Signals Inbox·August 22, 2026·Solar

What is Meteoric (YC S26) actually doing?

Meteoric is building drones that are supposed to make clouds over solar farms less reflective, so more sunlight reaches existing panels and more electricity comes out. The solar pitch is surprisingly coherent; the much louder hurricane ambition is still several scientific steps away.

Q1What is Meteoric actually building today?

Meteoric is currently building drones that are supposed to make clouds over solar farms less reflective, letting more sunlight reach the panels and increasing electricity production.

That is the clearest description of the company today. Meteoric’s own website now walks through five steps: deploy drones near a solar plant, fly them into low- and mid-altitude clouds, alter the droplets, reduce the cloud’s reflectivity, and let more sunlight through. It also promises no chemicals, no new infrastructure at the solar site, and a measurable increase in output.

The wording has become much more specific lately. A version of Meteoric’s website crawled only recently still led with the broad slogan “Engineering Reliable Weather” and showed a weather-balloon field test from May 2025. The current site leads directly with “Drones that clear clouds over solar farms.” Y Combinator now describes Meteoric the same way, while its latest launch material adds the longer-term ambition of weakening hurricanes.

Meteoric is a two-person S26 company founded by Cambridge engineers Mete Karslioglu and Eric Nilsson. YC currently lists it as active in San Francisco.

So if someone discovers Meteoric through the hurricane headline, they are starting several steps ahead of where the company actually is. The product being pitched to customers today is about getting more electricity from solar farms during cloudy conditions.

Q2Is Meteoric really a solar startup or a weather-control company?

Meteoric is using solar farms as the first place to prove that it can deliberately change clouds in a useful and measurable way.

Solar gives Meteoric an unusually clean experiment. The target stays in one place. The desired result is obvious: more sunlight reaches the ground. Solar plants already measure their production continuously, so a successful intervention should eventually show up in irradiance data and electricity output.

That is much easier to test than a broad claim such as “we can control weather.”

There is also a reason the founders chose this particular application rather than agriculture or rainfall. A solar farm turns extra sunlight directly into money. If Meteoric changes a cloud and a 200 MW plant suddenly produces more electricity than an untreated comparison site under similar conditions, there is a number attached to the result.

The underlying technology could eventually have uses far beyond solar. For now, though, solar is doing more than providing a convenient first market. It gives Meteoric a way to find out whether its basic idea works at all.

Q3How are Meteoric’s drones supposed to clear clouds without chemicals?

Meteoric says its drones alter cloud droplets without releasing chemicals, but it has not publicly explained the physical mechanism in enough detail for us to know exactly how the drones do it.

The public explanation currently stops at “alter the cloud droplets.” Meteoric does not say what the aircraft emit or apply, which droplet properties they change, how much energy the process uses, or how those changes make a cloud optically thinner.

There are scientifically plausible ways to affect droplets without conventional cloud-seeding chemicals. Electric charge, for example, can change how droplets collide, stick together and evaporate.

A useful precedent comes from researchers at the University of Reading and the University of Bath. They equipped a small unmanned aircraft with electrical charge emitters and flew it through natural fog. The experiment produced detectable changes in droplet behavior and altered the fog’s reflectivity by as much as about 2%.

There is an important catch: that experiment increased reflectivity, whereas Meteoric wants to reduce it. It shows that a drone can influence natural droplets without spraying conventional seeding chemicals. It does not reveal Meteoric’s method or demonstrate Meteoric’s desired effect.

Until the company publishes more, its core mechanism remains one of the biggest unanswered questions.

Q4Has Meteoric actually cleared a real cloud yet?

Meteoric has not publicly shown a controlled field result proving that its drones can clear a natural cloud above a solar farm and increase the plant’s electricity production.

That is where the company stands today.

We know the founders have been working with atmospheric hardware for longer than the YC launch might suggest. Meteoric previously showed a weather-balloon field test dating to May 2025, and both founders say they have built atmospheric measurement equipment before starting the company.

What is missing publicly is a very different type of result: drones entering a target cloud, a measurable change appearing in that cloud, more sunlight reaching the ground, and the solar plant producing additional electricity that can reasonably be attributed to the intervention.

