Net zero, part 5: Stopping climate change

Recapping the series, an international treaty, and some asides.

Sam Harsimony

Aug 08, 2025

This is part of a series of posts on how to pause climate change using technology and policy:

Net zero, part 1: energy

Net zero, part 2: transport

Net zero, part 3: production

Net zero, part 4: agriculture

Net zero, part 5: stopping climate change (you are here)

In the short term, it’s better to pause climate change than revert to a pre-industrial climate. To pause climate change, it’s sufficient to achieve net zero emissions across the economy.

Fixing climate change is easy, just inject SO2 into the stratosphere to balance out the radiative forcing from all greenhouse gas emissions. You have to stay on a 1000 year schedule for each gigaton of carbon released and if you ever stop it will fry the planet.

Too hard? Fine: use enhanced weathering at scale to reach net zero. At $100/ton that will cost $4 trillion annually for the foreseeable future. That’s around 4% of world GDP.

Too expensive? Then we’re stuck employing a host of technologies and policies to reach net zero without too much effort or expense. This post organizes the series into a hopefully coherent plan.

Every piece needs to be addressed (source)

Clean tech makes stopping climate change cheap

Let’s start by recapping the series.

In Energy, I argued that renewables will just win when it comes to electricity generation. In the short term, natural gas peaker plants can provide the 10% flexible power needed to dance with renewables. Longer term, point source capture of natural gas emissions or cheaper renewables can fill the gap. Green industrial heat is harder, it needs point source capture for now, but may be overtaken by renewable sources of heat in the future.

In Transport I argued that a carbon tax would have minimal impacts on most industries. For commuter cars and short-haul trucking, batteries are just better than fuel. Electric trucks can eventually move into long-haul trucking. Only air freight and shipping see significant (15%) price increases, though there are dozens of ways to adapt.

In Production, a carbon tax would have little effect on most segments. Cement needs a few innovations to scale and would pay green premium, but should be fine. Ammonia however is not ready. We need to switch away from producing urea and generate clean hydrogen as cheap as natural gas. That’s hard.

In Agriculture I concluded that paying to clean up agricultural emissions is feasible. Short lived pollutants can be mitigated with SO2 injection for $146B/yr at most. Agrochemicals would see little cost increase from a carbon tax. Emissions from land use change are expensive, but likely to go away as population falls mid-century.

Renewable energy, point source capture, enhanced weathering, and SO2 injection feature heavily in the solutions to these problems. Together, clean technologies can stop climate change for a low cost. With a few developments, the green premium in most industries would be less than 5%.

Just one problem: why would anyone switch to a slightly-more-expensive process?

Coordination problem: some clean tech isn’t competitive

Some industries will continue to emit CO2. For instance, cement production requires us to bake CO2 out of rocks. And electric long distance flights won’t happen in the foreseeable future. Also, some people just like driving gas cars; we should let them.

If we want to reach net zero, we need to pay extra to make these things green. Not everyone will want to pay. The costs are surmountable if everyone splits the bill, but free-riding is unsustainable.

So we need a coordinating device to get (almost) everyone to pay their share towards net zero. The most practical way to do that is an international treaty.

International treaty to pay for climate change

International treaties are the best global coordinating force we have. They mostly sort of work. Trying to broker some agreement without government buy-in is impractical; building some sort of global governing body seems impractical at best and dangerous at worst.

Governments are more amenable to spending money on stuff than changing domestic policy. So the best treaty is an agreement to collectively pay the cost to reach net zero via carbon capture and SO2 injection. The payments should be proportional to emissions1. That way, states can save money by subsidizing green technology or charging a carbon tax domestically.

There’s no guarantee that states will sign on to such a treaty2, but it’s crucial if we want to reach net zero in a reasonable time frame. I think that if the U.S. and China can reach an accord, most major economies would fall in line. Add Japan, South Korea, India, Australia, the EU, the UK, Canada, and Mexico to the agreement and you have united over 78% of world GDP. These countries (plus international flights and shipping) account for 71% of annual emissions3.

Brokering such an agreement is the key challenge to reaching net zero. The next few sections detail what it might look like.

The transition plan

Switching to a green economy on a short time frame is impractical. Instead, the countries in the agreement need to take steps to bring the cost down. I imagine this happening in stages in order of difficulty:

Each country’s energy sector aims for 90% renewables, with natural gas making up the remainder.
Countries pay the SO2 injection cost for global agricultural emissions.
Phase in a capture tax for gasoline and diesel used in road transport. Use the proceeds to pay for carbon capture.
Electrify or capture all industrial heat production.

