Besós and San Roque: same owners, opposite outcomes

Besós 4 and Besós 5 share gas pipes, a grid connection, and a postcode in Sant Adrià de Besós. Unit 4 is Naturgy’s. Unit 5 is Endesa’s. After the Iberian blackout, one gained; the other collapsed. But there’s a third unit nobody talks about — BES3, also Endesa’s — and 300 km south, the same two companies own another complex where the roles reverse.

A decade of data. From 2014 to 2024, all three Besós units moved in a similar band — 3-10 €M/month in technical constraint revenue, with occasional spikes during the 2021-2022 energy crisis. The blackout of April 28, 2025 breaks the pattern. BES4 (Naturgy, green) holds at 10-17 €M/month. BES5 (Endesa, red) drops to 0-2 €M by late 2025. And BES3 (Endesa, blue) — the unit nobody mentions — rises to 11-16 €M/month.

Two Endesa units. Same owner. Opposite trajectories. The company name doesn’t explain this.

Three turbines, two owners, one complex

The Besós complex sits on the coast north of Barcelona, at the mouth of the Besòs river. Three combined-cycle gas turbines, built in the early 2000s:

• BES3 — Endesa Generación. Active in I90 data since January 2014.
• BES4 — Gas Natural Comercializadora (Naturgy). Active since January 2014.
• BES5 — Endesa Generación. Active since January 2014.

Same physical site, same gas supply infrastructure, same grid node, same ambient conditions. The only structural difference is ownership — and within Endesa’s own two units, the trajectories diverge sharply post-blackout.

The program-level breakdown reveals what changed. Pre-blackout, all three units earned primarily through PDVP (day-ahead technical constraints) and RTR (real-time constraints), with BES3 and BES5 also carrying significant BS (secondary balancing) and PDBF (day-ahead market) volume. Post-blackout, PDVP grows for BES3 and BES4 — Operación Reforzada channels more constraint dispatch through the day-ahead process — while BES5’s PDVP and RTR both shrink dramatically.

The divergence isn’t in price — it’s in volume. REE stopped calling BES5 for routine constraint resolution and shifted that dispatch to BES3 and BES4. Cheap to schedule beats expensive to activate.

When does REE call each unit?

The hourly profile makes the shift visible.

BES4 (Naturgy) pre-blackout showed a classic bimodal pattern: morning ramp (6-9h), midday trough (solar hours), evening peak (17-22h, prices up to 220 €/MWh). Post-blackout, the profile flattens — dispatch spread across all 24 hours, 25-40 GWh per hour, prices compressed to 155-185 €/MWh. BES4 has become Barcelona’s baseload constraint unit.

BES5 (Endesa) tells the opposite story. Pre-blackout, it had broad dispatch across most hours. Post-blackout, volume drops sharply (15-28 GWh vs 30-60 GWh pre). REE simply stopped calling BES5 for routine constraint resolution.

The price distribution confirms this is a volume story, not a price story. All four distributions (both units × pre/post) cluster in the 100-300 €/MWh range with similar medians (BES4 pre: 197, BES4 post: 153, BES5 pre: 181, BES5 post: 158 €/MWh). What changed isn’t what REE pays per MWh — it’s how many hours each unit gets called. BES4 post-blackout gets thousands more dispatch hours. BES5 post-blackout gets far fewer.

San Roque: same owners, opposite result

300 km south, in the Campo de Gibraltar near Algeciras (Cádiz), sits another multi-owner CCGT complex with the identical ownership split:

• SROQ1 — Gas Natural Comercializadora (Naturgy). Barely dispatched: ~€10-12M/year in technical constraints.
• SROQ2 — Endesa Generación. Heavily dispatched: ~€85-90M/year.

The mirror is immediate. At Besós, Naturgy’s unit dominates and Endesa’s collapses. At San Roque, Endesa’s unit dominates and Naturgy’s is barely visible. Same two companies, same ownership structure — reversed outcome.

The explanation is geographic. San Roque sits in the southwest cluster — Cádiz, Huelva, the zone where every CCGT wins regardless of owner. The constraint flow there is structural: massive solar photovoltaic penetration in western Andalusia creates a daily need for synchronous capacity, and REE activates all available units. SROQ2 dominates SROQ1 not because of commercial strategy but because of capacity, availability, and operational history — in a cluster where geography overwhelms everything, the larger unit gets more dispatch.

At Besós, Barcelona is a mixed cluster where the owner and pricing strategy determine the outcome. Geography puts you in the game; the operator decides if you win or lose.

Bilbao: independent operators, same dynamic

Vizcaya offers a third variation. Amorebieta (Axpo/Castleton, +€81M post-blackout delta) and Bahía de Bizkaia (Gunvor+EVE joint venture, -€63M) — two separate plants, two independent operators, no shared infrastructure. Yet the same pattern: one unit captures the constraint flow, the other doesn’t.

This case was profiled in detail as part of the post-blackout series. The differentiator there is commercial aggressiveness: Axpo operates Amorebieta with a trading-house mentality, offering competitive constraint prices; BBE’s joint venture structure creates different incentives. Not geography, not ownership structure — operating philosophy.

What the data says about multi-owner complexes

Three complexes, three variations of the same thesis:

Complex	Location	Cluster type	Who wins	Why
Besós	Barcelona	Mixed	Naturgy (BES4)	Pricing strategy: BES4 offers cheap day-ahead; BES5 priced as emergency peaker
San Roque	Cádiz	Uniform (south)	Endesa (SROQ2)	Geography: all units win; larger unit gets more dispatch
Amorebieta/BBE	Bilbao	Mixed (coast)	Axpo (Amorebieta)	Commercial aggressiveness: trading house vs conservative JV

In uniform clusters, the owner barely matters — geography selects everyone in. In mixed clusters, the owner is everything. The unit that prices itself as cheap schedulable capacity wins under Operación Reforzada; the unit that prices itself as an expensive last-resort loses.

The broader implication: the value of a CCGT is no longer about megawatts or location alone — it’s about how the operator positions the unit in the constraint market. The same company (Endesa) wins at one node and loses at another. The same ownership split (Endesa/Naturgy) produces opposite outcomes depending on the grid topology around the plant.

For anyone modeling the economics of Spain’s energy transition — including the case for energy storage at expensive nodes — this unit-level data is where the signal lives. Aggregate fleet numbers hide more than they reveal.

---

The data is open at ESIOS Data — query any unit, any complex, any program.