Meteoric’s current 10% to 30% regional uplift figures are explicitly described on its website as maximum estimates based on modeled cloud types. They are modeling outputs, not published field results.

That distinction should stay front and center. Meteoric is currently trying to prove cloud clearing at useful scale. Public evidence that it has already done so is still absent.

Q5Is the science behind Meteoric actually real?

The physics Meteoric is building on is real, while the jump from changing individual cloud droplets to economically clearing cloud cover remains wide open.

Cloud microphysics is sensitive to droplet size, concentration, electrical charge, temperature and collisions. Researchers have repeatedly shown that changing these properties can affect precipitation and the way clouds interact with radiation.

The UAV experiment mentioned above is especially relevant because it moved the idea out of a laboratory. A real aircraft released charge into natural fog, and instruments detected a change in its radiative properties roughly 25 seconds later.

But the measured effect was small: maximum reflectivity changed by about 2%. The aircraft was circling at roughly 20 meters altitude in fog. Meteoric wants to operate in low- and mid-level clouds roughly 1 to 5 kilometers above solar plants and recover enough sunlight to change annual power production by double-digit percentages.

Those are very different scales.

We can be confident that droplets can be manipulated. We cannot yet be confident that Meteoric can manipulate enough of them, over enough area, for long enough, to create a useful solar product.

Q6Do clouds really hurt solar farms as much as Meteoric says?

Yes, heavy cloud cover can wipe out most of the sunlight that would otherwise reach a solar farm while the cloud is overhead.

Meteoric and Y Combinator use a figure of roughly 75% for low- and mid-level overcast clouds. That is broadly consistent with atmospheric measurements.

Studies of solar transmission through different cloud types have found very low transmission under dense overcast conditions. Nimbostratus and stratocumulus can leave only around 15% to 25% of comparable clear-sky solar radiation reaching the surface, while some middle-level overcast clouds also remove well over half.

The 75% figure becomes misleading only if we treat it as an annual production loss. Solar plants do not spend every daylight hour under thick overcast clouds.

Meteoric’s real opportunity is the fraction of annual solar generation lost during cloud conditions that its drones can actually treat. That number depends heavily on location, cloud type, season, wind and the amount of time the intervention lasts.

This also explains why Meteoric’s own modeled annual uplift varies so widely across U.S. power markets.

Q7Can Meteoric really increase solar output by 10% to 30%?

Meteoric’s 10% to 30% annual output figures are plausible as modeled upper bounds, but today there is no public field evidence showing that a solar farm can actually achieve them.

Meteoric publishes different estimates by electricity market. Cloudier regions come out much higher: 30% in NYISO, 22% in PJM and MISO, and 20% in ISO-NE. Sunnier CAISO is 12%, while the non-ISO Northwest is 10%.

That ordering makes sense. If cloud cover is the resource Meteoric is trying to remove, markets with more recoverable cloud losses should offer more upside.

The difficult part comes after the weather model. Meteoric still has to reach the relevant cloud, alter enough of it before it moves, keep the effect alive long enough to matter, and do that repeatedly through the year.

Some cloudy hours will also be useless. The wrong cloud type may be present. Winds may be too strong. The cloud may be too thick, too high or moving too quickly. The solar plant may already be curtailed. A cloud could also disappear naturally shortly after the drones arrive.

So we should read Meteoric’s figures as estimates of available upside under favorable assumptions. Realized output uplift is the number that still needs to be discovered.

| Power market | Meteoric’s maximum modeled uplift | | ------------------ | --------------------------------: | | NYISO | 30% | | PJM | 22% | | MISO | 22% | | ISO-NE | 20% | | SPP | 15% | | ERCOT | 15% | | CAISO | 12% | | Northwest, non-ISO | 10% |

Q8Does Meteoric’s $5,000 to $28,000 per MW claim actually add up?

Yes. Meteoric’s claimed annual value per solar megawatt survives a basic independent check surprisingly well.

The U.S. Energy Information Administration currently puts the 2025 capacity factor for utility-scale photovoltaic generation at 24.4%. One MW operating at that capacity factor produces roughly 2,137 MWh in a year.

