Why field data is weeks old by the time anyone acts on it

A thermal flight over a utility-scale solar plant takes hours. The decision it should inform, which strings to send a crew to and in what order, usually comes weeks later. Public industry figures show where that time goes and what would shorten it.

01Before the flight

Most utility-scale plants are flown once or twice a year; one 2026 industry guide recommends twice a year for sites above 10 MW ^[1]. A fault that appears the week after a flight can run until the next one. Raptor Maps' 2026 report, covering 373 GW of US solar, found that sites inspected quarterly had 36% lower power loss than sites inspected once a year ^[2]. That is a correlation across sites, without a controlled test behind it. Read with the cadence figures, it puts the longest wait in the cycle before the flight, in the months between inspections.

02After the flight

Once the drone lands, the work moves to the office. One guide for inspection operators puts total turnaround for a utility-scale site at five to ten business days from inspection to final report, and finds that processing and report generation take 40 to 60% of total project time ^[3]. Another puts AI-assisted reporting at 24 to 72 hours and manual-only review of large sites at a week or more ^[1]. Weather adds further delay. Valid thermal data needs at least 600 W/m² of irradiance ^[1], and operators should expect 20 to 30% of scheduled flight days to be lost to conditions ^[3].

Then the report has to become work. A common severity scale asks for further testing of moderate anomalies within 30 days ^[3]. A finding that is classed, filed, ticketed and scheduled can wait at each step in turn, and the average technician now covers 70% more megawatts than five years ago ^[2].

03Why it stays slow

The delay accumulates at the hand-offs. Each step waits for the one before it, in a different tool, owned by a different team. The IEA's survey of energy companies names the lack of digital skills as the largest barrier to AI adoption in the sector, alongside fragmented data ^[4]; in industrial operations, data quality is the obstacle cited most often ^[5].

04What would close the gap

• –Read the data where it lands. Processing on site, next to the drone dock, removes the upload and the queue behind it.
• –Tie every finding to a component. A finding located to a module and a string can go straight into a work order. Standard drone GPS is accurate to two to five metres; RTK correction to one or two centimetres ^[3].
• –Rank by consequence. A crew can work a handful of findings a day. Ranking them by the energy each fault costs sets the order of that work.
• –Look more often, where it matters. Docked drones make partial, targeted flights practical ^[2], which turns inspection planning into a choice of where to look next.

Lumerian Core is built to close that gap. Aperture reads each flight against the site's surveyed layout, assigns every detection to a module, and creates a follow-up automatically when an expert verifies a finding, so the next campaign reads it. Ranking by energy cost switches on once a site supplies energy data; until then, a labelled baseline stands in.