Someone sold your city a projection. Who checked it against what the array actually made?

By Olivier Beauchemin, founder · September 2026

The gap in every public solar deal

A city, county, school board, or public agency signs a power purchase agreement or a lease for a solar array on public land. The deal is sized around a projected annual output, in gigawatt-hours or megawatt-hours, that the developer put on a spec sheet or in a council presentation years before a single panel went up. Once the array is built, the PPA meters what gets delivered and the utility or offtaker pays for it, month over month, for the life of a 20-year contract. That metering answers "how much did we buy this month." It does not answer "did this array make what it was supposed to make, and who is checking."

The same dynamic that costs a homeowner a few hundred dollars a year on a residential guarantee costs a public body far more, because the numbers are bigger and the review is thinner. A facilities team with a full docket, an O&M vendor paid to keep panels running rather than to audit them against the sales pitch, and an operator who owns the plant have no reason to independently reconcile the deal. OwlWatt is built to be that independent party. This is the same reconciliation OwlWatt runs for residential production guarantees and for fleet owners (see what we do for fleet owners), sized for a municipal PPA instead of a homeowner's contract.

A public example, with the arithmetic shown

We are not a client of the array below. Every number in this section is public record, cited to its source, and we were able to verify the plant specifications and the 2025 output figure ourselves from two federal filings rather than take a news article's word for it.

The Annapolis Renewable Energy Park sits on a capped municipal landfill in Annapolis, Maryland. It is fixed-tilt solar on concrete slabs (the landfill's cap cannot be penetrated by driven posts), and it has been operating since August 2018.

We pulled the 22,897 MWh figure ourselves from EIA's own monthly generation workbook rather than rely on the trade-press coverage that first surfaced this example, which cited the same EIA data as compiled by a third party. It matches.

Here is what nobody appears to have reconciled: the city's own project page states an expected annual range of 23.8 to 27.4 GWh. Actual 2025 output, 22.9 GWh, comes in below the bottom of that range. A trade-press piece covering this array separately describes "roughly 24 GWh" as the developer's original projection when the park opened in 2018; we could not independently trace that specific figure to a named PPA exhibit, feasibility study, or council document, so we are not asserting it as confirmed. Both readings point the same direction. Depending which number you treat as the one the park was supposed to hit, the size of the miss moves:

At the city's own 7.75¢/kWh PPA rate and its 33% offtake share, the value of that gap works out to roughly $23,000 at the low reference and $69,000 at the midpoint, for one year, on a contract that runs 20 years. That is not money the city overpaid or is owed back; a PPA meters and pays for what is actually delivered, so a smaller harvest simply means a smaller bill. It is the value of clean, budgeted-rate energy the city planned to receive and did not, and the point is not the dollar figure, it's that the range spans three times depending which projection number you pick, and as far as we can find in the public record, nobody has published which one is the one that governs.

To be direct about what this is and is not: a shortfall in this range is unremarkable for a fixed-tilt array and is plausibly ordinary weather variance, not equipment failure or a park underperforming its design. The city's own page frames the plant's target around a ~20% AC capacity factor; 2025's actual output works out to about 21.8% (22,897 MWh ÷ 12 MW × 8,760 hours), a bit ahead of that target. Per-panel yield, using the plant's actual DC nameplate from its panel count (52,416 × 345 W = 18.08 MW DC), comes to about 1,266 kWh per kW DC for the year, in line with what a fixed-tilt array at this latitude produces. Every sign here points to a healthy plant. We are not asserting fault, a broken component, or that the city is owed anything. We are pointing out that a public body signed a deal sized on a specific number, and as far as public records show, nobody has been tasked with checking actual output against that number on an ongoing basis and saying, in public, whether the difference is weather or something worth a closer look.

What a design-partner engagement looks like

1. Scope and access. You point us at one array: site details, monitoring platform, and the PPA or lease documents that state a production figure. OwlWatt does not buy or hold any hardware, ever. See how we approach commercial sites in general at our fleet-owner page.
2. Reconciliation. We compare actual output, read from the site's own monitoring platform, against every production figure we can find in your contract and public documents, with the source cited for each one.
3. The deliverable. A written reconciliation, the methodology behind it, and a public case study with your agency's name attached, published on owlwatt.com.

Today's live data source is Enphase. String and central inverters (SolarEdge, SMA, Fronius, and other SunSpec platforms common on municipal and commercial roofs) are on our roadmap and not live; we would rather assess feasibility on a call than sell an engagement we cannot run on your site's actual equipment.

Four conditions we hold on ourselves

Honest limits

Questions public-sector buyers ask first

Become our first municipal design partner →

OwlWatt measures and documents production against a projection or contract figure. It does not determine breach, recover money, or provide legal advice. Read more on the Annapolis example in our companion write-up.