A 10% production increase adds about 214 MWh. A 30% increase adds about 641 MWh.

Berkeley Lab’s latest utility-scale solar dataset found that the middle 50% of PPAs for projects entering operation in 2024 were roughly $22 to $40 per MWh, with an average around $29. At those prices, 214 extra MWh is worth roughly $4,700 to $8,500. An extra 641 MWh is worth around $14,100 to $25,600.

The highest-priced regions push the calculation further. Berkeley Lab reported recent average PPA prices around $44.50/MWh in PJM, $51/MWh in ISO-NE and $59/MWh in NYISO, which are also among the regions where Meteoric models its highest cloud-related gains.

The economics behind the headline are therefore believable. Meteoric’s real risk sits upstream: producing those extra MWh in the first place.

| Per 1 MW of solar | 10% uplift | 30% uplift | | --------------------------------------------------- | ----------------------: | ----------------------: | | Baseline annual generation at 24.4% capacity factor | 2,137 MWh | 2,137 MWh | | Additional generation | ~214 MWh | ~641 MWh | | Value at $22/MWh | ~$4,700 | ~$14,100 | | Value at $40/MWh | ~$8,500 | ~$25,600 | | Meteoric’s stated range | ~$5,000–$28,000/MW/year | ~$5,000–$28,000/MW/year |

Q9Why would a solar farm pay Meteoric instead of just adding more panels?

Meteoric becomes much more attractive when we look at how hard it currently is to put additional generating capacity onto the U.S. grid.

Berkeley Lab’s latest interconnection update shows 773 GW of solar still sitting in U.S. grid queues. That number actually fell 19% during 2025, so the backlog has improved slightly lately, but it remains enormous.

Projects that reached operation in 2025 had typically spent more than five years between their original interconnection request and commercial operation in regions where Berkeley Lab had complete data.

The long-term success rate is even harsher. Among projects requesting interconnection between 2000 and 2020, only 13% of proposed capacity had reached commercial operation by the end of 2025. Three quarters had been withdrawn.

A solar farm that already exists has something increasingly valuable: an operating grid connection.

Meteoric wants to increase production from that existing asset without adding another field of panels or waiting years for another project to connect. Batteries can shift existing electricity into more valuable hours, but they do not recover sunlight lost behind clouds.

If Meteoric can add meaningful generation while staying inside a plant’s existing electrical and contractual limits, that is a compelling pitch. The appeal comes from squeezing more out of infrastructure that has already cleared the hardest development bottlenecks.

Q10How large could Meteoric’s solar market become?

Meteoric could eventually address billions of dollars of annual solar-production value in the U.S. alone, although we still have no defensible way to know what fraction is actually reachable.

American Clean Power’s latest utility-scale solar update puts the operating U.S. fleet at about 161 GW. The same organization expects roughly 300 GW to be operating by 2030.

If we mechanically apply Meteoric’s $5,000 to $28,000 annual value claim across today’s entire 161 GW fleet, we get roughly $800 million to $4.5 billion of additional electricity value per year.

At 300 GW, that theoretical range reaches about $1.5 billion to $8.4 billion.

Those figures are useful for scale, but they would be terrible estimates of Meteoric revenue. Plenty of plants will have little recoverable cloud loss. Some clouds may prove untreatable. Curtailment and PPA structures can reduce the value of extra midday generation. Regulations will remove some locations entirely.

Even after heavy discounts, however, Meteoric does not need the hurricane business to justify working on the solar problem. A technology that economically improves generation across only a modest slice of a 161 GW and rapidly growing fleet could already support a serious company.

Q11Could the drone-fleet economics kill Meteoric’s business?

Yes. Meteoric could prove that cloud modification works scientifically and still fail if each extra megawatt-hour requires too much aircraft time.

This is currently one of the largest blanks in the public story.

Meteoric does not disclose how many drones a typical solar plant would need, how much cloud area one aircraft can influence, how long an intervention lasts, how often drones must return, or what the fleet costs to buy and maintain. It also has not published a customer price.

Those variables are essential.

