Grid data the way operators see it.
Open-source infrastructure for real-time grid demand, interconnection queues, wholesale electricity prices, and renewable curtailment. Built for the people working the energy transition.
Open tools for open problems.
Real-time CO₂ intensity across every US grid region.
Weighted average lbs CO₂/MWh from live fuel mix data. Covers all 7 primary ISOs (CAISO, ERCOT, MISO, PJM, NYISO, ISONE, SPP) plus 20+ smaller balancing authorities via EIA-930. EPA eGRID 2023 emission factors. 60-second polling.
Every US power project waiting to connect to the grid.
Unified search across all 7 major ISO/RTO interconnection queues: ERCOT, MISO, PJM, CAISO, SPP, NYISO, ISO-NE. Daily refresh via GitHub Actions. One interface for the data that used to require 7 browser tabs.
Real-time US electricity demand across 15 balancing authorities.
Microservices pipeline: Redis Streams, Postgres, Kubernetes with 17 HPAs. 95% CONUS coverage. 23ms chaos recovery, 215 req/s under load.
How much clean energy gets thrown away. Every day, by ISO.
Daily solar and wind curtailment across CAISO (California), SPP (Great Plains), and ERCOT (Texas). 90-day rolling history. Reveals where the grid is congested and where storage is needed most. Refreshed each morning.
Live electricity prices on a map. Every pricing node, every ISO.
Full-screen nodal price map across NYISO, ERCOT, MISO, SPP, and CAISO. Every node color-coded from green to red by real-time LMP. Click any node for 14-day history: energy, congestion, and loss breakdown. RT and DA markets. Auto-refreshes every 60 seconds.
Electricity spot prices across the US grid. Every 5 minutes.
Real-time and day-ahead locational marginal prices across NYISO, PJM, CAISO, and SPP. Energy, congestion, and loss components per hub node. Fuel mix by source, Henry Hub gas prices, grid temperatures, renewable curtailment, and battery storage, all in one terminal.
Our model calls the price spread. See if it was right.
A temporal fusion transformer publishes day-ahead RT-DA spread forecasts for 15 ERCOT hubs and load zones every day, before delivery. Forecasts are written once and never revised, then scored in public against realized prices, including a paper-traded trading signal with published, fee-adjusted P&L.
Open source on GitHub
Source code, data pipelines, and infrastructure, all public under kardashev-lab.
Energy and compute, used with unprecedented wisdom.
Evidence replaces guesswork. Efficiency becomes the default interface to power. We build the tools that help teams make data-driven decisions that scale, toward a civilization that uses every joule and every FLOP with intention.
“If we can't explain it, we don't deploy it.”
Three pillars.
Plan
Concrete schedules and policies. Evidence first, assumptions labeled, advisory before automation.
Prove
Baselines, assumptions, and measured deltas. If we can't explain it, we don't deploy it.
Scale
Automation only where data supports it. Systems that compound without compounding risk.
Stay close to the work.
Occasional notes on what we're building, what the data shows, and where we're going. No spam. Unsubscribe anytime.