Tiny Glass Balls in the Sky: The Science, Politics, and Peril of Silica Sphere Geoengineering

In a windowless conference room, investors are betting $75 million that microscopic glass spheres — each one-125th the width of a grain of sand — can buy humanity time against a warming planet [1]. The company behind the bet, Stardust Solutions, plans to begin releasing proprietary silica particles from high-altitude aircraft in 2026, lofting them roughly 18 kilometers into the stratosphere where they would scatter incoming sunlight back toward space [1]. If the approach works as modeled, it would represent the first commercial-scale test of a material other than sulfur dioxide for stratospheric aerosol injection (SAI), the most-studied form of solar geoengineering.

The proposal arrives at a charged moment. Academic publication on solar geoengineering aerosols has surged — over 3,195 papers since 2011, peaking at 449 in 2025 [2]. Philanthropic funding has scaled in parallel: the Simons Foundation committed $50 million over five years to solar radiation management research [3], while Silicon Valley figures including former Google executive Alan Eustace have joined Bill Gates and George Soros as prominent backers [4]. Yet the governance infrastructure to regulate any of this remains fractured. A 2024 UN Environment Assembly resolution on solar geoengineering collapsed after the United States, Saudi Arabia, and Brazil blocked consensus, while African nations and small island states pushed for an outright ban on outdoor experiments [5].

Source: OpenAlex

Data as of Jan 1, 2026CSV

The Case for Silica Over Sulfur

Stratospheric aerosol injection as a concept dates to the 1970s, but most research has focused on sulfur dioxide — the compound volcanoes inject during large eruptions. The 1991 eruption of Mount Pinatubo, which temporarily cooled the planet by roughly 0.5°C, serves as the primary natural analogue [6]. Climate models estimate that sustained injection of approximately 10 teragrams (Tg) of SO₂ per year could offset about 1°C of warming [6]. The IPCC's Sixth Assessment Report concluded with "high agreement" that SAI could limit warming to below 1.5°C [6].

But sulfur has known drawbacks. Sulfate aerosols accelerate ozone depletion, produce acid rain when they eventually settle, and heat the lower stratosphere in ways that alter atmospheric circulation patterns [7]. These side effects have driven a search for alternative materials. Research led by David Keith, John Dykema, and colleagues at Harvard examined solid aerosol particles — alumina (Al₂O₃), diamond, and silica — that could produce equivalent radiative forcing with less ozone damage, less stratospheric heating, and less forward light scattering than sulfates [7][8].

Silica spheres occupy a middle position among these alternatives. Amorphous silica — essentially the same material as glass — is chemically inert at stratospheric temperatures and does not catalyze the chlorine-activation reactions responsible for ozone destruction the way sulfate particles do [8]. A 2016 paper by Keith and Dykema proposed that alkaline metal salt aerosols could actually reverse ozone depletion while providing cooling, a finding that reframed the entire risk calculus for non-sulfate approaches [9].

Source: Compiled from Keith et al. 2016, Dykema et al. 2016, IPCC AR6

Data as of Jan 1, 2025CSV

Stardust Solutions' 0.5-micron silica spheres are engineered to "recycle into existing natural cycles after they settle to the ground" [1]. The company is also developing a variant with a calcium carbonate core for enhanced solar blocking [1]. Precise cooling-per-teragram figures for silica specifically remain less established in the peer-reviewed literature than for sulfate or alumina. Model estimates compiled across Keith et al. (2016), Dykema et al. (2016), and IPCC data suggest silica yields roughly 0.09°C of cooling per Tg deployed annually — somewhat less than alumina (~0.12°C/Tg) but comparable to sulfate (~0.1°C/Tg) [7][8]. Diamond particles show the highest theoretical efficiency at roughly 0.15°C/Tg, but their production cost makes them impractical at scale [8]. These estimates carry wide uncertainty bands across models; the Geoengineering Model Intercomparison Project (GeoMIP) has shown that different climate models disagree on regional temperature responses even when they converge on global means [6].

What Happens After Deployment

Once injected at roughly 18-20 km altitude, stratospheric aerosols of any type persist for one to two years before gravitational settling and stratospheric circulation bring them back to the troposphere, where rain washes them out [10]. For silica, the atmospheric residence time is expected to fall within this range, though the precise duration depends on particle size and injection latitude [8].

At stratospheric temperatures (approximately -50°C to -80°C), amorphous silica is largely unreactive with ozone, water vapor, or nitrogen oxides — a significant advantage over sulfate, which provides surfaces for heterogeneous chlorine chemistry that destroys ozone [9]. However, the interaction between silica particles and sulfuric acid droplets already present in the stratospheric Junge layer has not been fully characterized in laboratory conditions. Researchers at ETH Zürich and other institutions funded through the Simons Foundation program are now investigating these interactions [3][11].

A novel composite material — diamonds dispersed in a silica aerogel matrix — has also been studied for its potential to combine high refractive index scattering with ultralow density, though this remains at the laboratory stage [8].

