Revenue per Megawatt-Hour in Energy: Wholesale Power Market Performance
Direct Answer
Revenue per Megawatt-Hour (MWh) in wholesale power markets is the core unit economics of an energy trader, generator, or utility. Unlike SaaS ARPU, it is highly volatile, geographically fragmented, and subject to regulatory and physical constraints. This guide defines the metric, its drivers (LMP, capacity, ancillary services), and the operational KPIs that govern performance.
A 1% improvement in realized price per MWh can swing a 1 GW portfolio's annual revenue by $5M–$10M+.
Why Energy Measures Differently
Wholesale power markets are not like SaaS or retail. Revenue per MWh is:
- Location-dependent – A MWh delivered at a congested node (e.g., PJM West Hub) can be $40/MWh more than the same MWh at a non-congested node. LMP varies by node, not by customer.
- Time-dependent – The same MWh at 5 PM on a July heat wave can be $500/MWh; at 2 AM in spring, it may be -$20/MWh (negative prices due to oversupply).
- Product-mix dependent – Revenue includes energy, capacity (e.g., PJM Base Residual Auction), ancillary services (regulation, reserves, reactive power), and renewable energy certificates (RECs). A combined-cycle gas plant might earn 60% from energy, 25% from capacity, 15% from ancillaries.
- Regulated vs. Deregulated – In regulated markets (e.g., Southern Company), revenue per MWh is set by cost-of-service tariffs. In deregulated markets (ERCOT, PJM, CAISO), it's pure market-driven.
Key difference from other industries: You cannot "upsell" a MWh. The product is a commodity. The only levers are timing, location, and hedging strategy.
The Most Important KPIs to Track
1. Realized Revenue per MWh (RR/MWh)
Formula: Total Wholesale Revenue ÷ Total MWh Sold Depth: This is the headline KPI. It includes all revenue streams: energy sales (day-ahead, real-time, bilateral), capacity payments, ancillary services, RECs, and any RINs (Renewable Identification Numbers) for biofuels.
Benchmark: For a typical PJM combined-cycle plant in 2023, realized revenue averaged $45–$65/MWh (energy + capacity). A top-quartile operator might achieve $55–$75/MWh via optimized bidding and congestion arbitrage. Vendor example: Clari (used in energy trading desks for revenue forecasting) – pricing starts at $15/user/month, but energy firms use it to track deal-level revenue from PPAs.
However, the core tool is custom-built or via OpenLink (now part of ION Group) for trade capture.
2. Capture Rate (Price Realization)
Formula: (Realized Price / Day-Ahead Index Price) × 100 Depth: Measures how well your portfolio captures the benchmark index. A capture rate of 95% means you're leaving 5% on the table. Best-in-class operators (e.g., NextEra Energy Resources) target 100–105% for renewable portfolios by optimizing curtailment and PPA timing.
Benchmark: Wind farms in ERCOT averaged 85–95% capture in 2023 due to negative pricing during high wind. Solar farms in CAISO averaged 80–90% due to the "duck curve."
3. Capacity Revenue per MWh
Formula: Total Capacity Auction Revenue ÷ Total MWh (past or expected) Depth: Capacity markets (PJM, ISO-NE, NYISO) pay for availability, not energy. A 1 GW gas plant might earn $100–$150/kW-year in PJM Base Residual Auction (2024/2025 clearing price ~$28/kW-month). That adds $3–$5/MWh to realized revenue.
Vendor: Genscape (now part of S&P Global) provides capacity auction data and forecasts – subscription ~$50k/year.
4. Ancillary Services Revenue per MWh
Formula: (Regulation + Reserves + Reactive Power Revenue) ÷ Total MWh Depth: In PJM, a fast-ramping battery can earn $15–$30/MWh from regulation alone. Gas plants with fast start times can earn $5–$10/MWh from synchronized reserves. Benchmark: Top-quartile hydro or storage assets earn 20–30% of total revenue from ancillaries.
5. Congestion Cost per MWh
Formula: (Real-time LMP at Node – Hub LMP) × Volume Depth: Negative congestion (your node is cheap relative to hub) hurts revenue. Positive congestion (node is expensive) helps. Real operators use Financial Transmission Rights (FTRs) to hedge.
Vendor: Lattice (energy trading and risk management) – pricing ~$100k+/year for mid-sized traders.
6. PPA (Power Purchase Agreement) Mark-to-Market
Formula: (Contract Price – Current Forward Price) × Volume Depth: For renewable developers, PPAs lock in revenue. A 10-year PPA at $35/MWh when forward prices are $50/MWh means you're leaving $15/MWh on the table. Mark-to-market is a risk KPI.
Benchmark: In 2023, solar PPAs in ERCOT averaged $25–$35/MWh; gas PPAs in PJM averaged $40–$55/MWh.
Real Operators
NextEra Energy Resources (NYSE: NEE)
- Scale: 30+ GW of renewables + gas.
- Approach: Uses machine-learning price forecasting (from Verdantix and in-house models) to optimize bidding across 100+ nodes. Achieves 102% capture rate on wind portfolio.
- Tool: Gurobi (optimization solver) – ~$50k/year per license.
NRG Energy (NYSE: NRG)
- Scale: 23 GW of fossil + retail.
- Approach: Uses Clari for PPA pipeline tracking and Salesforce for retail customer CRM. Their trading desk uses OpenLink Endur for trade capture and risk.
- Benchmark: Realized revenue per MWh in 2023 was $62/MWh (energy + capacity + retail margin).
