When Startups Want to Change the Weather

I titled this landscape “Climate Intervention Startups” rather than geoengineering startups because geoengineering usually refers to large scale attempts to modify the Earth’s climate. But what I noticed, while searching this space, is that several startups are intervening in individual weather events, while others are trying to alter physical conditions within precise regions. What connects them is the same underlying idea: instead of only adapting to environmental conditions, they are developing technology to actively change them.

As a consequence I organized the landscape by the scale of the intervention:

• Local intervention targets a specific weather event, such as increasing rainfall.

• Regional intervention tries to change conditions within a defined ecosystem, such as preserving Arctic sea ice. 

• Global intervention aims to influence planetary temperature itself. The distinction matters because moving up this scale radically changes the technical ambition and, even more importantly, the political constraints around deployment.

Below you’ll find an analysis of each category.

Local Weather Event Intervention

What is this category about?

• This category covers technologies that deliberately change the outcome of a local weather event rather than simply predict it. The most established use case is increasing precipitation when suitable clouds are already present. The intervention remains geographically bounded and depends on existing atmospheric conditions, which makes it very different from attempts to influence the climate at planetary scale.

• The most obvious problem being addressed is water scarcity in regions where additional precipitation has significant economic value. A cloud can contain moisture without efficiently producing rain or snow over the area that needs it. Weather modification tries to increase the amount of that existing moisture that reaches the ground rather than create weather from nothing.

• A big challenge is to prove that the intervention actually worked. Weather is naturally variable, so observing more precipitation after an intervention does not establish that the intervention caused it. This makes weather modification unusually difficult to evaluate and has historically limited trust in the industry. Newer systems increasingly treat measurement as part of the product rather than something done separately after the operation.

• Traditional weather modification is also operationally heavy. Crewed aircraft need to be available at exactly the right moment and operators need to decide where an intervention has the best chance of working. New companies are trying to make these operations more repeatable by improving targeting or replacing parts of the aircraft based workflow with more autonomous systems.

What do products in this category do?

• The most established product model is to deliver a seeding material directly into clouds. The underlying idea is that some clouds already contain the conditions needed for precipitation but do not convert enough of their moisture into rain or snow. Introducing particles that encourage ice formation can improve that process under the right conditions. Rainmaker is modernizing this approach with drones that enter the target cloud, while Cloud Seeding Technologies sells delivery systems that can be deployed from aircraft or other platforms.

• A second approach tries to intervene from the ground instead of entering the cloud directly. Removing the aircraft from the operation could make deployment easier across large areas and allow systems to remain installed for longer periods. Rain Enhancement Technologies takes the most distinct approach here with WETA, which uses ground based ionization arrays designed to influence precipitation formation in existing weather systems. Cloud Seeding Technologies also sells ground based generators, although those still disperse conventional seeding material.

• The real product innovation is about turning weather modification into a closed loop system. The intervention itself has existed for decades, but operators historically had limited ability to determine exactly when to intervene or quantify what happened afterwards. Rainmaker combines weather modeling with radar based validation to identify suitable clouds and measure the result. Recast is pushing further in the same direction by building its system around better sensing and identifiable seeding material, so each operation can produce data that improves future targeting.

• Some companies extend the same infrastructure beyond rainfall enhancement. Once a company can identify suitable atmospheric conditions and intervene at the right moment, similar infrastructure can potentially be applied to other localized weather problems. Project Rainmaker positions cloud seeding around hail reduction in addition to rainfall, while Rain Enhancement Technologies has discussed fog mitigation. These use cases remain less established than precipitation enhancement, so the product opportunity is broader than the scientific evidence currently supporting every application.

How has cloud seeding technology evolved?

• The technology started in the 1940s with a relatively simple discovery about cloud microphysics. Researchers found that supercooled water inside clouds could be encouraged to form ice crystals by introducing substances such as dry ice or silver iodide. Once those crystals grew large enough, they could fall as precipitation. This established the basic mechanism that still underpins most conventional cloud seeding today.

• From the 1950s through the 1970s, cloud seeding moved from laboratory experiments into large government programs. Aircraft became the main delivery mechanism and silver iodide emerged as the dominant material for cold clouds. Governments funded increasingly ambitious field experiments because the technology appeared to offer a way to increase water supply. The physical intervention became operationally feasible, but researchers struggled to determine how much additional precipitation had actually been caused by seeding rather than natural weather variation. U.S. federal spending peaked during this period before falling sharply.

