refreshed continuously.

03
Asset-aware

Per turbine, per bidding zone, or per market summary. Whatever shape your book is.

Forecast skill · 2024 held-out evaluation

Trading edge starts with forecast skill. These are the numbers it runs on.

Wind 10m · RMSE (m/s) · lower is better
Held-out 2024 evaluation
0h1d2d3d4d5d6d7dLead time1.82.252.73.153.6
Jua
ECMWF IFS
ECMWF AIFS
Microsoft Aurora
NOAA GFS
16%
Lower RMSE at 7 days
vs ECMWF IFS
24%
Lower RMSE at 7 days
vs Microsoft Aurora
21%
Lower RMSE at 7 days
vs NOAA GFS
Model
1d
3d
5d
7d
Jua
1.913
2.134
2.354
2.539
ECMWF IFS
1.935
2.212
2.552
3.023
ECMWF AIFS
2.084
2.294
2.573
2.889
Microsoft Aurora
1.991
2.235
2.650
3.319
NOAA GFS
2.115
2.412
2.830
3.196

RMSE of 10m wind speed above ground (m/s) on a held-out 2024 evaluation. Lower is better. EPT-2 Reasoning at 7 days is 2.539 m/s, ahead of every named alternative. Methodology and full tables are in the EPT-2 technical report.

Athena is in production. Bring it onto your book.

Open a trial, point it at your assets, ask it the question you ask your team every morning.

AccessWeb dashboard·REST API·Python SDK·CLI