The merchant solar caricature
The standard objection to merchant solar is real but incomplete. A standalone solar plant sells into the hours when every other solar plant is also selling, which is precisely when the marginal value of energy collapses. In high-penetration grids the midday clearing price can fall toward zero or negative, so the plant earns its worst prices exactly when it produces most. Capture price - the volume-weighted price the asset actually realises - drifts well below the average spot price, and the gap widens as more solar is built.
This is the phenomenon that makes allocators demand a contracted offtake before they will fund solar. The caricature holds for a bare solar plant. It does not hold once a battery is bolted onto the same interconnection.
How storage rebuilds the revenue stack
A co-located battery converts a single degraded revenue line into three distinct ones. The first is temporal arbitrage: charge when midday prices are depressed, discharge into the evening ramp when prices spike as solar drops off and demand peaks. In systems with a steep duck-curve the spread between midday and evening prices can be substantial and, crucially, structural rather than incidental - it is created by the same solar build-out that hurts bare solar. The battery monetises the cannibalisation that damages the standalone plant.
The second line is capacity or firm-power value, where the market or a bilateral counterparty pays for dispatchable capability at peak. The third is ancillary services - frequency regulation and reserves - which pay for fast response rather than energy volume and are often the highest per-MWh revenue available, albeit in a thinner, saturable market. Revenue stops being a single number multiplied by a collapsing capture price and becomes a portfolio of streams with different and partly uncorrelated drivers.
Co-location also captures real structural savings: a shared interconnection, shared land and shared balance-of-plant lift the blended return above what standalone solar and standalone storage would earn separately.
What diversification does to the downside
The investment point is not that storage raises the expected IRR, though it typically does. It is what storage does to the shape of the distribution. A bare merchant solar plant has a return distribution with a long, fat left tail: its downside case is a grid so saturated with solar that daytime prices stay near zero for years. That same saturation is the battery's upside case, because it widens the arbitrage spread the battery harvests. The two assets are negatively correlated on the single risk factor that matters most, so combining them compresses the left tail even if the central estimate moves less.
Modelled honestly, the storage-linked configuration narrows the band around the return. Illustratively, a standalone merchant asset we might underwrite in the high single digits with a wide dispersion becomes, with co-located storage, a low-to-mid teens central case with a distinctly tighter distribution.
The risks that remain
None of this makes the asset riskless. Battery degradation is a genuine cost that must be reserved against; augmentation capex to maintain contracted capacity over the asset life is a real line item, not a footnote. Ancillary-service markets are shallow and saturate quickly as more storage arrives, so revenue assumed from that stack should be modelled as decaying. Merchant price forecasts remain the largest single uncertainty.
The disciplined conclusion is that co-located storage does not eliminate merchant risk - it restructures and diversifies it. For patient capital willing to underwrite a portfolio of revenue streams rather than a single contracted offtake, the storage-linked merchant asset offers a return profile that is both higher and, more importantly, less fragile than the standalone caricature.