The uncomfortable promise: we could cool Earth quickly, but we might not like the side effects
What if the fastest way to slow global warming is not cutting emissions, but dimming the Sun a little? That idea sounds like science fiction until you learn that nature has already run the experiment for us. Big volcanic eruptions have cooled the planet measurably by lofting reflective particles high into the atmosphere.
The real question for the next hundred years is not whether humans can influence the climate. We already do. The question is whether we can do it deliberately, predictably, and fairly enough that the world would accept it, and whether it would actually reduce risk rather than reshuffle it.
What "geoengineering" actually means, and why the definition matters
Geoengineering is an umbrella term that hides two very different toolkits. One aims to reduce how much sunlight the Earth absorbs. The other aims to reduce how much carbon dioxide is in the air. They are often discussed together, but they behave differently, scale differently, and carry different kinds of risk.
Solar Radiation Management, often shortened to SRM, tries to cool the planet by reflecting a small fraction of incoming sunlight back to space. The most discussed approach is stratospheric aerosol injection, which would release reflective particles high above weather systems. Other proposals include brightening marine clouds or increasing the reflectivity of surfaces on land.
Carbon Dioxide Removal, or CDR, tries to lower atmospheric CO directly. That includes direct air capture machines, enhanced weathering using crushed rock, bioenergy with carbon capture and storage, and a range of ocean based ideas. CDR is closer to "cleaning up" than "turning down the thermostat," but it is slower and usually more infrastructure heavy.
The physics is not the hard part for SRM
SRM's appeal is speed. If reflective material is placed in the stratosphere, climate models and historical evidence suggest global average temperatures could respond within months. That is why SRM is often framed as an emergency brake for a world that overshoots temperature targets or faces abrupt climate impacts.
The basic mechanism is well understood. Increase planetary reflectivity and you reduce the energy imbalance that drives warming. After the 1991 Mount Pinatubo eruption, global temperatures dipped for roughly a year, offering a real world demonstration that stratospheric particles can cool the surface.
But "can cool" is not the same as "can safely manage." Cooling the global average does not guarantee stable rainfall patterns, healthy monsoons, or predictable storm tracks. SRM changes the distribution of energy in the climate system, and distribution is where politics and livelihoods live.
SRM could be technically feasible within decades, yet still be practically impossible
From an engineering perspective, lofting material into the stratosphere is not beyond human capability. Aircraft, balloons, or other delivery systems could in principle maintain a steady layer of reflective particles. Compared with the scale of global energy systems, the direct operational costs are often described as relatively low.
The barriers are more about uncertainty and legitimacy than hardware. Even if a program could be run for a few billion dollars a year, who decides the target temperature? Who decides how to respond if one region experiences drought while another benefits? And what happens if a future government stops the program abruptly, causing rapid warming as the masking effect disappears?
This "termination shock" problem is one of SRM's defining risks. SRM does not remove greenhouse gases. It potentially buys time, but it also creates a dependency if emissions remain high. A century is long enough for political systems to change many times, which makes long term continuity a central concern.
CDR is the opposite: slow, expensive, and ultimately unavoidable
If SRM is a fast acting painkiller, CDR is physical therapy. It is slower, harder, and more expensive, but it addresses the underlying cause. Removing CO reduces not only warming but also some related impacts such as ocean acidification, which SRM does not fix.
The challenge is scale. Today's direct air capture plants operate at relatively small volumes compared with global emissions. To make a noticeable dent in atmospheric concentrations, the world would need to remove billions of tonnes of CO per year for decades. That implies vast clean energy supply, large industrial buildouts, and reliable storage sites where captured carbon stays locked away.
Enhanced weathering has a different profile. It uses natural chemical reactions between silicate rocks and CO, accelerated by mining, grinding, and spreading rock on land or in the ocean. The chemistry is promising, but the logistics are enormous. Moving and processing billions of tonnes of rock each year would rival major global commodity supply chains, and it would bring local environmental tradeoffs that cannot be hand waved away.