• The 1980s and 1990s were largely a period of retrenchment rather than technological breakthrough. Many experiments produced inconclusive results, which weakened scientific confidence and caused U.S. federal funding to decline substantially. Cloud seeding nevertheless continued operationally in several regions. The central technical limitation was becoming clear: dispersing seeding material was relatively easy, but knowing whether the right cloud had been seeded and measuring the resulting effect was much harder.

• From the late 1990s onward, most of the important progress happened around the intervention rather than inside it. Better radar and numerical weather models made it easier to understand the structure of a cloud before deciding whether to seed it. Improved sensors then allowed researchers to observe what happened during an operation with much greater precision. Reviews of the field describe this period as a shift toward better decision support and evaluation rather than a fundamental reinvention of seeding itself.

• The current generation is trying to make cloud seeding much more precise and measurable. Drones can enter clouds without putting pilots at risk while collecting atmospheric measurements during the mission. More sophisticated software can continuously adjust where an intervention occurs based on observed conditions. Research is also improving seeding materials, but the biggest change is that deployment is becoming connected to real time observation. Cloud seeding today therefore looks increasingly like a controlled atmospheric operation rather than simply an aircraft releasing silver iodide into a promising cloud. The science is still conditional, however, and even modern systems cannot make unsuitable clouds produce meaningful precipitation.

• Cloud seeding is an old physical technology entering a much more interesting engineering phase. The core mechanism has changed surprisingly little in eighty years. What is changing quickly is our ability to decide when to intervene and establish what happened afterwards. That matters because the historical weakness of cloud seeding was never simply the delivery mechanism. It was the inability to close the loop between intervention and outcome.

What types of customers are using these products?

• Public water agencies are the clearest customer because additional precipitation can directly increase available water supply. Idaho Water Resources Board materials identify Rainmaker as the operator for the Bear River Basin interstate cloud seeding pilot. Rainmaker has also worked with the West Texas Weather Modification Association. Recast says it has completed more than 100 cloud seeding flights for government partners across Texas and New Mexico, although it does not publicly name those customers.

• Research institutions are important early users because commercial adoption still depends heavily on proving that the technology works. ETH Zurich's CLOUDLAB project uses seeding flares from Cloud Seeding Technologies on an unmanned aircraft as part of its experiments. For suppliers in this market, research deployments can therefore serve as product validation before the technology reaches larger operational programs.

• Agriculture is a natural commercial market wherever precipitation changes have a direct impact on crop economics. Project Rainmaker positions its service toward Australian agriculture, while Rain Enhancement Technologies explicitly targets agricultural water users. Public evidence of scaled private sector adoption remains limited, however, so government backed water programs currently appear further along than direct enterprise demand.

What are the major trends shaping this category?

• Demand for weather modification could increase as water stress becomes harder for governments to ignore. The WMO says operational weather modification programs now exist in more than 50 countries and notes that demand continues to rise in response to drought. In the United States alone, GAO found that public agencies across nine states were spending or planning to spend at least $20 million on cloud seeding in 2024. This is still a small market, but weather modification is increasingly moving from occasional experimentation into the water management toolkit.

• The scientific bar is rising at the same time as commercial interest. WMO considers the evidence strongest for certain forms of cloud seeding under well defined atmospheric conditions, while approaches such as ionization still lack comparable proof. The implication is that the industry cannot simply compete on who claims the largest increase in rainfall. Companies increasingly need to demonstrate under which conditions their intervention works and how confidently the effect can be separated from natural variability.

• Drones could materially change the economics of cloud seeding. A 2026 review of precipitation enhancement technology found that drones can reduce operational risk while collecting atmospheric measurements directly inside clouds. More advanced systems can also adapt their flight path using live conditions. The remaining constraints are meaningful, particularly around flight endurance and aviation regulation, but the direction is toward smaller autonomous platforms rather than relying exclusively on conventional seeding aircraft.

• Regulation is likely to become more visible as the industry grows. Weather modification operators in the United States already have federal reporting obligations, but GAO found in 2026 that more than half of the reports submitted to NOAA likely contained errors or missing information. GAO subsequently called for stronger oversight. Better reporting may create additional compliance work, but it should also favor operators that can document exactly what they deployed and where the intervention occurred.

How does the funding environment look?

• Most of the venture capital in this small landscape has concentrated around Rainmaker. The company raised a $6.3 million seed round in 2024 and followed it with a $25 million Series A led by Lowercarbon Capital in 2025. That progression shows that investors are willing to finance weather modification when it is presented as a scalable technology platform rather than a traditional cloud seeding service.

