It's interactive, and I'd recommend visiting it before reading this post. Check it out here: electricitybills.uk

Here are three interesting things about the data.

#1: The wholesale power cost is only one third of an electricity bill.

The wholesale price is the actual cost of buying electricity on the open market. But the average bill is triple that amount.

The remaining 2/3 of the bill is made up of three parts:

• Network costs: paying for the wires and substations of the power grid
• Generation costs: subsidising strategically important generation, like offshore wind, household solar, and firm gas.
• Miscellaneous: running a utility company customer service department, various taxes, etc.

#2: These charges are about to rise, a lot.

Network costs are about to skyrocket. Investment in the UK power grid has been stagnant for 20 years, because UK power demand has been flat for 20 years. But now, the UK urgently needs to expand the grid. Energy that used to flow through pipelines will need to flow through wires.

Contracts for Difference are the UK's flagship scheme for supporting renewables, and they will add an increasing cost to electricity bills. More than half of the contracts already allocated have yet to be activated, and the next contract allocation round is expected to be the biggest yet.

Note: CfDs do insulate consumers from high wholesale prices (during the energy crisis, CfDs reduced consumer bills), but on average they add cost.

The Capacity Market pays firm generation & demand response to be on standby, to prevent blackouts. Contracts are mostly allocated four years in advance, and the cost will ~triple to 2028.

#3: Existing costs are locked in for a long time.

The UK ran two pretty expensive green subsidies in the 2010s.

The Renewables Obligation mandates utilities to buy credits from wind and solar farms. It closed to new projects in 2017, but payments to existing projects will continue until 2037. It makes up around 10% of an average bill.

The Feed in Tariff pays households with solar panels a very tasty rate for exported energy. It closed to new applications in 2019, but payments will continue up to 2044 for some projects.

Being an early adopter of renewables has been expensive for the UK. You might argue that we should have waited an extra decade, since renewables would now be much cheaper. But the reality of learning curves means that renewables only got cheap because people built them. If everyone waits for someone else to decarbonise first, we won't get very far.

But we can learn from policy mistakes in the past. Schemes like the Feed in Tariff did not correct quickly enough when solar prices fell, leading to spiralling policy costs that were completely decoupled from market dynamics.

The future of cheap, clean power.

Let's say that we want electricity to be radically cheap in future - say £50 / MWh.

Well, network costs are already ~£70 / MWh and will increase steadily. We've missed our target, before we've even paid to generate electricity. There are only two solutions here:

• Ditch the power grid. Use local solar generation, and a lot of batteries. This works well in most of the world, but less so in northern Europe (it's not very sunny).
• Utilise the existing power grid better. Instead of expanding the grid just to service peak demand, improve our grid utilisation by "filling in the rectangle" throughout the day. (This is part of what we're working on at Axle)

The UK has achieved the fastest rate of grid decarbonisation among advanced economies. A lot of this progress occurred when renewables were still expensive, so we are stuck with a cost hangover. Luckily, renewables are getting much cheaper, so the tradeoffs in future policy are very different.

Thanks for reading, I'd love to hear your thoughts on the site. electricitybills.uk

Warmly,

Ben

Aerosols

“Aerosols” is a fancy name for “tiny particles in the air”. Aerosols are the most potent climate lever in the world, because they act near instantly.

Some aerosols warm the planet, for example, soot (‘black carbon’). Some aerosols cool the planet, for example, sulfur dioxide.

CO2 and methane have to accumulate for decades to cause the warming impact shown above. But sulfur has a short lifetime in the atmosphere, so its cooling effect is determined only by our recent emissions. It’s more like a flow rather than an accumulated stock. Any increase or decrease will quickly affect the earth’s temperature. Changing the amount of sulfur dioxide in the atmosphere is the most direct temperature control that we have over the planet.

Sulfur dioxide cools the planet a lot. In fact, it’s currently masking around 0.5 degrees of warming. Without it, we’d already be close to 2 degrees of warming.

The amount of sulfur dioxide in the atmosphere is changing for two reasons:

Reason #1: we are removing sulfur from fossil fuels

Sulfur dioxide is put into the atmosphere by burning fossil fuels.

But in 2020, the International Maritime Organisation mandated that ships must use low-sulfur fuels.

This has reduced sulfur dioxide in the atmosphere, which is un-masking existing warming. This is bad!