Take a 100 MW solar farm. Using Meteoric’s own range, the theoretical annual value created could be around $500,000 to $2.8 million. That sounds generous until we remember that cloud treatment might require repeated missions throughout the year, autonomous flight hardware, batteries, maintenance, weather forecasting, communications, insurance and regulatory compliance.

The real unit metric we want is simple: how many dollars does Meteoric spend to create one additional saleable MWh?

Today, we cannot calculate it. If the company eventually publishes a successful field test without publishing the fleet effort required to produce that result, half of the business case will still be missing.

Q12Can Meteoric legally operate cloud-clearing drones at scale?

Meteoric faces a serious regulatory problem in the U.S. because its intended flights collide with ordinary drone rules and, in some states, with weather-modification laws themselves.

The altitude alone makes standard Part 107 operations unsuitable. Meteoric says it wants to reach clouds around 1 to 5 kilometers high, while normal small-drone operations are generally capped at 400 feet above ground unless specific exceptions or approvals apply.

Ordinary Part 107 rules also require visual line of sight and substantial clearance from clouds. Meteoric’s whole idea involves deliberately entering those clouds.

The good news for the company is that these rules can be waived. FAA certificates already exist allowing combinations of beyond-visual-line-of-sight flight, operations above 400 feet, reduced cloud clearance and multiple drones under one pilot. Meteoric would need a convincing safety case rather than inventing a completely new legal category.

Weather-modification law is trickier.

Federal rules require weather-modification activities to be reported to NOAA, generally before operations begin. Several states have moved further. Tennessee prohibits intentional atmospheric activities intended to affect weather or sunlight. Florida now prohibits releasing chemicals, substances or even an “apparatus” into the atmosphere for the purpose of affecting temperature, weather, climate or the intensity of sunlight, with criminal penalties attached.

That wording reaches unusually close to Meteoric’s stated product.

So “no chemicals” helps with one environmental objection, but it does not give Meteoric a regulatory free pass. Where the company can operate may become part of the product strategy.

Q13Has anyone already used drones to modify weather?

Yes. Weather-modification drones already exist, including commercial systems operating under current U.S. weather-modification programs.

Rainmaker gives us a useful comparison. A recent NOAA filing shows the company using unmanned drones for cloud seeding over Idaho’s Bear River Basin. Each drone can release aerosolized dry silver iodide into target clouds for up to 30 minutes.

Rainmaker is trying to encourage precipitation, while Meteoric wants more sunlight and says it will use no chemicals. The physical methods are different.

Still, Rainmaker proves several things that are directly relevant: specialized drones can be sent into weather systems, regulators can approve unmanned cloud-modification operations, and there are paying institutions interested in manipulating clouds for economic reasons.

Add the University of Reading experiments, where researchers changed natural fog properties from an unmanned aircraft, and Meteoric starts to look less isolated technologically.

The unusual part is the combination. We have precedent for drones entering clouds, precedent for deliberate weather modification and precedent for nonchemical droplet manipulation. We have yet to see those pieces combined into a solar-generation service.

Q14Does Meteoric’s founding team actually fit this problem?

Meteoric’s two founders have unusually relevant backgrounds for a company trying to manipulate cloud droplets.

Mete Karslioglu studied aerospace, energy and mechanical engineering at Cambridge and previously worked at the Cambridge Centre for Climate Repair on seawater spray nozzles designed to increase cloud reflectivity.

That previous research is almost the reverse optical problem from Meteoric’s solar pitch. Marine cloud brightening tries to make clouds reflect more sunlight. Meteoric now wants certain clouds to let more sunlight through. Both require understanding how small changes to cloud droplets change radiation.

Eric Nilsson specialized in electrical, control and computer engineering at Cambridge. Meteoric says he previously built systems that remove droplets from LiDAR lenses and researched electronic chips for controlling tiny volumes of liquid. Both founders also have experience building atmospheric measurement equipment.

That does not prove they can clear clouds, but the connection between their earlier work and Meteoric is unusually direct.

The obvious limitation today is team size. YC still lists only two people. Building the eventual system will require far more than drone engineering: cloud physics, meteorology, autonomy, aviation safety, field operations and probably regulatory specialists.

