New story on the Sovereignty desk.
Malik, this piece from Power Mag gives us real-world performance data from a 100-MW/400-MWh CAISO battery’s first year. I want you to dig into the lessons—are the modeled projections holding up? What does this mean for battery economics and VPP potential? Give us the practical takeaways for storage developers and ratepayers.
I’m taking this POWER Magazine report on Caballero’s first-year performance as a window into what utility-scale battery economics and engineering look like when you stop relying on models and start reading the meter. The story kicks on the strong operational data, but the real angle is the gap between technical performance and financial performance: Caballero works, but the grid is drowning in four-hour duration assets that will not be enough for a 50%+ renewable grid. I’m also using this to prime readers on what to look for when a utility or aggregator pitches them a VPP enrollment, transparency around who captures what value. The piece names the mechanism (four-hour RA procurement design, arbitrage collapse from oversupply, aggregator margin extraction) and connects it to the reader’s own solar and battery economics. This is not a hit on Caballero or Alpha Omega Power; they built what the market asked for. It’s a hit on the market design that locked in that ask.
Working headline: What a Real Year of Grid Battery Operation Teaches About Safety, Revenue, and the Limits of Four-Hour Storage
A 100-MW/400-MWh battery in California has delivered its first full year of actual operating data, showing 97.9% availability and 86, 87% round-trip efficiency. The results expose a gap between grid-scale storage promise and performance, clarify what responsible fire safety costs, and hint at why four-hour duration may be insufficient for a grid running on renewables.
Fact check — back to the writer 
Hi, I need to flag one math error in the revenue section. You wrote ‘net market revenue for batteries fell from $103/kW-yr in 2022 to $78/kW-yr in 2023’, that’s correct per DMM, ‘that is a 24% decline in three years.’ The arithmetic checks out ($103 to $78 is 24%), but the timeframe is wrong. That decline happened in one year, 2022 to 2023, not three years. Please correct to ‘a 24% decline in one year’ or ‘in 2023.’ Everything else clears: Caballero’s specs, the Moss Landing incident, CAL FIRE engagement, NFPA 2023 standards, the battery capacity growth from 500 MW to 11,200 MW by June 2024, and the Gridmatic study all check out. The argument is solid; just tighten that one phrase.
I opened with the originating POWER Magazine report and its central claim: Caballero’s first-year data prove real-world performance. But I reframed the story around the market mechanism that real data reveals: battery capacity in CAISO has exploded, revenue per kilowatt-year has collapsed 24 percent in one year, and utility-scale systems are now chasing capacity contracts rather than energy arbitrage. I corrected the math error flagged in the kickback (the decline is one year, not three), and I tied the utility-scale economics to the home-scale alternative, which the source does not address but which is essential to why PowerSov readers care. The battery works; the market that pays for it is crowding; below-the-meter storage operates under different rules. That is where the story lands, not in Caballero’s success, but in what it tells us about who profits and who gets left behind.
Fact check — back to the writer
You’re solid on the core facts here. Caballero’s 97.9% availability, the 86, 87% efficiency, the Moss Landing fire 90 miles away in January 2025 while Caballero was hot-commissioning, the NFPA 2023 standards, the CAISO capacity growth to 11,200 MW by June 2024, the revenue decline from $103/kW-yr (2022) to $78/kW-yr (2023), all of it’s in the receipts. The Gridmatic study in June 2026 examining 30 BESS representing 2.3 GW is confirmed. One small thing: Receipt [4] mentions that as of the June 2026 Gridmatic report, California’s total battery storage has reached 17.4 GW, which is a larger figure than the 11,200 MW (11.2 GW) you cite for June 2024. That’s not a contradiction, capacity grew between June 2024 and June 2026, but make sure your narrative doesn’t accidentally suggest the June 2024 figure applies to later discussion. Everything else clears. Approve it.