Step 1 is essentially free as states can enjoy a cheaper and more resilient energy system. With advancements in SO2 injection, step 2 is a pure payment, but trivial for states to cover, especially considering how many agricultural subsidies there are already. My initial estimate is $146B/year, but that could fall by 10x or more.

Step 3 will be politically challenging. But it would be unreasonable for states to cover these emissions directly, costing $450B/yr worldwide. Without a tax, people won’t switch to electrified transport, leaving states to pay indefinitely. This is precisely the sort of hard-but-necessary political decision that an international treaty is supposed to bring about. Subsidies for electric vehicles and updated car taxes may blunt the impact.

Step 4 requires aggressive technological scaling and nuanced policy. Industrial heat encompasses all sorts of use cases. A different approach may be required for each industry. Heat pumps, electric arc furnaces, on-site renewables, and thermal storage will feature heavily here and need more support. Cheap point source capture could be retrofitted to existing plants.

After these steps are complete, paying for CO2 capture to reach net zero is far more feasible.

Paying the remainder

The plan above leaves out some important domestic emitters that treaty states will pay for. Going off of this breakdown of major emissions sources, they are:

Natural gas energy (3%)4
Direct cement emissions (3%)
Ammonia (1.8%)
Airplanes (1.9%)
Shipping (1.7%)

That adds up to 11.4% of CO2 emissions which would cost $431B/year5. That is 0.52%6 of the GDP of participating countries. The percent will fall as these economies grow and more countries sign on. For example, if the agreement starts in 2060 and these economies grow at only 2%, the fraction of GDP falls to 0.26%.

Technological developments will lower this cost even further. I speculate that point source capture would halve the cost of cleaning up natural gas, cement, and ammonia. Hybrid electric airplanes running on natural gas could emit 4x less CO2. Electric cargo ships look feasible as well.

When you consider the benefits of a green economy, the net cost is smaller still. Cheaper electricity, energy independence, eliminating fossil fuel price shocks, less air pollution, and a stable climate are all worthwhile.

Smog in Beijing vs. in clear weather (source)

Getting other countries on board

While this plan is sufficient to cover the emissions of the participating countries (which is 71% of all emissions), we still need to account for nonparticipating countries.

A lot of countries might switch to a high renewables grid because it’s cheaper. And I’m optimistic that batteries can take over ground transport eventually. Global agricultural forcing can be paid for with pocket change from the rich countries. But that leaves us with three problems in nonparticipating countries:

Maybe EV’s don’t win and a significant fraction of vehicles emit CO2.
They might not capture the CO2 from natural gas plants.
Industrial heat might not electrify or capture emissions.

Consider Russia: it’s cold, uncooperative, and stagnant. It’s highly unlikely that Russia will switch to green technology any time soon.

While the treaty states can try to bring more countries on board, the best bet is to make climate tech competitive with fossil fuels in as many areas as possible. That way other states will switch on their own.

That said, we can’t assume that green technology will completely displace fossil fuels. If these other countries move to a 90% clean grid but do nothing to change industrial heat or road transport, I estimate they’ll produce the equivalent of 10% of today’s emissions7.

Treaty members would need $378B/year to capture that with DAC. I don’t think that’s worth it. Instead, consider what the treaty has already achieved; a 90% reduction in emissions gives us 10x more time to find solutions. Treaty members can bide their time, let clean technologies diffuse, and grow their economies until paying the remainder is trivial.

Scaling cleantech

We will need to continue scaling clean technologies that will lower the cost of the green transition. Getting CO2 capture and storage below $100/ton is critical to making an international treaty feasible.

Beyond carbon capture, the biggest challenges are in the production sector; industrial heat and point source capture stand out as particularly important (see the Production post for more). Reassuringly, renewables and batteries have already seen success. I expect them to dominate energy and transport in the future.

My hope is that these clean technologies will outcompete fossil fuels even in countries that aren’t part of the treaty. With enough time, the remaining emissions may be trivial for treaty members to cover.

Conclusion

So that’s it, my plan to reduce emissions by 90% and let time grind down the last 10%. With enough R&D, we can achieve that for a reasonable cost. This cost will be borne out by a small group of the most powerful and generous countries, who will be remembered for slaying the greatest collective challenge humanity has yet seen.

After looking at the future of cleantech, it feels like we’re nearing an equilibrium. The end of the century may see the end of terrestrial engineering. We found cheap energy sources that won’t be usurped in the forseeable future. The quirks of renewables will propagate into our production processes. After that, what’s left to change?

Ironically, our energy system and thus everything else will depend on the weather. A return to an almost agricultural way of life after centuries of bootstrapping from fossil fuels. But this time, the sun will provide us orders of magnitude more energy. A technology that will quite literally scale to the stars.