For an experiment-stage company, the founders fit the problem well. The organization required to run the product does not exist yet.

Q15What would Meteoric need to prove in a serious field test?

Meteoric needs to show that its drones cause a repeatable increase in solar output compared with what the same cloud would probably have done without intervention.

A dramatic video of blue sky appearing over panels would be weak evidence. Clouds change shape and disappear naturally all the time.

A convincing test would track treated and untreated cloud areas under similar conditions. Ground instruments should measure solar irradiance below both areas. Satellite, radar or lidar data should follow the cloud itself. The plant’s meter should then show whether the extra sunlight turned into extra electricity.

Duration matters just as much as peak effect. A large irradiance jump lasting 20 seconds might be scientifically interesting and commercially useless.

Meteoric would also need repeated experiments across different days and cloud types. One successful event would leave too much room for luck.

Then comes cost. We would want to know how many drones flew, for how long, how much electricity value was created and how much the mission cost.

The key number is eventually going to be very boring compared with the company’s headline: additional MWh per dollar of cloud treatment. That metric will tell us whether Meteoric has discovered an interesting atmospheric effect or built a useful energy product.

Q16Can clearing clouds over solar farms really lead to weakening hurricanes?

Solar cloud clearing could teach Meteoric a lot about atmospheric intervention, but hurricane weakening remains several major scientific leaps beyond anything the company has publicly demonstrated.

The connection is real at a basic level. Meteoric would learn how droplets respond to intervention, how autonomous aircraft behave inside clouds, how to target particular atmospheric structures and how to measure whether a change was actually caused by the treatment.

A hurricane adds an entirely different level of complexity. It is a huge rotating heat engine continuously drawing energy from warm ocean water. Changing the properties of droplets somewhere inside the system only helps if that local change eventually alters convection and circulation enough to reduce damaging winds.

We already have a powerful warning from Project STORMFURY, the U.S. government’s decades-long attempt to weaken hurricanes with silver iodide.

The early experiments looked surprisingly good. Seeding was attempted in four hurricanes on eight days, and maximum winds fell by 10% to 30% on four of those days.

Later research changed the interpretation. Scientists discovered that unmodified hurricanes naturally undergo eyewall replacement cycles that can produce very similar temporary weakening. They also found less of the supercooled water needed for the original seeding hypothesis than expected.

In other words, researchers observed the result they wanted before realizing nature could create the same result on its own.

Meteoric may eventually use a completely different mechanism, so STORMFURY cannot tell us whether its future approach will work. It does tell us how difficult the proof will be.

As we saw previously with solar clouds, causality is already hard to establish in ordinary weather. Inside a hurricane, that problem becomes much harder.

Today, hurricane weakening belongs in Meteoric’s research ambition rather than its product description.

Q17So what is Meteoric actually doing?

Meteoric is currently trying to turn cloud modification into a measurable way of getting more electricity from existing solar farms.

The basic idea is surprisingly coherent. Thick clouds really can remove most incoming solar radiation while they are overhead. The U.S. already has about 161 GW of utility-scale solar operating, while 773 GW of proposed solar remains in interconnection queues. Meteoric’s claimed $5,000 to $28,000 of annual value per MW also holds up reasonably well when we rebuild the calculation using current solar capacity factors and recent PPA prices.

The science underneath the idea has some real foundations too. Researchers have altered natural fog droplets using unmanned aircraft, and drones are already being used for commercial cloud seeding in the U.S.

Meteoric’s missing piece is also very clear now: there is no public evidence yet that its own system can take a real cloud above a solar farm, change it enough to increase generation, repeat the effect reliably and do it for less than the extra electricity is worth.

That is the company’s real bet.

The latest YC launch pushes the story all the way to weakening hurricanes, but that can distract from a much nearer and more interesting test. Meteoric first has to show that a solar operator can look at its production meter after a drone mission and see electricity that would otherwise have been lost behind clouds.

If it can do that repeatedly, Meteoric will have built a genuinely unusual solar technology and proved that at least one part of local weather can be treated as something an operator can actively change.

For now, that proof is still ahead.