Wakeline
Wind turbines turning on forested hills under a blue sky

ENERGY & UTILITIES

Our first deployment, in production since April 2026.

We deployed our AI in the European power market first. It is our largest deployment to date. Market Edge forecasts day-ahead electricity prices in production. It learns from 83 live streams as the market changes, with no retraining.

What we are doing in energy today

  1. Production

    Market Edge, day-ahead price forecasts

    24-hour electricity price forecasts for European day-ahead markets, in production since April 2026, currently from 83 parallel input streams. Battery storage operators and energy traders use them to plan their operations.

    Market Edge
  2. Empirical

    The CLS Forecasting Benchmark

    In a historical backtest covering 2,097 days in the German-Luxembourg power market, Market Edge achieved 14 percent higher cumulative trading P&L than the TabPFN baseline. Most of the advantage comes from regime shifts. This is a backtest result, not a measure of live trading returns.

    Read the benchmark report
  3. Production

    No retraining

    The instances behind Market Edge keep learning as the market changes. Nobody schedules a retraining run, and each instance runs on an ordinary processor.

    How it learns

RESEARCH

What we have run so far

Deployed

wholesale power market

Market Edge has forecast day-ahead electricity prices continuously in production since April 2026. It learns while operating, currently ingesting 83 parallel input streams. Battery storage operators and energy traders use its forecasts to plan their operations.

Backtested

wholesale power market

Historical backtest, Market Edge: 14 percent higher cumulative trading P&L than the TabPFN baseline across 2,097 days in the German-Luxembourg power market. The advantage concentrates in regime shifts. A backtest result, not live trading returns.

Demonstrated

simulated stock market

In a simulated stock market, an instance acted and learned what its actions cost: over 2,177 trades it recovered a hidden 25 percent tax schedule from the consequences of its own trades, with no reward function, and recovered it again each time the rule changed.

Demonstrated

simulated road network

In a simulated road network, an instance learned to predict delays and to choose a route before the delay happened, and kept improving how it acted on its own forecasts as conditions changed. We are looking for design partners to take this into operational use.

Aerial view of a wind turbine beside an access road

We work with a select few companies per industry.
If yours should be one,
talk to us.

Talk to us