A startup backed by Silicon Valley accelerator Y Combinator wants to send swarms of drones into clouds above solar farms to thin them out, so more sunlight reaches the panels below. Meteoric Technologies, started by 2 engineers from the University of Cambridge, says its drones would change the water droplets that form low- and mid-level clouds, without using chemicals.
"This reduces their reflectivity and brings back measurable amounts of sunlight to the panels below," said chief executive Mete Karslioglu.
Insights on Meteoric's cloud-clearing drones
The plan is to fly fleets of dozens or even hundreds of autonomous drones into clouds between about 1km and 5km above a solar farm. Karslioglu says low and mid-level overcast clouds cut incoming sunlight by 73% to 82%, and that the drones could partly reverse that loss.
"Our cloud-loss model shows that this process can increase annual generation from existing assets by 10-30% across major US grid regions, worth around $5,000-28,000/MW, without building any new infrastructure," he said, about R82,000 to R460,000 per megawatt.
Meteoric also compares costs: it says conventional cloud-clearing flights by aircraft cost $2,000 (about R33,000) an hour or more, against $30 to $60 (about R490 to R990) an hour for its electric drones.
The evidence so far is thin. Meteoric has not explained exactly how its drones disperse clouds. It says a working prototype thinned an artificial cloud by 13% in a cloud chamber, but has published no details about the chamber, how it measured the result, repeated tests or independent review. The 10% to 30% figure comes from a computer model, not from a working solar farm.
Its longer-term goal is even bolder: "Our ultimate goal is to reduce the intensity of severe storms and hurricanes," Karslioglu said. The next step is to show it works on a real cloud.
What others are saying about Meteoric's cloud-clearing drones
TechSpot reported Meteoric's plans and pointed out the gaps in its published evidence. Existing ways of clearing fog, used mainly at airports, work only in specific weather conditions, according to the American Meteorological Society, and the World Meteorological Organization notes that heating the air or mixing it mechanically is often too expensive to be practical.
The WMO also says there is no generally accepted evidence that hurricanes can be modified, and warns that technologies promising large or dramatic effects on weather lack a sound scientific basis.
A bold idea that still needs a real cloud
The appeal is easy to see. Solar farms lose a lot of output to cloud, and anything that adds 10% to existing panels without building new ones would be valuable, especially as the world races for more power, with some AI data centres now burning gas in old jet engines while they wait for grid connections.
But the claims are far ahead of the evidence: a lab test on an artificial cloud and a computer model, with no published method.
For South Africa, the use case is also weaker than it sounds. Much of the country's large-scale solar is built in the Northern Cape, one of the sunniest places on Earth, where cloud is less of a problem than in Europe or parts of the US.
Our bigger obstacle is getting power onto the grid, as we found with South Africa's onshore wind. It is worth watching, but as an experiment, not a solution.
You might also like our piece on a robot building in space, AI data centres in Africa, and Eskom's falling sales.
Get more SA tech and business news and subscribe to The Open Letter.


