KL.Grid is live — one market down, a region to go!

After months of calibration work, KL.Grid — K Labs’ electricity spot price simulation platform — has completed a full 51-month backtest across Panama’s power market (Jan 2022 – Mar 2026), spanning three distinct ENSO regimes.
📊 The results are solid:
· Mean Absolute Error: 4.7% across 36 hard-confirmed months
· 49 of 51 months within ±15% of actual CND prices
· 51 of 51 within ±20% — across La Niña, El Niño, and transition years
· Validated through the structural break of Gatun’s commissioning, which dropped average market prices by ~37% year-on-year
⚡ What KL.Grid does:
Hourly resolution dispatch (8760 hrs/year) across the full generation fleet· Stochastic price tubes with ENSO regime uncertainty· Hydro ICF calibration with SCADA data· Endogenous computation of the value of water — the opportunity cost of stored reservoir energy, derived from market conditions and ENSO regime· Zonal transmission network module with line flow limits and nodal pricing potential· Gas pipeline network module with capacity constraints and congestion premiums· AI-generated market commentary· Full audit trail.
💧 Value of water — a first-class output:
In hydro-dominated markets, the single most consequential modelling decision is how much a MWh of stored water is worth. KL.Grid computes this endogenously by month and ENSO regime. In La Niña conditions with full reservoirs, water values fall toward Gatun’s variable cost. In an El Niño drought, they spike toward higher fuel costs — the true scarcity value of every MWh in storage. Directly relevant for reservoir operators, PPA negotiations, and DFI revenue assessments.
🌐 Network-aware modelling:
The transmission module enforces line flow limits and computes zonal nodal prices — directly relevant for Chile, where north-to-south grid constraints regularly create material price differentials between zones. The gas pipeline module enforces capacity limits and congestion premiums — directly applicable to Colombia, where pipeline constraints between producing regions and thermal hubs affect dispatch and spot prices. Both activate via Excel configuration — no code changes required.
🌎 What it’s ready for:
The platform is market-agnostic. Panama is the first fully calibrated market, but the engine is configured via standard Excel input files — no code changes needed to model Colombia, Costa Rica, the Dominican Republic, Chile, or any other LAC power system.
The charts below show the 2026 forward price tube, stochastic generation distributions for Penonomé wind and Solar Chiriquí, the full 2022–2026 backtest validation panel and others.

💬 Which market should we model next?
Drop your market suggestions in the comments 👇