Solar farm output depends not only on panel conversion efficiency but also on how much sunlight clouds let through. Meteoric, a Y Combinator Summer 2026 startup, has unveiled a plan to boost the annual output of existing solar farms by acting directly on those clouds using drones. Tom's Hardware reported on August 23 on this concept, along with the company's ultimate goal of weakening storms and hurricanes.
The company's pitch includes a figure claiming annual output could increase by 10–30%. However, this number isn't based on actual generation results from natural clouds. It's an estimate from a "cloud loss model" that uses region-specific cloud types, with 30% representing the model's upper bound. Separating the published test results from flight regulations and cloud physics reveals that what Meteoric is offering isn't so much an add-on device for power plants as a weather intervention whose validation design has only just begun.
The 30% Figure Is a Model Ceiling, Not a Measured Result
Meteoric envisions flying dozens to hundreds of autonomous drones into low- and mid-level overcast skies to alter cloud droplets and reduce cloud reflectivity—without dispersing any chemicals. The goal, according to the company, is to redirect sunlight that would otherwise be reflected by clouds back toward solar panels. The company models that this could boost annual output by 10–30% across major U.S. grid regions, valuing this at $5,000–$28,000 per MW annually.
The 30% figure cannot be read as a current performance metric. It's a region-specific estimate derived from a cloud loss model. The model's methodology, input data, and baseline year have not been disclosed. Nor have the success rate of the treatment process, flight duration, or confidence intervals. Therefore, the 10–30% range isn't uniform across the U.S., isn't a constant increase, and isn't a guaranteed value for customers.
The only published test is the company's claim that a prototype dispersed 13% of an artificial cloud in a test chamber. The dimensions of the test apparatus, the cloud's liquid water content, and what exactly the 13% measured haven't been disclosed. Neither have control trials, repetition counts, nor peer-reviewed materials. At this stage, it's impossible to determine whether cloud area, optical thickness, or liquid water content changed—or some combination.
The roadmap must also be treated separately from actual results. Meteoric targets large-scale cloud-clearing flights for 2027, with the first storm flights planned for late 2028. Flights to verify output gains using natural clouds are still ahead.
Does Altering Cloud Droplets Actually Return Sunlight?
There is a baseline observation for how much clouds reduce power generation. The 1980 paper by Kasten & Czeplak, which Meteoric cites on its Y Combinator profile page, reported global solar radiation transmittance by cloud type under fully overcast conditions, based on observations made in Hamburg, Germany, from 1964 to 1973. For cloud types ranging from altocumulus and altostratus to stratus, transmittance ranged from 0.27 (a 73% reduction) to reductions of up to 82% depending on cloud type—figures that underpin the company's claims.
However, this transmittance data reflects specific locations, specific cloud types, and fully overcast conditions. It doesn't directly indicate annual output gains. The percentage of blocked sunlight and the extent to which drones can alter that cloud—and for how long—are separate questions.
Cloud brightness isn't determined by liquid water content alone. The number and effective radius of cloud droplets also matter; generally, more numerous smaller droplets create more reflective surface area, making clouds brighter. While enlarging droplets to promote collision-coalescence and precipitation could reduce reflectivity, the outcome depends on the cloud's thermodynamic state, updrafts, and surrounding moisture supply. Meteoric hasn't disclosed its specific method for altering cloud droplets, and the term "mechanical modification" doesn't allow us to infer whether the approach involves mixing, ionization, heating, or something else.
Natural clouds differ from a sealed test chamber. They exist in open systems where wind and turbulence continuously replenish water vapor and droplets, meaning the 13% figure from the chamber test can't simply be scaled linearly to treatment area or duration. The World Meteorological Organization (WMO) states that fog can in principle be dispersed through sufficient heating or mechanical mixing, but notes this is often impractical and expensive. It also states that cloud-scale ionization methods remain scientifically unproven.
The scale of stratocumulus clouds offers another gauge of how difficult verification will be. According to a 2012 paper by Wood, regions dominated by marine stratocumulus have an average liquid water path of 40–150 g/m², with typical cloud thickness of 200–500m. A simple conversion suggests a 1 km² column of cloud contains 40–150 tons of liquid water. This isn't a measured value for the clouds Meteoric is targeting, nor does it indicate the energy required to evaporate the entire volume. Still, it illustrates that validating results with natural clouds operates on a different scale than a test chamber.
What the $30–60/Hour Figure Doesn't Cover
Meteoric compares its approach favorably against traditional weather modification aircraft, which cost over $2,000/hour, claiming electric drones cost $30–60/hour—roughly a 97% reduction in the cost of reaching clouds. This $30–60 figure isn't a service price. It's the company's comparison of hourly costs for electric drones, and doesn't represent the total cost of operating a fleet of dozens to hundreds of drones.
Whether this comparison includes required flight hours, charging, and personnel costs hasn't been disclosed. Nor have insurance or permitting costs. To compare this against a power plant's annual revenue increase, one would need the treated area, duration, and number of drones deployed—calculating total operational costs over the same period, rather than the hourly cost of a single aircraft. The modeled value of $5,000–$28,000 per MW and the $30–60 per hour figure cannot simply be subtracted from one another.
Aviation operations are also central to this concept. Under the U.S. Federal Aviation Administration's (FAA) Part 107, standard small unmanned aircraft operations are generally limited to altitudes of 400 feet (about 122m) above ground level and require visual line of sight, with one operator or visual observer able to oversee only one aircraft at a time. Waivers are required for operations exceeding this altitude or beyond visual line of sight. Waivers are also needed if a single person is to control multiple drones simultaneously, or if drones fly within 500 feet vertically or 2,000 feet horizontally of clouds.
The 1–5 km altitudes Meteoric describes translate to roughly 3,281–16,404 feet—far exceeding the standard altitude ceiling. If dozens to hundreds of drones are to be flown into clouds within the U.S. under Part 107, multiple waivers would be required depending on the oversight structure and visual observation method, with airspace authorization potentially needed as well depending on the location. Not using chemicals doesn't eliminate aviation safety requirements.
Four Metrics That Should Define the 2027 Trials
The WMO calls for randomized comparisons between treated and untreated cases when evaluating weather modification. Effect estimates should include confidence intervals and be supported by supplementary analysis grounded in physical processes. If evaluation relies on modeling, uncertainty must be quantified and constrained by observation. For Meteoric's 2027 flights to be evaluated as viable technology for solar farms, results will need to meet this standard.
First is treatment area: without knowing which cloud types were treated, over what area, and by how many drones, it's impossible to compare the 13% chamber result with outdoor outcomes. Second is effect duration: in the open system of natural clouds, how quickly wind and turbulence restore the cloud to its original state will determine how much output can actually accumulate.
Third is net output gain per drone deployed: beyond changes in sunlight, the increase in generation must be demonstrated relative to an untreated control area. Fourth is uncertainty: because clouds vary even in the same location depending on time and surrounding conditions, a single successful case cannot substitute for actual measurements replacing the modeled 10–30% annual estimate.
The late-2028 storm flights represent an even more distant goal. The WMO states there is no generally accepted evidence that tropical cyclones can be modified, and that technologies claiming large-scale effects capable of eliminating severe weather events lack sound scientific grounding. At this stage, there is no basis for claiming hurricanes can be weakened.
If clouds above solar farms can indeed be thinned, the amount of power extractable from existing infrastructure would increase. But the number that would turn this possibility into a viable business isn't the 30% figure. What the 2027 flights must demonstrate is whether they can show controlled output gains and effect duration, and calculate everything from required drone counts to total operational costs.