This single policy change by the IMO will cause 0.07 °C of temperature rise. That’s about the same as the warming caused by all of France’s emissions, in history [footnote 1]. Such is the leverage of aerosols.

Reason #2: someone might geoengineer the earth and add more sulfur.

Humans could artificially inject more sulfur into the atmosphere quite  easily. This would be risky, as we have high uncertainty about the side effects of geoengineering. But we do have certainty that it would be effective at reducing temperatures.

You don’t need an opinion on geoengineering to appreciate that it is the most potent temperature control that we have. If someone decided to deploy it, it would eclipse the importance of other climate solutions by an order of magnitude.

I think that it’s quite likely that someone deploys geoengineering. I explain why here.

Methane

Here’s the warming impact of different greenhouse gases (GHGs).

Methane doesn’t stick around as long as CO2, but the short term matters. We could hit many climate tipping points in the next 20 years.

Unfortunately, methane’s trajectory is much more worrying than CO2. Our CO2 emissions are relatively stable, even if they haven’t started falling yet. By contrast, methane is soaring, with no signs of slowing down.

Figure adapted from Spark Climate, data from NOAA Global Monitoring Lab, UNEP, and CCAC

The primary cause of this is rising fugitive emissions. Natural gas is methane. Because we use a lot of natural gas, a lot more leaks out.

Additionally, methane has some strong feedback loops.

1. As the earth gets warmer, our natural methane emissions increase. This isn’t included in climate models (even CMIP6), which means that we’re underestimating future methane emissions by quite a bit.
2. The more methane that we add, the longer that the existing methane takes to decay. If methane concentrations double, the atmospheric lifetime of all methane increases by roughly 35% (h/t Climateer).

What to do? Thankfully there are some easy methane solutions:

• Clean up fugitive emissions (methane leaks). You really can just fix up some pipes, compressors, and fracking wells, with virtually no downsides. Incredibly high climate impact per $.
• Capture methane from landfills. This ain’t hard.

Other solutions are tricky. Cows are 1/3 of methane emissions and it’s hard to feed them things that reduce their burps (although these additives do exist).

Methane removal is grossly underexplored. Methane gets naturally “removed” from the atmosphere by oxidation into CO2 (which is much less potent as a GHG). Methane removal means finding ways to accelerate this. It’s difficult, because methane is very dilute in the atmosphere - only 2 parts per million.

Solutions here are very early, and very under-resourced:

• The world is full of organisms called methanotrophs, which break methane down. Enriching soils and crops with methanotrophs would remove more methane.
• We could enhance natural methane oxidation throughout the atmosphere by adding powders like Iron Salt Aerosols.
• For more, read Spark Climate’s excellent Atmospheric Methane Removal Primer.

CO2

All solutions to reduce CO2 rely on one of three ingredients: clean energy, biomass, and carbon capture.

Almost every proposed CO2-reduction technology falls into one of these buckets.

This was the first graphic I ever made for this blog, back in 2023. Forgive the graphic design.

The main requirements of carbon capture and engineered carbon removal are a lot of energy, so our critical ingredients mostly reduce to just two: clean energy and biomass.

Therefore, we must scale:

Clean energy (grid-connected)

We are going to electrify everything, from vehicles to heating. Energy that previously flowed through pipelines will flow through the power grid. Anyone who says that electric vehicles and heat pumps won’t strain the power grid is incorrect. We must use flexible demand to match renewables, and streamline permitting for new generation.

Clean energy (off-grid)

Solar will get too cheap to connect to the power grid. A lot of future energy will be off-grid, for hydrogen production, synthetic fuel production, and for data centres.

Biomass

Engineering plants could help us scale clean fuels and carbon removal. I do not work in Biotech, but I am very excited by what’s possible. As Niko McCarty puts it:

“this is really the first generation where direct molecular observation and manipulation of living cells is possible. [...]  The cost to sequence a nucleotide of DNA fell from about $20 in 1990 to fractions of a penny today. The cost to “write”—or synthesize—a base of DNA fell by four orders of magnitude between 2000 and 2017. It’s now relatively cheap to sequence and make strands of DNA that can, in turn, be used to engineer cells.”

Conclusion

Our levers to reduce temperature are simple. We should:

• Spend more effort reducing uncertainties on aerosol effects
• Urgently deploy carrots and sticks for methane leak reduction
• Spend more effort investigating enhanced methane oxidation
• Integrate clean power into the grid - reduce permitting and use energy flexibility
• Scale off-grid clean power
• Work on solutions that increase the abundance of biomass.