Why "possible" depends on what you mean by success
There is a version of geoengineering that is clearly possible within a century: limited, targeted deployment that measurably affects climate variables. SRM could likely achieve that sooner than CDR because it acts quickly and does not require building a carbon removal industry the size of today's fossil fuel sector.
There is another version of geoengineering that is far harder: a stable, globally coordinated climate management system that reduces overall risk, is perceived as legitimate, and can be sustained across generations. That is not just a technical milestone. It is a governance milestone, and history suggests governance is the slower technology.
Even for CDR, "possible" can mean different things. It might mean reaching a few gigatonnes per year of removal by late century, which could be plausible with strong policy and investment. Or it might mean removing enough CO to reverse decades of high emissions, which would require sustained effort on a scale the world has rarely maintained for anything other than war.
The governance problem is not a footnote, it is the main plot
Geoengineering forces a question that climate policy has often avoided. Who gets to set the global climate dial? Emissions cuts are distributed across countries and sectors, messy but familiar. SRM, by contrast, could be deployed by a small coalition with outsized influence on everyone else.
That creates a risk of unilateral action, whether by a nation, a group of nations, or even a wealthy private actor. It also creates a risk of paralysis, where fear of conflict prevents even careful research. Both outcomes are dangerous. A world that refuses to study SRM could still end up using it in a panic, with less evidence and weaker norms.
Over the next hundred years, the most important "technology" may be the creation of credible rules. That includes transparency for experiments, shared monitoring, liability frameworks, and clear thresholds for what counts as deployment versus research. It also includes representation for regions most vulnerable to rainfall shifts, not just the countries with the biggest research budgets.
What the next century could realistically look like
In the near term, expect more modeling, more lab work, and more political argument about whether outdoor experiments should proceed. Small scale tests, if they happen, are likely to be designed to validate atmospheric chemistry and particle behavior rather than to cool anything meaningfully.
By mid century, CDR is likely to expand because it fits more comfortably within existing climate policy. It can be regulated like an industry, financed through carbon markets or public procurement, and measured in tonnes. If costs fall and clean energy becomes abundant, direct air capture could move from boutique projects to a meaningful contributor, though still not a substitute for emissions cuts.
SRM's trajectory is harder to predict because it depends on climate impacts and geopolitics. If warming triggers severe, widely felt disruptions, pressure for an emergency response could rise. In that scenario, SRM might be considered as a temporary measure to reduce peak temperatures while emissions fall and CDR scales. If impacts are less abrupt, SRM may remain largely in the research domain, controversial and constrained.
The risks that don't fit neatly into a spreadsheet
SRM's biggest scientific worry is regional disruption. Cooling the planet on average could still shift precipitation patterns, potentially affecting agriculture and water security. The uncertainty is not an argument for ignoring SRM, but it is an argument for humility about what models can guarantee.
CDR's biggest risk is overpromising. If policymakers assume future removal will be cheap and massive, they may delay emissions cuts now. That is not a hypothetical. Many net zero plans already lean heavily on removals that do not yet exist at scale.
Both approaches also carry a trust problem. Communities asked to host CO pipelines and storage sites will demand safety and accountability. Countries asked to accept altered skies will demand voice and veto power. Without trust, even technically sound projects can fail.
So, is geoengineering the climate possible by 2125?
If "possible" means the ability to measurably cool the planet for a period of time, SRM could be within reach well before 2125. The physics is established, and the engineering is not unimaginable. The real uncertainty is whether the world will build the governance to use it without turning climate policy into a new arena for conflict.
If "possible" means lowering atmospheric CO enough to materially reduce long term warming and related harms, CDR can do it in principle, but only with sustained industrial scale growth and political commitment across decades. Within a century, meaningful CDR is plausible. Planet cleaning at the scale of past emissions is a much taller order.
The most realistic future is not a single silver bullet, but a tense coexistence of tools. Rapid emissions cuts to stop adding to the problem, CDR to unwind part of the damage, and SRM sitting on the shelf as a potential emergency measure that the world hopes it never needs, yet cannot afford to ignore.
Because the hardest part of geoengineering is not learning how to change the sky, it is learning how to share it.