--- title: "58 Days to Break Ground, 10 Days to Build" slug: "cheapest-solar-schedule" language: "en" date: "2026-10-28T05:34:49.719Z" updated: "2026-10-28T05:34:49.000Z" description: "A 1.8 MW off-grid solar build went up in ten working days. The other 48 were permits, lead times and first-timer mistakes - and that is the lesson." canonical: "https://fix.slopit.io/cheapest-solar-schedule/" tags: ["energy", "construction", "engineering", "economics"] --- Terraform Industries built a 1.8 MW solar array, off-grid, in Southern California, and published the numbers: 58 days from closing on the land to permits and groundbreaking, 10 working days to install the array, 50 days in the field, 2,984 panels, $1.3 million all-in including land and site work, and $0.016/kWh amortized. They call it the cheapest solar in North America. The number to be careful with is the last one. Amortized power cost for a single off-grid R&D site, with land included and no storage in the denominator, is not the same as LCOE, and a project built by the company that will consume the power is not a market price. Treat $0.016/kWh as an internal cost, not a benchmark. The article is worth reading for a different reason: the schedule, and what it cost to compress it. ## Construction was the easy part The array went up in ten working days. Nearly everything else was permitting, sequencing, and the cost of being a first-timer. The timeline: pick land that is flat, secluded, fenced, near a town, and zoned heavy industrial. Make an offer, and submit the conditional use permit as soon as the tract goes into escrow, so you learn whether the project is dead before the diligence period closes. Split the project into multiple permits and submit them in parallel so one piece can't hold up the rest. Contract the long-lead item — the solar array — first, because in 2026 you're competing with datacenter builders for the same electronics. That last point is the one that decides schedules. The array was the longest lead time, so it was ordered first. Everything else was arranged around its arrival. ## The three mistakes, priced The most useful part is the post-mortem on what went wrong, with a number attached: about $60,000 attributable to errors and first-timer fees. **Too collaborative with one contractor.** They let an electrical contractor influence the system architecture, discovered the ideas weren't good and the execution wasn't there, and replaced them. The lesson is stated well: form an opinion before engaging a contractor on architecture, and always have a backup. **A large capital order before the permit.** One system needed 3-phase AC power rather than HVDC, which pushed them into a much larger battery class, which triggered a far more rigorous fire-department review in a county with a history of large battery fires. The permit was denied even though the equipment met the National Electrical Code and California Electrical Code, because the county can demand a higher standard. That forced a redesign and a sunk cost. **Design churn.** The systems weren't finalized when construction started. The fix is the most transferable idea in the piece: they standardized around a shipping-container footprint (8 by 20 feet) and overbuilt the structural steel to absorb changes in the equipment. As the author puts it, sometimes the best thing you can do for engineers in a design phase is give them constraints to build around rather than a moving target. That's a software lesson wearing a hard hat. A stable interface with margin absorbs churn; a bespoke fit to current requirements breaks on the first change. ## The unglamorous thesis The essay's argument is that "pleasant persistence" works. They logged 2,500 miles of visits, collected business cards, followed up on submittals, brought cards and chocolates at Christmas. A permit was delayed a month because the one employee who manages it was out and the paperwork wasn't filed. Accelerationists like the author tend to vilify permitting. His conclusion is better: the bureaucracy is made of normal people, most of whom don't want to block you, and the binding constraint is often attention rather than policy. That's not a reason to like the process. It is a reason to call the office. ## What actually made it fast Three things, I think, and none of them is "move fast." **Parallelism.** Multiple permits, three structural engineers, an electrical engineer, all submitted concurrently. The critical path was the slowest approval, not the sum of the approvals. **Sequencing around lead times.** Order the thing that takes longest first, even before you can build anything. **Standardization under uncertainty.** Pick a footprint and overbuild it, so design changes don't force new site work. The $60,000 in errors is the cost of learning those lessons in public. The schedule is what they bought with it. ## The caveat A single project, built by a motivated company on land it selected, with an in-house consumer for the power, is a proof of concept, not a general cost curve. The claim that any new US solar build takes four years and $1.40/W may be wrong, but one 1.8 MW array doesn't establish the new number either. What it does establish is that the schedule is negotiable. The construction took ten days. The other 48 were bureaucracy, lead times, and learning — and those are the parts a repeat builder gets to keep. *Source: [How Terraform built the cheapest solar in North America — Matt Weickert, Terraform Industries](https://terraformindustries.wordpress.com/2026/10/27/how-terraform-built-the-cheapest-solar-in-north-america/).*