Footnotes

Footnote 1 - back of the envelope on France historical emissions

(Rough calculation alert!!)

• France cumulative CO2 to date: around 40b tCO2
• Weight other GHGs at an extra 30%: 52 tCO2e total
• 52 tCO2e = roughly 6.7 ppm concentration change
• Radiative forcing including France = 5.35 * ln(420/278) = 2.207 W/m^2
• Radiative forcing w/o France = 5.35 * ln((420 - 6.7)/278) = 2.122 W/m^2
• Forcing caused by France: 0.085 W/m^2
• Assume climate sensitivity of 0.8°C per W/m^2: 0.068 °C.
• So warming caused by France’s cumulative historical emissions: ~0.07 °C.

Someone willing to pay a premium to build solar+batteries, geothermal, and nuclear, in order to bring them down the cost curve and make them cheaper for everyone.

That is exactly what AI data centres are.

For the first time ever, securing clean 24/7 power is an urgent strategic priority for the most powerful organisations in the world:

1. Tech giants, who have made strong commitments to using clean energy, and are willing to pay through the nose for it
2. Governments who are determined to win the AI race. The only people who can fix the thicket of regulatory barriers to building energy projects quickly.

Many are worried that AI will increase emissions. This is true in the short term - emissions will rise for a few years. But it misses the bigger picture.

Clean energy technologies are famous for their learning rates - they get cheaper as you build more of them [footnote 1]. Data centres create price-insensitive clean power demand that will foot the bill of expensive early projects.

The perfect customer

Tech companies have the highest revenue per tonnes of CO2 emitted of any type of company. They are extremely cash-rich, and they are willing to pay extra for clean energy.

Tech giants upend the energy industry in four big ways:

1. Energy generation is the cheap part of an AI data centre.

Modern GPUs are pricey as heck. Eight Nvidia H100s in a single DGX box will run you $250,000.

Relative to compute, energy generation is cheap.

AI turns the electricity sector on its head. It places a premium on speed and low emissions, rather than cost.

Of course, tech companies will not wait a decade to build a nuclear plant before switching on their data centre. More on that later.

2. Urgency

The stakes could not be higher for big tech companies. If you are Google, you face the loss of your entire business within the next thousand days. Winning the AI race supersedes all previous priorities.

And let’s be clear, we will need a lot of power. A single H100 GPU uses the same electricity as two UK households. Meta is installing hundreds of thousands per month. Previously, data centres were chip-constrained, but now GPU production is ramping up. We are becoming energy-constrained. You don’t have to take my word for it - just follow the contracts.

3. Risk appetite

Traditionally, energy generation is financed by utilities, banks, and infrastructure funds. They build low-risk projects, with mature technology.

Tech companies are a bit more fun. They will yolo $hundreds-of-millions into fusion projects, linear generators, or whatever zany tech they like. They need power tomorrow, but they’re also willing to make big-if-true long term bets.

I don’t know if small nuclear reactors and advanced geothermal will ever get cost-competitive for widespread adoption. But I do know that until this year, they’ve never had such a well-suited customer.

There’s ideology here too. Tech companies are… techno-optimist. They exert their agency on the world, and they will attempt to venmo new technologies into existence. Sometimes this doesn't work, but sometimes it does.

4. Tech companies will become their own utilities.

Getting a new connection to the power grid takes years, often decades. Tech giants cannot wait that long.

In the short term, data centres will hoover up grid connections where they already exist. They will re-start Three Mile Island, or buy a bitcoin farm and convert it. If it is more profitable to use the grid connection of an aluminium smelter for AI, the aluminium smelter will die.

But existing grid connections are limited. In the medium term, tech companies will become their own utilities. They will build their own off-grid electricity generation at the sites of their data centres. This also saves money; about half the cost of energy is transporting it through the grid.

This is great news for solar and batteries, which can be thrown up in a field/desert in months instead of years.

AI training clusters are also much more locationally flexible than traditional servers, because they don’t need low-latency internet. Running a fibre line to the desert is a fraction of the cost of a data centre.

The president called, he likes your data centre.

The US and China will prioritise frontier AI development, because it will define the economic and military superpowers of the 21st century. (It is yet to be determined whether Europe will do the same.)