It’s of personal significance to be able to crystallize this stuff. After years thinking about climate change, I can finally see the ending from afar. A society whose power grows with each rising sun, secure in the knowledge that its way of life can continue indefinitely.

Appendix

A collection of minor points that didn’t fit into other parts of the series.

Hot takes on a warming earth

Developing countries might leapfrog fossil fuels. Much of their energy infrastructure remains to be built and they are often near the equator with good solar resources. As solar and batteries get cheaper, a renewables dominated grid will make economic sense. Industries adapted to intermittency with on site energy and storage will come about naturally. Look at what’s happening in Pakistan for example.
It’s better to let developing countries grow and pollute than to restrict their emissions.
Rich countries should subsidize the adoption of green technologies in developing countries to reduce emissions.
Clean ammonia and renewable industrial thermal are the biggest question marks. I believe they can win in sunny places. Wind thermal is another interesting approach.
Things are unpredictable and I’m probably wrong about whether certain technologies are viable. We should push Pareto fronteirs rather than dismiss climate technologies that don’t fit into this plan.
CO2 capture is the best way to transport solar energy. Simply paying someone to capture CO2 (using cheap solar) and then burning fossil fuels is the cheapest option and requires the least new technology.
If SO2 injection works, we can pause warming for a paltry $20B/year while climate technologies develop. I worry that states will opt for that rather than actually solve climate change.
Oilfield CO2 injection reduces the net emissions of fuels slightly. We should research this more as a scalable way to make lower carbon fuels.
Per capita emissions peaked in 2011. From 2026 to 2052, population will only increase 8.3% before it starts falling (2026: 8.26B, 2052: 8.95). And the world has (probably) passed peak pollution and per capita CO2 emissions. Taken together, a secular decline in emissions is already baked in. Though it will happen too slowly to have a big impact on its own.
The biggest unknown in forecasting future per capita emissions is future per capita meat consumption. Will GLP-1 inhibitors have an impact here?
Recycling atoms is just a “nice to have” rather than a critical part of the green transition. That’s disappointing because geothermal and nuclear could really get a boost from this application. I hope I’m wrong, I want to see a world where supercritical water, plasma arcs, lasers, and microwaves are used to make fuels and metals from waste. Perhaps we could mine desalination salts? See also: chemical looping gassification, InEnTech, and the three companies mentioned here.
The cheap heat from nuclear and geothermal might be valuable for carbon capture8.
Cheap neutrons from nuclear are more exciting for synthesis of precious metals. Marathon Fusion recently announced a demonstration of the latter turning mercury into gold.
Besides advocating for these policies or scaling these technologies, what can you do to address climate change? Switching to an electric car, reducing meat consumption, and reducing long distance travel helps.

Heating demand isn’t a problem

I was curious about cities that will have a lot of winter heating demand. As discussed in the Energy section, this demand can be tricky to address. It comes at a time when solar resources are low and is highly concentrated during one part of the year. That means leaning on wind (which is higher during the winter) and natural gas peaker plants.

I looked for cities9 with:

Average lows below freezing in the coldest part of the year.
more than ~5 million people in the metropolitan area.

There aren’t that many!

8 cities in northern China, including Beijing; see note for more10
North America: the Northeast Megalopolis, Chicago, Toronto
Russia: Moscow, St. Petersburg
Seoul, South Korea

So heavy winter heating demand will be in the northern hemisphere11. No major cities in the southern hemisphere get cold enough. As climate change progresses, heating demand will fall faster in these cities due to polar amplification. Increasing urbanization and the urban heat island effect (which mostly increases evening and nighttime temperatures) may also lower total demand.

Almost all of these places have substantial open space, wind resources and natural gas resources nearby. South Korea is the exception, and may need to lean on offshore wind and natural gas storage to meet winter heating demand. South Korea is already building a lot of LNG storage and has a falling population, so I don’t expect this to be a huge problem.

Shale oil and shale gas resources are globally abundant

What about capturing the carbon? It turns out that shale gas wells can be used for carbon capture. For example, America’s Northeast can get natural gas from the Marcellus and Utica Shale, burn it for electricity, and inject the CO2 back in. A BOTEC suggests that each winter would require ~100 spent shale wells to capture carbon and there are ~10,000 wells in the region.

Economics of carbon taxes

Green technologies need low inflation because they involve investing in new equipment with a longer term payoff.
Governments spending billions on DAC instead of other things might be deflationary.
Domestically, it’s easier to tax concentrated producers (oil companies, chemical companies, etc.) because they are more concentrated.
Carbon taxes are an opportunity to reassess severance taxes on oil.
For smaller industries and carbon sources, it may be easier to simply DAC on their behalf rather than regulate and investigate them.
Satellites for measuring emissions are an important tax technology for reaching net zero. Though there are potential privacy issues.
The carbon tax aligns what you pay with the social cost, so merely avoiding carbon taxes is sufficient to be green in some sense. But the point is that you no longer have to worry about being green because you’re paying for the clean up.