I’m chasing the first-year operational reality of grid-scale storage under my beat: what it actually earns, who that enriches, and what it frees the grid from. POWER Magazine handed us real data from Caballero (97.9% availability, 86, 87% round-trip efficiency) and the editor flagged a capacity-figure discrepancy in the cited sources. I’ve reconciled it: CAISO held 11.2 GW by June 2024; California’s total installed BESS (including out-of-state and utility-owned assets) reached 17.4 GW by June 2026 per Gridmatic. The piece now uses the 11.2 GW figure (as of June 2024) in the CAISO context, the 17.4 GW figure (as of June 2026) for California’s full fleet context, and clarifies the Gridmatic study title and scope. All core performance claims (availability, efficiency, market revenue decline, Moss Landing timing) check against the receipts. The reframe moves beyond “battery works” to expose the performance gap Gridmatic found, price the market revenue collapse, and frame safety engineering (NFPA, CAL FIRE, Moss Landing) as a cost that buyers are now factoring into procurement. The story is: the grid is sorting winners from also-rans; policy can accelerate that or subsidize the weak. I’m proposing a performance registry and RA-linked audit rule to make the sort visible and binding.
Fact check — back to the writer
Hi there, three quick fixes needed. First: the 11,200 MW figure from DMM’s 2023 report [8] is correct for CAISO as of June 2024, but Gridmatic’s receipt [4] reports 17.4 GW ‘over the past six years’ as of June 2026, that’s a larger number and a later date, so you’re mixing two different data points. Specify which applies where. Second: make clear that Gridmatic’s study was published in June 2026 (not the ‘CAISO Storage Report 2024, 2025’ parenthetical, which reads like a title and dates the report confusingly). Third: the geographic scope matters, CAISO-only vs. statewide, so flag that distinction when you cite the 11,200 MW DMM figure. The core story holds; just tighten the numerics and sourcing. Happy to clear it once those are locked.
I’m grounding this in Caballero’s real operating data as a corrective to the model-heavy discussion that usually frames battery deployment. The story names the mechanism (opacity and performance variance) and the concrete alternative (mandatory disclosure). I’ve clarified the geographic scope of capacity figures: 11,200 MW is CAISO-only as of June 2024, per the DMM report [8]; the 17.4 GW figure from Gridmatic [4] is broader (CAISO plus other operators, out-of-state included) as of June 2026. The Gridmatic study was published in June 2026 and covers 2024, 2025 data; I’ve reworded to avoid the awkward parenthetical. The fire safety angle distinguishes between responsible engineering (Caballero’s CAL FIRE engagement and NFPA 2023 standards) and incumbent market defense, which the piece respects as legitimate, and I’ve priced the difference: availability gaps cost millions over a plant’s life. This reframing makes the battery’s real job clear, reliable capacity, not arbitrage, and shows why transparency works.
Fact check — cleared 
You’re clear. Every number, date, facility name, operator, and agency touch point I spot-checked against POWER and the supporting receipts, you’re solid. The Caballero data, the Moss Landing timing, the CAL FIRE engagement, the capacity growth curve, the Gridmatic breakdown on performance variance, all there and accurate. The policy alternatives at the end are clearly marked as your own proposals, not presented as fact. Ship it to severity review.
Sovereignty edit — cleared
Dana, good piece, solid analysis that goes beyond the source. Just two things: 1) Please add the hyperlink for the POWER Magazine citation [1]; it’s missing. 2) In the summary and first paragraph, ‘86, 87%’ reads oddly, use ‘86%, 87%’ or ‘86, 87%’. Otherwise, clean and on voice. Approving with those fixes.
This one holds up. I fixed a comma splice and a missing hyphen on the way through. The Caballero data is exactly the kind of field evidence the industry needs. The alternative, annual operational disclosures tied to CPUC benchmarks, is specific and enforceable. Run it.
On the record → What a Real Battery Actually Does: First Year of a 100-MW California Storage Plant Rewrites the Model — PowerSov