The Monsoon Problem: Billions at Risk

The most consequential objection to any form of stratospheric aerosol injection — silica or otherwise — concerns precipitation. The South Asian summer monsoon supplies water and sustains agriculture for nearly two billion people [12]. Multiple model studies show that SAI would reduce mean and extreme summer monsoon precipitation, driven by lower-stratospheric warming that weakens the northern hemisphere subtropical jet [12][13].

A 2024 study published in npj Climate and Atmospheric Science found that SAI scenarios consistently weakened the Indian monsoon, with one model run producing outright monsoon failure — an outcome the authors described as "a disaster for millions of people" [12]. A separate study in Environmental Research Letters determined that a regional aerosol optical depth of 0.25-0.5 would trigger droughts across India, directly affecting over a billion people [13].

The populations bearing these risks are overwhelmingly in the Global South. The United Nations Human Rights Council's Advisory Committee has warned that geoengineering technologies "could significantly infringe on human rights for millions and perhaps billions of people," with disproportionate impact on communities that have contributed least to greenhouse gas emissions [14]. Developing nations dependent on monsoon-fed agriculture — India, Bangladesh, Pakistan, sub-Saharan African nations — face hydrological disruption from decisions they had no role in making [14][15].

This asymmetry is not hypothetical. At UNEA-6, the African Group of nations, joined by Colombia, Mexico, Fiji, and Vanuatu, explicitly called for a ban on outdoor solar geoengineering experiments, citing the risk that wealthy nations would impose climatic side effects on poorer ones without consent [5]. The resolution failed — withdrawn after high-emitting nations opposed precautionary language [5].

Termination Shock: The Withdrawal Problem

If large-scale silica deployment were halted abruptly — whether from geopolitical conflict, funding collapse, or a change in scientific consensus — temperatures would rebound toward levels they would have reached without intervention. This "termination shock" is frequently cited as one of the most severe risks of any SAI program [10][16].

The dynamics are straightforward: stratospheric aerosols persist for only one to two years after injection stops. As the aerosol layer thins, the radiative forcing it had been offsetting returns. The rate of warming during rebound could be "several times what it would have been under climate change alone," according to analysis from SRM360 [10]. A 2018 study by Parker and Irvine in Earth's Future examined the likelihood and severity of termination shock, noting that if SAI were masking a large degree of warming — say, 2°C — sudden cessation would produce rapid temperature increases that ecosystems and human systems would struggle to adapt to [16].

For silica specifically, the termination shock profile does not differ materially from sulfate aerosols. Both particle types have similar atmospheric residence times of one to two years. A one-year interruption in deployment would produce approximately a 20% loss of cooling effect, with recovery possible if injection resumed before the aerosol layer fully dissipated [10]. The key variable is not the aerosol material but the magnitude of warming being masked: a modest deployment offsetting 0.3°C would produce a manageable rebound, while a deployment masking 2°C or more would create conditions far more dangerous than gradual warming [16].

This creates what critics call a "commitment trap." Once deployment begins at scale, the political and physical costs of stopping become so high that continuation becomes effectively mandatory — for decades or centuries [15].

Follow the Money

The funding landscape for solar geoengineering research has shifted from a handful of individual donors to a broader ecosystem of philanthropies and, increasingly, for-profit ventures. Bill Gates began financing geoengineering research in 2007 through the Fund for Innovative Climate and Energy Research (FICER), providing at least $4.5 million for studies of stratospheric aerosols and related approaches [17]. The William and Flora Hewlett Foundation, the Simons Foundation ($50 million), and the Environmental Defense Fund, which has publicly supported geoengineering research since 2011, now form a significant institutional funding base [3][4].

Stardust Solutions represents a different model: a for-profit company that has raised $75 million from investors and projects revenues of $1.5 billion annually under "global full-scale deployment" by 2035 [1]. Critics have flagged the structural problem. "Self governance led by for-profit entities does not work," one expert told E&E News [1]. A company that develops, deploys, and monitors its own planet-altering technology has an inherent conflict of interest that no internal governance structure can resolve.

The fossil fuel industry's relationship to geoengineering funding is more indirect but persistent. Harvard's Solar Geoengineering Research Program adopted an explicit policy refusing donations from entities whose profits predominantly come from fossil fuels, citing concerns that "fossil fuel companies will seek to exploit solar geoengineering as a pretext for delaying reductions in greenhouse gas emissions" [18]. The ETC Group has documented Gates' simultaneous investments in fossil fuel technologies alongside geoengineering research, arguing that the two portfolios serve complementary interests [17]. At UNEA-6, Saudi Arabia — the world's largest oil exporter — joined the United States in opposing governance measures that would have restricted geoengineering experimentation [5][15].

The Case for Keeping Labs Open

Against this backdrop of concern, a counter-argument has gained traction: that suppressing geoengineering research is itself a form of harm.