Governments haven’t yet realised what prioritising AI entails, but it will quickly become clear that it means making it easy to build new energy projects.

This is important because nuclear and geothermal are both bottlenecked by a labyrinth of permitting, land rights, and regulatory approval. For decades, western governments have had no urgent motivation to streamline these processes - until now.

There are a lot of barriers to scaling clean power that can only be addressed by policy. AI could be the decisive factor that spurs governments into action.

Will AI really only use clean power?

No. In the short term we will burn more gas to meet extra demand. This year, both Microsoft and Google announced that their emissions had spiked, driven mostly by AI energy requirements that materialised out of nowhere.

Currently, tech companies’ clean energy supply lags their demand. But they are serious about their climate commitments, and they are training their cash bazookas on the problem. As a reminder, tech companies remain the biggest corporate buyers of renewable power in the world. Google, Microsoft and Meta have all committed to carbon neutrality by 2030 - and they plan to get there.

It is clear that tech companies will not sacrifice speed, but they will sacrifice cost - as demonstrated by the steady stream of clean energy purchase announcements. You do not pay to restart the Three Mile Island nuclear plant as a cost-effective way to source energy.

Even if their clean energy supply lags their demand temporarily, tech firms will pay to match their demand with clean generation purchases (often at an hourly level). That means that they have to purchase a lot of future clean energy contracts. Buying these contracts means that they pay to build new clean generation - whether directly or indirectly.

Not AI or climate - both.

It might seem naïve to think that AI and climate are complementary. But money has already started pouring into clean-firm power as a direct consequence of LLMs. Data centres will make clean energy cheaper just by building more of it.

There’s also the application of AI to climate solutions, which could be far more consequential than data centre energy use. That’s another blog post, but AI’s impact on science will undoubtedly be profound - from foundational chemistry and biotech, to materials discovery and catalysis. We stand at the brink of science 2.0, and it would be remarkable if this does not positively impact the energy industry within the next ten years.

One way or another, hundreds of millions of GPUs will crackle with the world’s data. Vast pools of capital will be brought to bear on the creation of thinking machines, and the opportunity for the energy industry is immense.

Done correctly, AI data centres will accelerate the economic and technological frontiers of clean-firm power. They have already started.

Thanks to Steve Newman, Rosie Campbell, William Bubenicek, and Rob Tracinski for their thoughtful comments on this piece. All opinions are mine, not theirs. Also, a shoutout to Nan Ransohoff for the original tweet that inspired this blog:

Follow me on Twitter/X, and read my other writings:

• The Big Guide to Batteries
• The Big Guide to Sustainable Aviation Fuel
• Actually, We Can Deploy Energy Infrastructure Very Quickly

Footnotes

Footnote 1

It’s still not clear how cheap nuclear fission will get as we build more of it. But it is true that all the cheapest nuclear plants in the world have been built as part of standardised, mass rollouts.
Small Modular Reactors have had little success so far, but they are only just encountering their dream customer: the AI data centre. If they are ever going to work, now is their chance.

1. I will be in San Francisco in mid October. Let me know who I should meet!
2. Over one hundred thousand people visited my blog during September, which feels very bonkers. People seemed to enjoy that solar is getting cheap.
3. I would love to chat to people who are thinking about: perovskite solar, bio-solutions for accelerating methane oxidation, geothermal, ML for chemistry in general, and sunlight reflection/geoengineering. Future blogs incoming!
4. At Axle Energy we recently raised a $9M seed round from Accel and are hiring the best people. Drop us a note if you’re interested in building the future of the grid. I’ve also been doing some writing at Axle: on UK power markets and the insane potential of residential energy flexibility.
5. I’m finally dragging myself into the 2020s and getting to grips with LLMs/transformers. Here is my artists impression of a basic transformer.
6. I might post on X/Twitter more.
7. This year I’ve been very lucky to be a Roots of Progress fellow and a Polaris fellow. In 2024 I've felt more grateful than ever to be around people whose "it's cool to care” meter is off the charts.
8. I printed up a booklet containing a few of my climate tech guides. Let me know if you'd like a copy and I’ll mail one.

As always, just reply to this email to say hi. I’d love to hear from you.

Warmly,

Ben

When a nation decisively commits to building energy infrastructure, it can build fast.

We perceive that “energy can’t change quickly” because of recency bias. Most people in The West today have not experienced a major energy transition in their lifetimes. We think it's impossible to build new energy infrastructure quickly, because we haven’t recently.