• The rest of the category is much less heavily venture funded. Recast has reported seed backing from Climate Capital and Asylum Ventures, but the amount has not been disclosed. Cloud Seeding Technologies and Project Rainmaker have no meaningful institutional funding recorded in the landscape. This is not yet a category where multiple venture backed companies are raising large rounds in parallel.

• Overall, I would describe the category as “selectively VC compatible” but still slow in terms of VC activity. The difficulty is that every company needs substantial real world validation before growth can become repeatable.

Regional Climate Intervention

What is this category about?

• This category covers technologies that deliberately alter physical conditions in a defined region in order to reduce a climate related risk. The intervention is larger than modifying an individual weather event, but it is still bounded to a particular geography. The objective is not to change the global climate directly. It is to protect a region where warming creates an especially damaging feedback or where extreme conditions have unusually high consequences.

• One problem is that some regional climate systems can amplify warming once they begin to change. Arctic sea ice is the clearest example. When reflective ice disappears, darker ocean water absorbs more solar energy and accelerates local warming. An intervention that keeps ice present for longer could therefore have an effect beyond the additional ice itself.

• Another problem is that some climate impacts are driven by physical conditions that can theoretically be altered locally. Tropical cyclones draw energy from warm surface water, while coral bleaching is strongly linked to marine heat. This creates the possibility of intervening directly in the local environment rather than waiting for global emissions reductions to change conditions decades later.

• The biggest challenge is that almost none of these approaches have been validated at the scale where they would matter. A method can work across a small patch of sea ice or within a controlled experiment without proving that it can influence a regional climate system safely. The companies in this category are therefore still solving basic questions around feasibility before they can solve commercialization.

What do products in this category do?

• One approach tries to preserve sea ice by artificially increasing its thickness during winter. The product logic is relatively simple: thicker ice should survive longer into the summer, keeping the ocean reflective for more of the year. Real Ice pumps seawater from beneath existing ice onto the surface, where it freezes in the Arctic winter. Arctic Reflections uses the same physical principle but is working on pumping systems that could eventually cover larger areas. Both are currently testing whether the additional ice actually survives longer and whether the intervention creates ecological problems.

• Another type of startups tries to weaken extreme storms by changing the ocean conditions that feed them. Tropical cyclones require sufficiently warm surface water, which creates the idea of temporarily cooling the water in their path. OceanTherm uses submerged bubble curtains to move colder deep water toward the surface. A Norwegian field test showed that this mechanism could lower local surface temperature, but the company has not yet demonstrated that it can operate across an area large enough to influence a cyclone.

• Marine cloud brightening approaches the problem by changing how much sunlight reaches the ocean rather than changing the ocean directly. More reflective marine clouds could reduce incoming solar energy over a targeted area and temporarily lower local temperatures. Ocean Cooling Technology proposes spraying very small seawater droplets into suitable marine clouds so the resulting salt particles increase cloud reflectivity. Unlike sea ice thickening, the intervention would be temporary and would need to continue whenever cooling was required.

• Across these approaches, the real product is currently the experiment rather than the final climate intervention system. Companies are building pumping equipment or spray systems, but much of the work goes into measuring what happens after deployment. The immediate objective is to generate enough real world evidence to determine whether larger scale intervention should even be considered.

What types of customers are using these products?

• There are essentially no conventional commercial customers yet. Real Ice and Arctic Reflections are operating primarily through research programmes rather than selling deployments. ARIA is funding their work as part of the RASI programme, while academic institutions provide much of the scientific infrastructure around the experiments. This makes the current market closer to funded research than commercial climate infrastructure.

What are the major trends shaping this category?

• Climate intervention research is starting to move from models into deliberately small outdoor experiments. ARIA is funding five controlled outdoor experiments across its climate cooling programme, including sea ice thickening and marine cloud brightening. The experiments are explicitly limited in scale and must either be reversible or have effects that dissipate quickly. This is a meaningful change for a field that has historically existed much more in scientific papers than in real world deployment.

• The scientific question is shifting from whether an intervention is physically possible toward whether it could ever work at meaningful scale. ARIA's Arctic programme is testing patches below one square kilometre while simultaneously studying how the intervention affects ice movement and local ecology. The programme explicitly says current models are not accurate enough to support decisions about larger deployment. The experiments are therefore designed to generate the real world data those models are missing.

• Public institutions are becoming more willing to fund research while remaining deliberately cautious about deployment. ARIA describes its goal as building evidence that could equally support a decision not to use these technologies. UNEP takes an even more cautious position on solar radiation modification, arguing that approaches such as marine cloud brightening remain speculative and cannot substitute for emissions reductions. This creates an unusual market where funding for experimentation can grow even while political support for actual deployment remains unresolved.

