Depending on whom you ask, the climate agreement that just came out of the COP28 conference is either an upset—having been agreed on in the United Arab Emirates, a petrostate—or a disappointment, or maybe something in between. (Climate change is complicated—and climate politics more so.) Regardless, for the first time ever, nations have agreed to transition away from fossil fuels. It’s not as ambitious as agreeing on a phase-out of those fuels, but it’s at least a path toward decarbonization.

The UN’s Intergovernmental Panel on Climate Change has for years pleaded that humanity needs to stop emitting greenhouse gases or else face ever-worsening climate disasters. But the IPCC has also stressed that we’ll need to remove carbon from the atmosphere too, to bring temperatures down—especially if we overshoot the Paris Agreement’s goal of keeping warming below 1.5 degrees Celsius, as is now looking virtually certain. The COP28 agreement briefly mentions accelerating these carbon removal technologies.

“Delaying carbon removal creates its own risks that we start to diverge from where the science is telling us that we need to go,” says Ben Rubin, executive director and cofounder of the Carbon Business Council, a coalition of carbon management companies. “Carbon removal unequivocally needs to work in tandem with the important work of reducing emissions.”

The COP28 agreement is strong in that it calls for a 60 percent drop in emissions by 2035, says Gregory Nemet, who studies low-carbon innovation at the University of Wisconsin-Madison. But its ambition also doesn’t reflect the emergency humanity is in: The more carbon we pump into the atmosphere, the more we’ll have to rely on carbon removal technologies to keep temperatures from skyrocketing further.

“The UAE Consensus is weak in not including an agreement to peak global emissions by 2025 and not including an agreement to stop investment in new fossil fuel infrastructure,” says Nemet, who coauthored a recent COP report that addresses carbon removal. “It’s also clear that there is nowhere near enough capacity for carbon removal to offset continued use of fossil fuels, so the imperative to ‘transition away from fossil’ is important for climate goals, even if it will require substantial carbon removal.”

Carbon removal brings with it a climate conundrum: Humanity has to do it, but done the wrong way, critics worry it could distract from that ultimate goal of rapidly reducing emissions, by pulling both funding and research resources away from clean energy. In the worst-case scenario, carbon removal might even encourage the continued burning of fossil fuels, if countries can say they’re sucking carbon out of the atmosphere to offset their emissions. (That’s what’s known as net zero. The COP28 agreement calls for reaching this globally by 2050.)

Carbon removal techniques come in two main flavors—the technological and the natural—but also increasingly a melding of the two. The dominant technology at the moment is direct air capture, or DAC. This involves giant machines sucking in air and filtering out the CO₂. Like an air purifier filters dust out of your indoor air, DAC facilities scrub the atmosphere of carbon. Technically, carbon removal is distinct from carbon capture, which intercepts the gas at the source, such as a power plant, before it reaches the atmosphere.

DAC, though, is a nascent technology, and it’s nowhere near operating at the scale needed to put a proper dent in global emissions. In 2021, researchers calculated that it would take a huge annual investment of between 1 and 2 percent of the global gross domestic product to remove about 2.3 gigatons of CO₂ annually by the year 2050. To put that in perspective, global emissions of CO₂ are currently around 40 gigatons a year—and are unfortunately still rising instead of going down. The 2021 study found that we’d need between 4,000 to 9,000 DAC facilities by the year 2075, and more than 10,000 by 2100, to theoretically be sequestering up to 27 gigatons of carbon a year. (The idea behind this rapid scale-up being that as the technology and industry progresses, it gets easier and cheaper to deploy more plants.)

So DAC could play some part in removing carbon from the atmosphere, and it will grow more impactful the more carbon we can stop emitting, since there will be less to clean up. But it’ll take a whole lot of cash. “Could we actually scale it up fast enough to go from a couple million tons a year now to, say, a billion tons a year in 2050?” asks Nemet. “That’s where I’m actually more optimistic. We could do it, but it’s challenging. That doesn’t change at all our policy right now, or what our goal should be: We need to really quickly start reducing emissions and get down to close to zero by 2050.”

Even if DAC were scaled up massively, it couldn’t alone save us from ourselves. If it’s removing a billion tons of CO₂ annually in 30 years, and humans are still emitting tens of billions of tons of the gas, it’ll be like trying to drain a bathtub with the tap still running. One promise of carbon removal, though, is that it could help offset future emissions from hard-to-abate sectors, like the steel industry, which require enormous amounts of fossil-fuel power to run. Unlike making a home fully solar-powered, you can’t just slap panels on these factories and call it a day.