People are particularly concerned about domestic energy transitions - how fast can you deploy things that touch households? Electrification relies on breakneck deployment of heat pumps, electric vehicle charging, and the associated upgrades to the local distribution grid (which are significant). Luckily, there is a rich history of domestic energy transitions going very rapidly.

Here are four examples. We’ll pick out some common features after.

UK, 1960s. Town gas to natural gas

Up to 1967, the only gas used in Britain was “manufactured” from coal. People called it town gas, because each city had a local gas works, which heated up coal to produce gas. It was dirty, expensive, and dangerous.

Then, Britain discovered 'natural gas' in the North Sea. Natural gas was much safer and cleaner than coal gas - but it would require new burners in gas appliances (heaters, cookers and lamps). Making the change would be disruptive for all gas users in Britain, but the government of the day fully committed to it.

In just ten years, 14 million homes were converted, and 40 million appliances were replaced. Each home cost £1,700 (inflation adjusted) to convert on average.

Thirteen dedicated training schools were set up across the country, and a national workforce was trained up to replace appliances and infrastructure. It was hailed as “the greatest peacetime operation in this nation’s history” by the chairman of the British Gas Corporation (then state-owned).

They ran showrooms in every town and city, which dealt with accounts and service queries as well as selling cookers, heaters and smaller appliances. They offered rental and hire purchase terms to those who could not buy outright, making new equipment available to a wide range of householders for the first time. - Resilience.org & Rapid Transition Alliance

Many people protested moving from coal to new-fangled gas. Their arguments sound very similar to those who resist electrification today.

In an entertaining transcription of a parlimentary debate from 1968, a young Margaret Thatcher vehemently argues against natural gas, resists the writing off of coal gasification plants, and telling of the horror of workers coming to upgrade town gas appliances.

We looked at the gas fires in the bedrooms and his face grew graver and graver. 'I'm afraid', he advised me, rather in the manner of a doctor diagnosing smallpox, 'that we shall have to take them away’

US, 1930s. Rural Electrification Act

In 1936, only 10% of US rural homes had access to electricity (the national average was 65%).

Here’s a map with states scaled to the proportion of their population with access to centrally-produced electricity

In 1936, FDR signed the ambitious Rural Electrification Act (REA), to connect all of America to the grid. It was extraordinarily successful. The proportion of rural homes with electricity jumped from 10% to 90% in just 14 years.

New transmission lines were planned, built and owned by local cooperatives, formed of rural residents. The REA encouraged these cooperatives to form by providing low interest government-backed loans. Cooperatives were also provided with technical assistance, qualified engineers, and standardised equipment. The REA standardised poles, power transformers, and electrical wires. Before the programme started, transmission lines cost up to $2,000 per mile. By 1939 it was $825 per mile.

Indonesia, 2007. Cooking fuels

In 2007, Indonesia launched one of the most ambitious cooking fuel conversion programs in history. At the time, kerosene was the primary fuel for domestic cooking, and state subsidies for kerosene imports cost $5 billion annually. LPG (Liquefied Petroleum Gas) was chosen as a cheaper alternative, that also emits roughly half the CO2 of kerosene when cooking.

In 2007, a nationwide conversion programme was launched, and within five years, almost all kerosene-burning households had converted to LPG. Many users of wood-for-cooking also switched.

The program distributed 57 million LPG starter kits. Each contained a 3kg LPG cylinder, stove, and regulator. The rollout was intense; at its peak, 180,000 packages were being distributed daily.

By 2015, the government had saved an estimated $11.8 billion in reduced subsidies, indoor air quality had dramatically improved, and millions of tonnes of CO2 were avoided. All in all, the conversion program reduced domestic kerosene use by 92% in less than ten years.

Nonetheless, LPG remains an expensive and imported fossil fuel. Indonesia is now drawing up plans for a large-scale deployment of electrified rice cookers and induction stoves.

France, 1970s. Nuclear power

The prominence of nuclear power in France is well known. But the speed that they deployed their nuclear fleet is often missed.

When the 1973 oil crisis hit, France realised that it was entirely dependent on imported oil. This 1974 government ad is extremely entertaining. It concludes "In France, we do not have oil, but we have ideas."

In 1974 France unveiled the Messmer Plan - which called for the fastest buildout of nuclear power in history. It worked astonishingly well.

