A constrained water treatment plant site became a live test of Earth Mount Solar architecture, running at 101% of expected output.
By Keith Reynolds | Publisher & Editor, ChargedUp!
862 megawatt-hours. That is how much energy ForeFront Power's new solar array delivered to the City of Fresno's Department of Public Utilities (DPU) in the months after the system entered commercial operation in late March 2026, running at an average of 101% of expected production.
The array sits at Fresno's Northeast Surface Water Treatment Plant, and it is the first project to pair ForeFront Power with Erthos' direct-to-ground Earth Mount Solar design at this scale. For owners evaluating onsite solar on constrained parcels, campuses, or industrial sites, Fresno offers a working example of how mounting architecture, not just panel cost, determines whether a project pencils.
Fresno Turns a Small Site Into a Design Test
The Northeast Surface Water Treatment Plant was the smallest of three DPU sites in a broader solar and storage portfolio, and a conventional ground-mounted array would have required extensive grading, steel racking, and civil work to fit the parcel. ForeFront senior project manager Erinne Davis said the site's constraints pushed the company toward Erthos' Earth Mount Solar system, which places 3,132 EliTe Solar 550-watt modules and Chint inverters directly on the earth rather than on elevated steel racks and trackers.
That substitution changed the capital math. ForeFront reported savings from avoiding structural steel, translating into both lower utility bills for the city and a stronger return on investment for ForeFront, which owns and operates the system under a 20-year power purchase agreement (PPA) at a fixed rate below the utility's price. The arrangement shifts the city's energy economics from a one-time equipment purchase to a long-duration operating contract, and it shifts ForeFront's economics toward keeping production and maintenance costs aligned with that PPA price for two decades.
The Portfolio Sets the Financial Frame
The Fresno array is one piece of a 27-megawatt-DC DPU solar and storage portfolio across three sites, expected to generate 47 million kilowatt-hours annually and to save more than $122 million in ratepayer dollars by 2045. That scale matters for how the numbers should be read. Dividing 47 million annual kilowatt-hours by 27,000 kilowatts DC works out to roughly 1,741 kilowatt-hours per installed kilowatt, a capacity-normalized planning figure that should not be read as a guarantee for any single site, but that does indicate the output behind the public savings claim.
For municipal and institutional owners weighing a similar portfolio approach, the lesson is that a single constrained site does not need to carry its own financial case in isolation. Bundling a hard-to-build parcel into a larger multi-site program can absorb the added coordination cost of a first-of-its-kind installation while still meeting the portfolio's blended return target.
Earth Mount Solar Changes the Construction Equation, Not the Discipline
Direct-to-ground installation reduced some construction costs at Fresno while introducing new execution requirements elsewhere. The installation crew trained at Erthos' headquarters before mobilizing, and Erthos provided onsite support during initial module installation. At peak, 12 union workers worked the site, and the project took longer than expected because it was the crew's first Earth Mount Solar deployment.
The most consequential challenge came from the ground itself. Davis said the team had to grade the site to flatten it while preserving the drainage characteristics of the surrounding facility, and heavy rains delayed that work. At a water treatment plant, drainage is not a cosmetic detail. It affects site resilience, maintenance access, erosion control, and the continued function of adjacent infrastructure, and it is exactly the kind of constraint that separates a portfolio-ready mounting decision from a purely cost-driven one.
ForeFront also encountered design-stage coordination friction between its engineering, procurement, and construction team and Erthos' team, along with logistics and procurement delays. The company said it expects to move faster on future projects once the lessons from Fresno are standardized into repeatable procedures, which is the execution signal worth tracking: first deployments of an unfamiliar architecture carry coordination costs that should decline with repetition, and the pace of that decline will determine how much of the avoided structural steel savings survives into future projects.
Operations Shift to Cleaning and Performance Discipline
Modules placed directly on the ground face more soiling exposure than elevated arrays, and Fresno addresses that through an Energy Services Agreement under which Erthos performs regular cleaning with its ErthBot autonomous cleaning robot, a dry nylon brush system rated to clean more than 3.5 megawatts DC of array, roughly twice Fresno's 1.72-megawatt-DC nameplate capacity. The PPA's economics depend on keeping production close to expectations over 20 years, and robotic cleaning converts what would otherwise be a soiling risk into a defined, contracted service function.
The Repeatability Question Sets the Next Test
ForeFront said it is evaluating further opportunities to deploy Erthos after completing its first Earth Mount Solar project, a cautious posture that fits the evidence. Fresno shows the architecture can work on a constrained municipal site. Whether it becomes a broader procurement option depends on three things: construction velocity improving on the next deployment, site selection that favors parcels where conventional ground-mount triggers expensive grading or steel work, and operating discipline that keeps the ErthBot cleaning model from becoming its own maintenance burden.
For owners of municipal facilities, campuses, and industrial sites facing similar land constraints, the diligence question is not equipment familiarity. It is site-specific delivered cost and long-term operating performance, measured against what a conventional ground-mount array would have required on the same parcel.
Sources