How does the funding environment look?

• The largest amounts in this category are coming from public research funding rather than conventional venture rounds. ARIA allocated roughly £3.5 million to the Real Ice workstream and about £3.3 million to Arctic Reflections within its wider Arctic sea ice programme. 

• I would classify the category as “pre VC” than anything else. There is activity and meaningful funding, but most of it is research capital rather than venture capital. A genuine VC market probably only appears if one of these approaches moves from an experiment into something governments are seriously considering for deployment. Until then, these companies look more like technology vehicles for proving a scientific hypothesis than conventional climate startups.

Global Climate Intervention

What is this category about?

• This category covers technologies designed to lower global temperatures by changing how much solar energy the Earth absorbs. The main approach is solar radiation modification, where reflective particles are introduced high in the atmosphere so that a small portion of incoming sunlight is redirected away from the planet. Unlike carbon removal, the intervention acts on temperature without reducing the concentration of greenhouse gases already in the atmosphere.

• The appeal of global geo-engineering comes largely from speed (in theory). Cutting emissions takes decades to change atmospheric concentrations enough to materially affect temperatures, while removing carbon at planetary scale would require enormous infrastructure. Stratospheric aerosols could theoretically influence global temperature much faster, which is why the technology is sometimes discussed as an emergency measure if warming becomes substantially more damaging.

• Scale changes the nature of the category. A regional intervention can potentially be stopped if a local experiment produces unexpected effects. Global solar modification would influence atmospheric systems across national borders, which means technical deployment cannot realistically be separated from international governance.

What do products in this category do?

• The most direct approach uses sulfur dioxide because volcanic eruptions have already demonstrated that sulfate aerosols can temporarily cool the planet. Make Sunsets takes this logic almost literally. It launches balloons containing sulfur dioxide and sells Cooling Credits corresponding to the amount released. The product is already commercially available, but the amount deployed remains tiny compared with anything that could materially influence global temperature.

• A more engineered approach is to design particles specifically for solar reflection rather than relying on sulfur dioxide. The product logic is that a purpose built material could potentially reflect sunlight while reducing some of the drawbacks associated with sulfate aerosols. Stardust is developing solid particles for this purpose and published candidate designs in 2026. Its work is still at the R\&D stage rather than operational deployment.

What are the major controversies and political constraints?

• The fundamental political problem is that deployment would create global effects without a clear mechanism for global consent. Aerosols released by one country would not respect national borders, and different regions could experience different consequences. UNEP explicitly raises the possibility that unilateral deployment by a state or private actor could create geopolitical tensions. This makes solar geoengineering fundamentally different from most climate technologies

• There is also strong institutional resistance to treating the technology as a normal climate solution. UNEP stresses that SRM does not address greenhouse gas concentrations and could create unintended changes in climate patterns. European scientific advisers go further by recommending a moratorium on large scale deployment and arguing that emissions reduction should remain the priority. A company can therefore make substantial technical progress without getting any closer to being allowed to deploy commercially.

• Make Sunsets shows that these constraints are already becoming concrete rather than theoretical. Mexico banned the company's activities after its early launches, while the US EPA formally demanded information about its sulfur dioxide releases in April 2025. EPA now says it is reviewing whether existing authorities could be used to halt the activity if it scales significantly. The regulatory precedent matters far beyond Make Sunsets because it shows that governments may intervene long before atmospheric releases become large enough to affect the climate.

What are the major trends shaping this category?

• The technical conversation is moving from whether aerosols can cool the planet toward whether the process can be controlled safely. There is little doubt from volcanic evidence and climate modelling that stratospheric aerosols can reduce global temperature. The much harder questions concern atmospheric chemistry and how different regions would be affected. UNEP's latest assessment stresses that those uncertainties remain substantial.

• Governance is starting to develop before the technology itself is mature. European scientific advisers have recommended continued research while simultaneously calling for an EU wide moratorium on deployment. They also argue that any future decision would require a global governance framework. This means regulatory architecture may form years before a scalable commercial product exists.

How does the funding environment look?

• This is barely a venture category, but Stardust has created one major exception. Stardust has raised around $75 million across its $15 million seed round and $60 million financing in 2025, while Make Sunsets started with only $720,000 and later raised another undisclosed seed investment. The gap is revealing: investors are not broadly funding solar geoengineering startups, but a small group is willing to make a very large bet on a company that could become the infrastructure provider if governments ever adopt the technology. I would therefore describe the category as highly speculative rather than trending. Venture compatibility exists, but it depends almost entirely on believing that a future government procurement market will emerge.