The Houston paradox
To understand where battery developers are placing their bets in ERCOT, you first have to understand how a battery makes money. The revenue stack has shifted sharply toward energy arbitrage: charging when wholesale prices are low, typically midday when solar is abundant, and discharging when prices are high, typically during the evening peak when demand rises and the sun has set. The difference between those two prices is the spread, and the spread is the starting point for arbitrage revenue. Higher spreads generally mean higher potential returns.
Energy arbitrage was not always the main revenue stream. Until recently, ERCOT batteries earned the bulk of their revenues from ancillary services, short-duration grid-stability products such as frequency regulation and contingency reserves that paid well when relatively few batteries competed for them. As more batteries came online, those markets saturated and prices fell, pushing operators toward arbitrage. But arbitrage can saturate too: once enough batteries chase the same midday-to-evening spread, they begin to compress the very price difference they are trying to capture.
The shift has been sudden. Energy arbitrage rose from 25% of ERCOT BESS revenues in June 2024 to 76% in June 2025, a complete inversion of the revenue stack in twelve months. ERCOT is now converging toward the CAISO benchmark, where mature solar-driven arbitrage delivers roughly 90% of battery revenues. If arbitrage is the game, the obvious question is: are batteries being built where the spreads are largest? They are not.
Houston has the lowest spreads but the densest pipeline
ERCOT has four competitive load zones: Houston, North, South, and West. In 2025, the average one hour daily real-time spreads in each were:
| Zone | 2025 spread | Rank |
|---|---|---|
| West | $106/MWh | Highest |
| North | $87/MWh | |
| South | $80/MWh | |
| Houston | $77/MWh | Lowest |
Table 1. 2025 average daily real-time spreads by ERCOT load zone.
If developers were chasing the highest current spreads, we would expect the largest pipeline in West Texas. Instead, the largest concentration of signed interconnection agreements (IA) sits in Houston, the zone with the lowest spreads in ERCOT.
Houston holds 8.7 GW of IA-signed battery capacity. West holds 6.8 GW. On the basic logic of arbitrage economics, this is the wrong way around.
The node-level data does not resolve the puzzle
Load-zone spreads can obscure local economics because batteries settle at specific nodes, not at the average zone price. A reasonable response to the zonal numbers above is that Houston might contain a handful of high-spread nodes hiding inside a low-average zone. If those exceptional nodes are where developers are siting, the paradox dissolves.
The data does not support that explanation. In 2025, Houston node averages were tightly clustered around the zonal mean. The median Houston node averaged roughly $78/MWh, the 90th percentile was around $80/MWh, and the highest Houston node averaged roughly $91/MWh. No Houston node exceeded the West zonal average of $106/MWh. The developer placing a 200 MW asset in Brazoria County is not selecting an outlier node with West-like economics. They are accepting Houston-zone economics.
Today’s fleet is balanced. The interconnection queue is not.
The clearest evidence that the pipeline reflects a forward bet, not a current-spread bet, is the contrast between what is operating today and what is signed for tomorrow. The current operating fleet is roughly balanced across the four zones. The forward-looking IA pipeline is heavily weighted toward Houston.
| Zone | Operating today | IA pipeline | Pipeline / fleet ratio |
|---|---|---|---|
| Houston | 2,312 MW | 8,699 MW | 3.8x |
| South | 4,226 MW | 10,051 MW | 2.4x |
| West | 4,067 MW | 6,761 MW | 1.7x |
| North | 4,952 MW | 8,634 MW | 1.7x |
Table 2. Operating fleet vs IA-signed pipeline by zone.
Houston’s operating fleet is the smallest of the four zones at 2.3 GW. Its IA pipeline is 8.7 GW. Developers are signed to nearly quadruple Houston’s battery capacity, while expanding the other three zones by roughly two-thirds. This is not incremental growth on top of an existing market. It is a structural transformation of the Houston zone.
Why Houston, and why now
Developers are betting on changes to the Houston grid that the spread data does not yet reflect. The case rests on three drivers grounded in CenterPoint Energy's published load forecasts and investor disclosures. CenterPoint’s service territory is not identical to the ERCOT Houston load zone, but it is the best public proxy for the demand growth affecting the Houston-area load pocket.
- Industrial load growth and reshoring. Energy refining, energy exports, and the reshoring of advanced manufacturing are expected to account for 40% to 60% of Houston-area demand growth. CenterPoint's CEO Jason Wells attributed the acceleration in its load forecast specifically to "reshoring of advanced manufacturing" alongside data center demand, calling it a diversified pipeline spanning more than a dozen customers across nearly 20 projects. Industrial load is the clearest single anchor for the Houston BESS thesis.
- Port and export electrification. Port of Houston electrification and EV adoption are expected to contribute another 10% to 20% of demand growth, adding load near the coastal industrial corridor. Houston, Freeport, and Sabine Pass LNG facilities continue to expand capacity.
- Data centers and large-load filings. CenterPoint attributes 30% to 35% of Houston-area demand growth to Texas Medical Center expansion and new data centers. As of Q1 2026, the utility expects to energize 8 GW of data center load by 2029, two years earlier than previous years forecast, and has a line of sight to 12.2 GW of firmly committed new load.
ERCOT’s own forecasts point the same way. The Long-Term Load Forecast (LTLF) is ERCOT’s annual projection of peak electricity demand over the next 5 to 10 years. It is published in two versions: a central case from ERCOT’s modelling, and a higher TSP-provided case built from the load growth that transmission service providers report seeing in their own queues, dominated by data centre filings.