• In 1974, France had just 8 operating nuclear reactors, and only one was French-designed.
• By 1990, France had built 56 new reactors.
• Between 1980-85, France was connecting a new reactor to the grid every 2-3 months.
• Between 1970 and 1990, France took nuclear power generation from 4% of electricity generation to over 70%.

Their playbook was very simple. Firstly, they standardised heavily and built the same reactor design repeatedly. Secondly, they streamlined regulation to make rapid construction possible. Finally, clear government support of the project made it easy for EDF (the company building the reactors) to raise cheap capital, which was crucial for the high upfront costs of large nuclear.

(These three points remain all that is necessary to foster a thriving nuclear industry today.)

The list goes on

• Between 1970 and 1990, oil use for heating in Sweden plummeted from 75% to 25%. It was replaced by district heating, heat pumps, and resistive heating.
• Between 1975 and 1985, Brazil went from making almost no ethanol, to replacing about 60% of gasoline consumption.
• Between 1950 and 1970, the UK quadrupled electricity supply from 50 TWh to 220 TWh.
• Between 2005 and 2010, the Cuban energy revolution eliminated blackouts, by distributing millions of efficient appliances within months, and installing decentralised electricity generation. They saved an estimated $1bn a year in oil imports.
• Between 2018 and right now, Vietnam has gone from 0 GW of installed solar to 10 GW.

Exceptions or the rule?

It's conspicuous that these examples come from (1) rich countries pre-1990, or (2) poorer countries that were getting richer.

It’s also true that most energy transitions are slower than these, so you might reasonably ask if these examples are the exception rather than the rule. In my opinion that’s the wrong question to ask, because these examples became exceptions based on how hard they tried. It is not mere luck whether your infrastructure project succeeds; you change the outcome by how you set it up.

I am not saying that all energy transitions are fast. But I am saying that it’s possible for them to be fast, if we create the right circumstances. So what are those circumstances?

Pattern-matching successful projects

Some clear patterns emerge from the examples above. In all cases:

• The technology was well established. Technical risk was not a concern.
• Policy was extraordinarily strong. Even though the state did not usually install the new infrastructure, they created a cast iron framework of incentives, loans, and support.
• A huge workforce was needed, and training up tens of thousands of people was a core requirement (and economic benefit)
• In most cases, the state did not own the final infrastructure. Local communities did.
• Deployed technology was ruthlessly standardised.
• National priorities were placed above NIMBYism, but locals often benefited from the technology too.

The West’s fundamental disadvantage in the energy transition is that it has forgotten what bold industrial development looks like, because it happened so long ago. The good news is that the necessary remedies, above, are very simple.

There’s one other thing. A lot of historical energy infrastructure was large, centralised projects. And we’ll certainly need more of them. But the core technologies of electrification are decentralised. We will not execute large projects in series, we will roll out millions of EVs, heat pumps, solar panels and batteries in parallel. We’ve already started.

If all you extrapolate is the past two decades of energy deployment in rich countries, the energy transition looks pretty tough. But electrification’s core technologies (batteries/EVs, solar, and wind) are just starting to hit the right price levels for mass deployment.

The technology is arriving. The question is: will we be bold enough to deploy it?

-

Follow me on Twitter/X, and read my other writings:

• The Big Guide to Batteries
• The Big Guide to Sustainable Aviation Fuel

This post explains why we'll keep building more, and what we'll do with the excess.

There's three parts to explain:

• Why solar will get cheap
• Why solar will saturate the power grid
• What we will do with the excess.

Solar power is going to get very cheap.

The solar panel is wizardry manifest. It literally prints energy from free shit that falls out of the sky.

A very flat rock that whispers to the light of heaven and coaxes it into bottled lightning. Without even moving.

Solar PV is the only way to make electricity with no moving parts. This makes it much easier to mass manufacture than other energy sources.

If you can make something that resembles a flat silicon pebble, manufacturers will look at it and say “hell yeah”. Humans are very good at driving down the cost of flat silicon pebbles - think semiconductors, windows, LEDs and touchscreens.

That means that solar has a tremendous learning rate. The more solar you make, the cheaper it gets.

This means that solar deployment is accelerating at breathtaking speed. Most of the world's solar power was installed in the past 30 months. In fact, China installed more solar in 2023 than the US has installed in history.

In the UK in 2024, I can go online and buy a solar panel with the same dimensions as a fence panel, for only double the cost. In five years, the cost of solar will have halved again.

