Mining energy economics
After the halving, every surviving miner is an energy trader first
Bitcoin mining margins collapsed to 20-30% post-halving, making energy cost the dominant variable in profitability. At current difficulty, a 2 cent/kWh difference in effective power cost separates profitable operations from shutdown candidates. The miners who survive are not the ones with the most hash rate. They are the ones who buy power like a trading desk.
Hash rate is a commodity. Energy position is the alpha.
Power is the only variable miners control
Bitcoin miners operate in a market where difficulty adjusts algorithmically and block rewards halve every four years. The only cost lever remaining is electricity — 60-70% of operating expense. A 10 MW facility at $0.04/kWh spends $3.5M annually on power alone. The difference between profitable and unprofitable mining is measured in single-digit mills per kWh, captured or lost through procurement timing and curtailment discipline.
After the halving, every miner that survives is an energy trader first.
How AI optimizes mining energy economics
Forecast power price windows
Predict 15-minute interval prices across wholesale markets. Mining profitability flips between positive and negative multiple times per day. Knowing which intervals to run determines survival.
Optimize hash rate against price
Dynamically scale mining operations up and down based on real-time and forecasted power prices. Not every hour is worth mining. AI identifies which hours generate positive margin.
Monetize curtailment capability
Register mining load as demand response capacity. When grid stress drives prices above mining economics, curtail and earn the spread. Mining as a flexible load earns twice.
Manage power procurement strategy
Blend PPAs, spot market, behind-the-meter generation, and curtailment credits into an optimal power portfolio. Static contracts leave money on the table when prices move.
Manual curtailment vs AI-automated load response
| Metric | Manual Process | AI-Optimized |
|---|---|---|
| Forecasting accuracy (MAPE) | 8-10% | 3.21% |
| Decision cycle time | 4-8 hours | 15 minutes |
| Billing query resolution | 2-3 days | < 5 minutes |
| Residual value model refresh | Quarterly | Daily |
| Operational data utilization | < 30% | 98%+ |
| Margin capture potential | Baseline | 5-12% uplift |
Survival of the most efficient
Post-halving economics create a binary outcome: miners operating below $0.04/kWh survive; those above exit. AI energy optimization — curtailment timing, 5-minute reforecasting, procurement intelligence — compresses operating cost by 15-25%. The miners investing in energy intelligence today acquire the capacity of those who did not when the market turns.
Every halving is a stress test. Energy intelligence is the thing being tested.
Key players
Riot Platforms
Largest US Bitcoin miner; 1 GW capacity in ERCOT, grid-responsive curtailment.
Marathon Digital
800 MW mining capacity; expanding into immersion cooling and curtailment revenue.
CleanSpark
600 MW hashrate; focused on low-cost power procurement and grid services.
Iris Energy
Renewable-focused miner; 510 MW capacity with AI/HPC diversification.
What we have shipped in this space
Attribution — TS2Vec-Similar Day forecasting
Production system forecasting ERCOT day-ahead prices every 5 minutes. Trained on 2 years of SCED interval data, weather, and transmission constraints.
Residuals — operational telemetry to financial instruments
Battery degradation curves, solar performance decay, and generation asset condition converted from operational telemetry into residual instruments that reflect actual state.
Our price forecasting system provides the signal that mining operations use for run/stop decisions. At post-halving margins, the accuracy difference between 8% MAPE and 3% MAPE determines which operations survive.
Mining survival is a forecasting problem. We built the forecasting system.
Ready to instrument your operations?
Get a specific power cost audit for your mining operation. We'll identify the exact hours where price spikes cost you the most and quantify the demand response revenue you could capture.
Schedule an auditExplore more
Related activities
Mining curtailment programs→
Bitcoin mining operations in ERCOT represent 4.2 GW of interruptible load that can shed within minut...
Mining power procurement→
Post-halving mining economics require all-in power costs below $0.04/kWh to maintain positive margin...
Grid-scale battery dispatch→
Grid-scale batteries co-located on the same node, with identical chemistry and capacity, show 30-40%...
Common questions about AI in energy optimization crypto mining
What kilowatt-hour efficiency requirement is needed for mining to remain profitable at $20,000 bitcoin?
At $20,000 bitcoin, operations consuming more than 0.25 kWh per dollar of expected revenue face negative unit economics at $0.06–$0.08/kWh power costs. Efficient operations achieving 0.18–0.22 kWh per dollar sustain 15–25% gross margins; less efficient rigs require sub-$0.04/kWh power to remain viable.
How much excess renewable generation (stranded curtailment) is technically harvestable by mining operations?
Rural areas with 200–400 MW of stranded wind/solar capacity can technically support 40–80 MW of mining infrastructure. Most regions restrict permanent mining connectivity to avoid grid reliability issues; co-siting with offtake agreements captures 60–75% of stranded generation potential.
What is the optimal mining operation size to absorb behind-the-meter generation from a 10-megawatt solar farm?
A 10-megawatt solar farm can reliably support 6–8 megawatts of mining load (given solar intermittency), with oversizing to 10–12 megawatts creating uneconomical curtailment (8–15% annual). Optimal sizing matches 70–80% of average solar generation to grid-stable loads.
Can mining flexibility reduce renewable curtailment by more than 15% in isolated grids?
Isolated grids with 50%+ renewable penetration can reduce curtailment by 15–25% with flexible mining loads, provided mining infrastructure represents 10–15% of total peak demand. Grids with less than 5% potential mining load see minimal curtailment reduction (3–5%).