The Brookings Institution has argued that without proper research, "responsible research programs will be sidelined and slowed, and ad hoc and commodified private-sector initiatives will likely take advantage of the gap" [19]. If climate change reaches a point where emergency deployment becomes politically irresistible — a sustained heat wave killing tens of thousands, a sudden ice-sheet collapse — making that decision with inadequate scientific data would be far more dangerous than the research itself.

This argument draws strength from the trajectory of private actors. Mexico banned solar geoengineering experiments in 2023 after the startup Make Sunsets released sulfur dioxide from weather balloons without government authorization [15]. The incident demonstrated that the absence of a research framework does not prevent deployment — it merely ensures that deployment, when it comes, is conducted by actors operating outside any scientific or regulatory structure.

Over 500 academics across more than 50 countries have signed an open letter calling for an International Non-Use Agreement on solar geoengineering [20]. But even some signatories distinguish between deployment and research. The Environmental Defense Fund has proposed that small-scale, transparent, and publicly governed research programs are necessary precisely to inform the governance decisions that non-use agreements require [4]. The question is whether society can maintain the distinction between studying a technology and normalizing its use.

A Governance Vacuum

The international legal framework for stratospheric silica injection is, in practical terms, a patchwork of instruments that were designed for other purposes.

The Convention on Biological Diversity (CBD) established a de facto moratorium on geoengineering in 2010, reaffirmed at COP16 in 2024. The moratorium permits only "small-scale scientific research studies conducted in a controlled setting" with full environmental impact assessment and no transboundary harm [20]. It explicitly bars commercial purposes. But the CBD moratorium is non-binding, carries no enforcement mechanism, and does not apply to nations that are not party to the convention — notably, the United States [20].

The London Convention/London Protocol prohibits ocean fertilization and has moved to regulate additional marine geoengineering techniques, but its 2013 amendments have not yet entered into force [20]. Stratospheric injection falls outside its marine scope entirely.

The Environmental Modification Convention (ENMOD) prohibits hostile use of environmental modification but was designed for military contexts and has not been applied to climate interventions undertaken with ostensibly beneficial intent [20].

No existing treaty specifically addresses stratospheric aerosol injection by any actor — state or private. The Center for International Environmental Law has documented that at least 598 outdoor geoengineering experiments have been proposed since 1971, with a tripling of solar geoengineering experiments between 2014-2018 and 2019-2023 [20]. This acceleration has occurred in a legal environment where, as a practical matter, a sufficiently resourced private company or a unilateral state actor could begin deployment with no international body empowered to stop them.

Precedents: From Haida Gwaii to SCoPEx

The governance vacuum is not theoretical. In 2012, the Haida Salmon Restoration Corporation dumped 120 tonnes of iron sulfate into the Pacific Ocean off Haida Gwaii, British Columbia, triggering a 10,000-square-kilometer plankton bloom visible from space [21]. The project was led by Russ George, former CEO of the defunct carbon-credit firm Planktos. Only one of the two local Haida band councils had approved the project; many Haida individuals rejected it [21]. The dump appeared to violate the CBD moratorium and the London Convention framework, but because those instruments lack enforcement teeth, no legal consequences followed [21].

Harvard's Stratospheric Controlled Perturbation Experiment (SCoPEx) met a different fate. Planned as a small-scale balloon test to release calcium carbonate particles in the stratosphere above Kiruna, Sweden, the project was halted in 2021 after the Sámi Council — representing Indigenous Sámi people across Scandinavia — objected that researchers had failed to engage in meaningful consultation [22][23]. A letter signed by 35 Indigenous groups worldwide called for SCoPEx's complete shutdown [22]. In March 2024, lead researcher Frank Keutsch announced the project would not move forward [23].

The contrast is instructive. The Haida Gwaii experiment — large-scale, irreversible, conducted by a for-profit actor with minimal consent — faced no legal consequences. SCoPEx — small-scale, designed to be reversible, conducted by a university with an advisory committee — was stopped by social pressure and Indigenous advocacy. The lesson many governance scholars draw is that the current framework penalizes transparent research while leaving rogue actors unconstrained [19][20].

Stardust Solutions' planned 2026 silica deployment sits uncomfortably between these precedents. It is larger than SCoPEx, commercially motivated unlike Harvard's program, and — absent new legislation — faces no clear legal barrier [1].

What Comes Next

The silica sphere proposal crystallizes the central tension in climate politics: the gap between what atmospheric physics makes possible and what political institutions can govern. The science suggests that engineered silica particles could provide measurable cooling with fewer ozone side effects than sulfur. The models also suggest that the same intervention could weaken monsoons that feed two billion people, create a commitment trap lasting centuries, and hand for-profit actors control over a planetary commons.

Current policies project roughly 3°C of warming by 2100 [15]. The first year to exceed the 1.5°C threshold was 2024 [15]. Against that trajectory, the pressure to deploy some form of solar geoengineering will intensify regardless of what any treaty says.

The question is whether the institutions catch up before the technology does.