Renewabl's model puts the cheap cross-border vPPA 29 percent above in-country deals in a bad year

A Spanish solar contract pays out against Spanish midday prices. The data center it covers pays a French or German price at four in the morning in February ^[7]. When those markets move apart the contract stops covering the bill, and the gap opens widest in the conditions the hedge was bought for: a still Continental evening as Nordic wind fades, or a 2022-style gas shock that lifts every market but not in step ^[8].

The annual report says the operator bought a year of renewable volume, retired against a year of consumption, 100 percent ^[9]. What it actually consumed was Spanish generation during Spanish daylight, then the spot market overnight and through the winter ^[9]. Renewabl puts hourly coverage for the cross-border structure at about 14 percent, against about 82 percent for an optimized in-country wind and solar mix ^[15]. Those are 68 percentage points of load hours one contract covers and the other does not ^[20].

On cover price the structures look identical. Across 1,000 ten-year price paths built on Pexapark data, the cross-border vPPA, a set of in-country solar contracts and the optimized mix all landed within about 1 euro/MWh of each other, near 50 euros ^[12]. On the modeled 100GWh load, 50 euros/MWh is about 5m euros of power a year ^[18]. Renewabl ran the model itself and published it as an opinion piece on Data Center Dynamics ^[22].

The separation sits at the P10, the outcome bad enough that only one path in ten comes out worse ^[11]. In-country procurement removed 85 to 91 percent of the ten-year price uncertainty; the cross-border deal removed about 40 percent ^[13]. In the bad year it cost roughly 29 percent more than the optimized in-country portfolio ^[14]. That gap is given as a percentage, and 29 percent of the 50 euro median is 14.50 euros/MWh, or 1.45m euros a year on a 100GWh load. Because the bad-year price per MWh sits above the median by definition, 1.45m euros is a floor ^[19].

The 2027 labels are why this comes up now ^[2], but on this evidence the label is not where the hourly gap gets graded. The author's own complaint about the June roadmap is that it deals mostly with efficiency and grid connection and says little about how operators buy their clean power ^[4]. What the rating scheme will measure is not in the piece ^[23]. The pressure on a 14 percent score comes from accounting instead: under today's annual Scope 2 rules the two portfolios are equivalent, and under the hourly accounting the GHG Protocol has proposed they are not ^[16].

The author's point is that an operator's hourly-matching score measures two things at once: how credible the clean-energy claim is and how exposed the bill is ^[17]. Strike price shows only the first. The second axis is the share of load hours the contract actually covers, and a cheap contract with low coverage is the one whose annual report and whose P10 disagree; 29 percent is what that disagreement cost in Renewabl's bad year ^[14].