Calpine Corporation
- Scale: 26 GW of gas + geothermal.
- Approach: Focuses on ancillary services – their fast-start gas plants earn 15% of revenue from regulation and reserves. Uses ABB's Ventyx (now part of ABB) for asset optimization.
- Benchmark: $7.50/MWh from ancillary services alone in 2023.
EDF Renewables
- Scale: 10 GW+ of wind/solar.
- Approach: Uses PPA hedging with LevelTen Energy (PPA marketplace). Achieves 95–98% capture rate via curtailment optimization.
- Tool: LevelTen – transaction fee ~2% of PPA notional.
Failure Modes
- Over-reliance on Day-Ahead Markets – Day-ahead prices often diverge from real-time. A plant that only bids day-ahead misses congestion arbitrage. Case study: In CAISO summer 2022, real-time prices were $150/MWh higher than day-ahead for 12 hours – plants that didn't rebid lost $20M+.
- Ignoring Transmission Constraints – A generator at a congested node may face negative LMP even when hub prices are high. Failure: A 500 MW wind farm in ERCOT's West Zone lost $15M in 2023 due to curtailment and negative pricing.
- Poor Hedging of Capacity Revenue – Capacity auctions clear annually; if you don't hedge with FTRs or swaps, a drop in capacity prices (e.g., PJM 2025/2026 clearing price fell 40% from 2024) can wipe out $3–$5/MWh of revenue.
- Underestimating Ancillary Services Revenue – Many operators treat ancillaries as "bonus" instead of optimizing for them. Failure: A 1 GW gas plant in PJM that didn't bid into regulation lost $8M/year in potential revenue.
- Over-hedging PPAs – Locking in a 10-year PPA at $30/MWh when forward prices are $60/MWh creates opportunity loss. Case study: A solar developer in CAISO signed a $25/MWh PPA in 2020; by 2023, spot prices were $55/MWh – a $30/MWh loss per MWh.
Reporting Cadence
| KPI | Cadence | Owner | Tool |
|---|---|---|---|
| Realized Revenue per MWh | Daily | Trading Desk | OpenLink Endur |
| Capture Rate | Weekly | Portfolio Manager | Custom SQL/BI |
| Capacity Revenue per MWh | Monthly | Risk Manager | Genscape + Excel |
| Ancillary Services Revenue | Daily | Dispatch | ABB Ventyx |
| Congestion Cost per MWh | Real-time | Trading Desk | Lattice |
| PPA Mark-to-Market | Weekly | Finance | Salesforce + Clari |
Best practice: Daily morning meeting to review prior day's RR/MWh vs. Index. Weekly deep dive on capture rate by asset. Monthly board review of capacity and ancillary revenue trends.
30-60-90
First 30 Days: Audit & Baseline
- Task: Pull 12 months of historical data for each asset: realized price, day-ahead index, capacity revenue, ancillary revenue.
- KPI to establish: Baseline RR/MWh – e.g., $48/MWh for a PJM gas plant.
- Tool: Excel or Tableau (pricing ~$70/user/month for Tableau Creator).
- Vendor: Clari can be used to map PPA pipeline to revenue.
Days 31–60: Optimization & Gap Analysis
- Task: Identify top 3 underperforming assets. Run capture rate analysis. If capture rate < 90%, investigate bid strategy.
- Action: Implement real-time bid optimization using Gurobi or Lattice.
- Benchmark: Target 95% capture rate for all thermal assets, 90% for renewables.
Days 61–90: Hedging & Reporting Cadence
- Task: Set up daily RR/MWh dashboard in Tableau or Power BI (pricing ~$10/user/month). Implement weekly capture rate review.
- Action: Hedge capacity revenue using FTRs or swaps. Vendor: ICE (Intercontinental Exchange) for futures – transaction fees ~$0.50/MWh.
- KPI target: Achieve $2–$5/MWh improvement in realized revenue vs. Baseline.
FAQ
Q: What is a good Revenue per MWh for a gas plant in PJM? A: For 2023, $55–$65/MWh is typical (energy + capacity). Top quartile hits $70+/MWh via ancillary services and congestion arbitrage.
Q: How do renewables compare? A: Solar in CAISO averaged $35–$45/MWh (energy + RECs) in 2023. Wind in ERCOT averaged $25–$35/MWh due to negative pricing. Capture rate is the key differentiator.
Q: What tools do energy traders use for RR/MWh tracking? A: OpenLink Endur (trade capture), Lattice (risk), Genscape (market data), and Clari (PPA pipeline). Pricing ranges from $50k/year (Genscape) to $500k+/year (OpenLink).
Q: How does congestion affect RR/MWh? A: A plant at a congested node can see $5–$20/MWh difference vs. Hub. Positive congestion (node is expensive) boosts revenue; negative congestion hurts it. FTRs hedge this.
Q: What is the biggest mistake operators make? A: Ignoring ancillary services. A 1 GW gas plant can earn $5–$10/MWh from regulation and reserves. Most operators leave 50%+ of that on the table.
Q: How often should I report RR/MWh? A: Daily – it's the most volatile KPI. Weekly for capture rate, monthly for capacity/ancillary breakdown.
Sources
- PJM Base Residual Auction Results 2024/2025
- ERCOT Real-Time LMP Data
- NextEra Energy Resources Investor Presentation 2023
- Genscape Power Market Analytics
- OpenLink Endur – Energy Trading and Risk Management
- LevelTen Energy – PPA Marketplace Trends
- Clari – Revenue Operations for Energy
- Lattice – Energy Trading Software