Optimal temperature

A slightly warmer world might be optimal if we can transition slowly enough. Polar amplification means lower heating demand without counteracting cooling demand. And cooling demand is easy to meet with solar. Other benefits include shipping in the arctic, fewer cold deaths, more rain (though higher variability from droughts and floods), CO2 fertilization, and longer growing seasons12. For this to work, we would need specific mitigations for heatwaves, ocean acidification, flooding, and fires, which may not be worth it.
We eventually must ask what the optimal temperature is. That would involve a combined climate and economic simulation. Perhaps incorporating models of human behavior in response to weather. Such a simulation would be akin to Asimov’s psychohistory.
AMOC collapse is the climate tipping point I’m most worried about. Going over a 3.5°C temperature anomaly for a sustained period would be concerning.
Higher temperatures may lower wind speeds. This is very important to investigate because wind power falls non-linearly with lower wind speeds.

The geopolitics of solar

In a world with solar industrial production, which countries win? Panels are cheap but capital equipment and to a lesser extent battery storage are expensive. You want the sun to show up for consistent times at consistent angles (i.e. near the equator13) with very little cloud cover. Moderate temperatures help.

A few places stand out:

Marsa Alam, Egypt has some of the most hours of sun per year with moderate temperatures. You could even use nearby seawater to cool the panels or provide resources. It’s also in a favorable location for shipping.
Australia has an advanced economy, stable politics, and plenty of sun. But some seasonality.
Calama, Chile has lots of sun, stable politics, and near-perfect weather for solar panels. Though some seasons as well and the high elevation may be a problem.
Oman has good solar resources, few seasons, and hyperalkaline brines which are useful for carbon capture.

Direct food synthesis

Food production, particularly animal protein, takes up a lot of land. Could we directly synthesize food? I covered this in more detail here, but I think it’s feasible.

One way to reduce land use is to turn fossil fuels into food via direct chemical synthesis of oils, enzymatic synthesis of starch, or fermentation into proteins like synthetic meat. Processes like Calysta that produce animal feed from natural gas are a step in this direction. Food made in this way would require carbon capture to remain neutral.

Could food products be made from CO2 capture and green hydrogen? The economics just barely pencil out. Oils and starches are easier and use much fewer resources today so let’s look at protein. Our benchmark is soybeans, one of the cheapest protein sources around at $450/tonne14.

Carbon is ~50% of protein dry mass. 1 kg of carbon requires ~3 kg of CO2, so you need 6 kg CO2 per kg of protein. At $100/tonne DAC15, that adds $0.60/kg.
Hydrogen is only 7% of protein mass, but you need a lot of hydrogen to reduce the CO2 and nitrogen present in the protein. I used Claude to estimate hydrogen requirements for each amino acid and weighted by commonality. For each kg of protein you need 0.27 kg of hydrogen. With $1/kg hydrogen, that adds cost of $0.27/kg.
Capex for conducting the reactions adds a lot of cost. Easily a 2x increase for ammonia production and even greater for peptide synthesis. Let’s optimistically assume a 5x multiplier.

Together the costs are $4.35/kg of protein, slightly better than soy. After accounting for the huge R&D costs and the risk of single cell protein processes like Calysta overtaking the feed market, it doesn’t seem worth it.

The world isn’t starving, people want high quality calories more than cheap calories and competing with soy is silly. Synthetic protein synthesis should focus on replacements for high quality animal protein, which commands a 10-100x higher price than soy meal per gram of protein.

Potential gamechangers

Some speculative technologies that would be game changers for the energy transition

Geological hydrogen. Would make green ammonia possible much sooner and pyrolysis transport of hydrogen.

Underground kerogen gassification plus CCS. Similar to underground coal gassification. Kerogens or SMR plus point source plus pyrolysis transport equals cheap H2. kerogens + steam (or supercritical water/co2?) into cheap hydrogen into cheap protein for millennia.

Reflect orbital could reflect light down onto solar panels, reducing intermittency dramatically.

Dramatically cheaper CO2 capture. These are things beyond enhanced weathering like nuking the ocean, iron fertilization, or farming the ocean.

Laser generation of hydroxyl radicals. Perhaps we could use lasers to generate hydroxyl radicals in the atmosphere. These would scavenge for N2O, methane, and halogenated gases, shortening their lifetime. The nice part about this is that it can’t supercool the Earth like too much SRM would.

See also my important research areas list.

Splitting Infinity

Discussion about this post