But with carbon removal and carbon capture come the dreaded “moral hazard”: If we’ve got technology, why worry so much about decarbonization? Why bother with solar panels and wind turbines if we can just cancel out our carbon emissions? “Where it is not helpful at all—and in fact is really counterproductive and has been counterproductive for years—is as a PR stunt. That’s what we’re seeing mostly now,” says Jonathan Foley, executive director of Project Drawdown, which advocates for climate action. “It’s become a talking point for Big Oil. It’s clear that they are using that to delay the phase-out of fossil fuels. So that’s where carbon capture goes wildly off the rails, is these big, expensive industrial techniques that really don’t work.” Indeed, instead of the COP28 agreement calling for a phase-out, like scientists and climate activists had hoped, it calls for a transition away from fossil fuels, which is better for Big Oil.

(And keep in mind that burning fossil fuels doesn’t produce just greenhouse gases, but also particulate air pollution that’s responsible for one in five deaths worldwide. Phasing out these fuels will save a whole lot of lives and stop global warming.)

As an alternative, Foley points to the other variety of carbon removal—the natural kind. Just as they’ve done for hundreds of millions of years, trees suck in CO₂ and sequester the carbon in their tissues. The idea behind “nature-based solutions” is to protect as many ecosystems as we can, especially wetlands and rainforests like the Amazon, so they can naturally remove carbon from the atmosphere. But unfortunately, humans are headed in the opposite direction: The Amazon is now so degraded due to deforestation, parts of it are switching from being carbon sinks to carbon sources.

While removing carbon with DAC is easily quantifiable—machines run for x number of hours to collect y tons of the gas—nature doesn’t work so neatly. Scientists are still figuring out just how much carbon can be stored in a given ecosystem—in all the plants and even the soil—and for how long. And while DAC can store carbon underground, locking it away for the long term, natural carbon isn’t so safe. If you restore a forest only for it to be obliterated by wildfires—which are growing ever more powerful as the climate warms—your carbon goes right back into the atmosphere.

Also, one political regime might take pains to restore an ecosystem, only for another to come along and destroy it again. “The obvious problem here is the question of permanence: We planted the trees, we put the soil carbon back. But how long is it going to really stay there in a warmer, more dangerous climate?” says Foley. “With the nature-based solutions, we’re kind of putting ourselves on a treadmill. We might have to continually restore nature. But a treadmill where the world is engaged in restoring nature doesn’t sound so bad to me.”

To address this problem of permanence, researchers are pursuing hybrid approaches to carbon removal: engineering combined with nature. With “enhanced rock weathering,” for instance, you sprinkle basalt dust across fields before planting crops. The rock reacts with CO₂ in the air, sequestering it as bicarbonate, which then washes out to sea. Similarly, turning biomass waste into biochar and adding it to soils sequesters the carbon that the biomass had itself pulled from the atmosphere as it grew. (You might also bury the biochar underground.) Elsewhere, researchers are gene-editing tree species to grow faster and therefore capture CO₂ quicker, potentially supercharging tree-planting campaigns.

Like a good investor would diversify their portfolio so they don’t lose all their money if one company fails, so too are carbon-removal types diversifying their options. If one technique doesn’t end up scaling well enough, maybe others will in the coming decades. “We’re supporting a diversity of pathways, because each one of them does offer a diversity of different benefits,” says Rubin of the Carbon Business Council. “So biochar, as an example, can help be fertilizer for soil.”

A diversity of interventions would also create a diversity of jobs—making biochar or basalt dust for enhanced rock weathering, or simply planting trees. (In September, the Biden administration announced the mobilization of the American Climate Corps, which will put young folks to work preparing the landscape for the pain ahead.) Restoring ecosystems boosts biodiversity, in turn attracting tourists and their money to communities.

The ever-present danger with carbon removal, Foley stresses, is that we may lose sight of the ultimate goal: reducing emissions. “I think we need to be able to be more nuanced in this conversation, saying there’s bad carbon removal and there’s the possibility of some good carbon removal,” Foley says. “But no matter what, 90 to 95, 99 percent maybe of the real work here is still going to be cutting emissions—no matter what you believe about carbon removal.”