We can use the LTLF to test whether the IA pipeline is sized for tomorrow rather than today. Storage intensity is a simple ratio: MW of signed BESS development divided by GW of zonal peak demand, indexed so the system average equals 100. Above 100 means a zone is more densely built than the ERCOT average. Below 100 means less densely built.
BESS developers are seeing the same future. Houston demand is growing quickly, but the battery queue is growing faster. That means the buildout is not just keeping pace with load; it is increasing the storage intensity of the zone. The implication is that developers are underwriting a long-term view of Houston as a durable flexibility market, where industrial load growth, congestion, and reliability needs could create revenue opportunities well beyond the current low-spread environment.
The risk is development timing
The argument above is a bet on demand growth and grid evolution. If those materialize on schedule, Houston spreads should widen, congestion should emerge, and the dense battery fleet finds revenue. If they do not, the bet inverts.
That risk is partly mitigated by the fact that Houston demand now appears to be arriving earlier than previously expected. CenterPoint has pulled its 50% peak-load growth forecast forward by two years, to 2029. The Houston thesis therefore depends less on whether load growth appears at all, and more on whether the BESS pipeline reaches COD in step with it.
THE CENTRAL RISK
If batteries arrive before demand and congestion materialize, Houston cannibalizes its own spreads.
Nearly 9 GW of new Houston BESS into a zone that today shows the lowest spreads in ERCOT is a meaningful supply addition. If industrial electrification, data center buildout, and transmission congestion arrive slower than developers expect, the new fleet competes for a thin arbitrage opportunity and revenues compress further. The thesis depends on demand growth and grid constraints arriving on time.
Houston’s pipeline is a coherent demand-driven bet, supported quantitatively by the LTLF and qualitatively by visible industrial expansion, port growth, and data center filings. The bet is defensible, concentrated, time-sensitive, and large.
Sources & methodology
Realized ERCOT real-time settlement-point spreads aggregated to load-zone and node-level monthly averages, January 2023 to November 2025. Used for the 2025 zonal spread figures in Table 1, the Houston node-level statistics, and the spread context throughout.
ERCOT public data
ERCOT Generator Interconnection Queue (GIM), Q4 2025. Source for the IA-signed BESS pipeline in Figure 1 (200 projects, 35 GW) and Table 2. Available at https://www.ercot.com/gridinfo/resource
ERCOT 2025 Long-Term Load Forecast (LTLF), published April 2025. Source for the 2030 demand projections (central and TSP-provided cases) used in Figure 3.
ERCOT 2024 Grid Reliability and Resiliency Assessment, Summer Hurricane Case load assumptions (January 2025). Source for the 2024 zonal peak demand figures (Houston ~23 GW, North ~38 GW, South ~28 GW, West ~16 GW).
ERCOT competitive load zone map. The official four-zone topology used as the basis for county-level zone assignment.
Realized ERCOT real-time settlement-point spreads aggregated to load-zone and node-level monthly averages, January 2023 to November 2025. Used for the 2025 zonal spread figures in Table 1, the Houston node-level statistics, and the spread context throughout.
Operating fleet data
US Energy Information Administration Form EIA-860 (operating generator data) and ERCOT operational reports. 223 operating BESS projects totalling 15.6 GW in ERCOT as of May 2026, used for Figure 2 and Table 2.
External research
Modo Energy. ERCOT and CAISO BESS revenue composition (ME BESS ERCOT index, ME BESS CAISO index). Source for the energy arbitrage share trajectory cited in footnote 1.
Modo Energy, ERCOT BESS revenues outlook: load growth, thermal retirements, and bridging solutions. Background context on the 2023 to 2025 revenue decline and Far West load growth dynamics. https://modoenergy.com/research/en/ercot-bess-revenues-outlook-load-growth-thermal-retirements-bridging-solutions
Modo Energy, ERCOT & CAISO BESS: The evolving revenue stack - June 2025, July 11, 2025. https://modoenergy.com/research/en/ercot-caiso-june-2025-revenue-stack-batteries-bess-energy-arbitrage-nodal-price-locational-marginal-price-transmission-congestion-price-spreads
Houston Chronicle. CenterPoint Energy demand-growth attribution for the Houston area, including industrial / port / data centre breakdown. Source for the percentage figures in drivers 1, 2, and 3, cited in footnote 2. https://www.houstonchronicle.com/business/energy/article/houston-electricity-demand-growth-20281415.php
US Energy Information Administration. ERCOT large flexible load growth context (data centres, cryptocurrency mining), 2024 to 2025 forecasts. https://www.eia.gov/todayinenergy/
Utility Dive, "CenterPoint to energize 8 GW of data center load by 2029," April 23, 2026. https://www.utilitydive.com/news/centerpoint-energy-data-center-load-earnings/818293/
CenterPoint Energy, Q4 and FY2025 Earnings Release, February 19, 2026. https://investors.centerpointenergy.com/news-releases/news-release-details/centerpoint-energy-reports-strong-q4-and-fy-2025-results-updates
Methodology
Pipeline data is filtered to battery projects (Technology = BA) with a signed Interconnection Agreement in the Q4 2025 ERCOT GIM queue, restricted to projects in counties mapped to one of the four ERCOT load zones. Operating fleet data is restricted to projects with a commercial-operation date on or before May 2026. ERCOT competitive load zones are defined by transmission topology rather than county lines; for this analysis we approximated each zone with its closest-fitting set of Texas counties, following the published ERCOT zone map. Storage intensity is calculated as MW of IA-signed BESS divided by zonal peak demand in GW, indexed to the system mean (= 100). Node-level Houston statistics are based on average daily real-time spreads at each settlement point within the Houston load zone, January to November 2025.
This report is for informational purposes only and does not constitute investment advice. Trading data Comity Inc’s proprietary book.