Solar will saturate the power grid.

Today, most solar gets connected to the electricity grid. But the more that we connect, the harder it is to hook up new solar. This is for two reasons.

Saturation reason #1: Solar competes with solar (“cannibalisation”)

Solar panels all generate power at the same time of day: when the sun’s out.

Adding more solar drives down electricity prices when it’s sunny, because energy supply in sunny hours increases. This means that all solar panels get paid less $$$ per unit of energy produced. If you add a lot of solar, at some point it becomes uneconomical to build the next marginal panel. You are just adding electricity at times that nobody needs it (although we are still very far from this point).

When there’s enough surplus energy supply from renewables, prices can even go negative. That's great news for consumers - who get paid to use energy - but it's bad news for renewable generators. Too much negative pricing will mean that new renewables don't get built, because they won't break even.

Saturation reason #2: Grid capacity

A surprising thing about the power grid is that it’s very difficult to connect up new electricity generation. Even if solar farms are quick to build, they are not quick to integrate with the grid.

In the UK, the average wait for a grid connection is more than 5 years, and over 40% of queuing projects have a connection date beyond 2030.

In the US, the size of the grid connection queue has increased over 7x in the past decade, and 95% of the queue is solar, wind and batteries. The capacity in the queue is now more than twice the installed capacity of all existing U.S. power plants.

What to do with all the solar?

Solar will saturate the power grid, but that doesn’t mean that we’ll stop building it. It just means that we’ll use it off-grid.

We can opt to either store the magic lump for later in the day (#1), or use it at the time of production (#2).

#1: Solar and storage

Free electricity in the middle of the day, followed by high prices a few hours later? That is a dream day at the office for the lithium-ion battery. The more solar we install, the more we will be deluged by batteries - bathing in volatility like pigs in mud. If you are not yet sold on the idea that batteries will eat the world, you should read these posts. Like solar, the costs of lithium-ion batteries are plummeting every year.

The problem: we won’t have enough grid capacity to dump solar energy into batteries through the high voltage transmission system. We’ve already established that when the sun is out, we’ll blow through the limits of what the grid can handle.

This means that solar and batteries will have to zippity-zap each other without touching the transmission grid. Hence: solar and storage will have to be locally installed. This means we’ll get co-located utility solar and storage, homes with personal batteries for their solar, and new batteries springing up next to existing solar generation.

To be clear, a lot of the electricity transferred between solar farms and big batteries will still go through parts of the local distribution grid. And a lot of this solar will be under-gridded, rather than entirely off-grid. But it would be foolish to rely on new transmission lines to absorb our entire future solar capacity. Transmission lines are expensive and slow to build; batteries are cheap and quick to deploy.

#2: Solar for intermittent processes

Using batteries to backup solar is intuitive. Less obvious is the idea of surrendering to the sun, and only using energy during the magic lump.

If you only use energy for one quarter of the day, your solar system gets much cheaper, because you don’t need a battery. This is actually the cheapest electricity in the world.

These prices are already quite cheap, but we’ve seen nothing yet. The cost of solar energy in a sunny place is trending towards virtually-free.

This is solar’s opportunity to not just displace electricity supply, but also primary energy supply. Rather than simply supplying energy in the form it’s consumed (electricity), intermittent solar is so fricking cheap that it could manipulate atoms into fuels for subsequent consumption.

We’re talking about using solar to create synthetic kerosene for planes, clean ammonia for fertiliser, clean methanol for shipping, and maybe even synthetic natural gas for general purpose use.

These synthetic and ‘green’ fuels all rely on green hydrogen as a base ingredient. Green hydrogen is extraordinarily expensive to produce, and the only cost-competitive way to make it is off-grid solar.

Off-grid solar is not just one way that synthetic fuels like Sustainable Aviation Fuel can be made - it is the only way that they can be cost competitively produced.

And it’s not just fuel production. Any process that can be tuned to run cost effectively at 25% utilisation is up for grabs. Direct Air Capture, industrial heating, electrochemical-everything; all are fair game.

Taking solar off the grid also has a few other major cost advantages. If you are ripping solar straight into a DC application, you can skip the costs and efficiency losses of inverting that power into AC. If you lose most of the balance of plant, power electronics, and the paperwork of a grid connection, you’re getting really cheap and fast.

Conclusion

The question “can we build processes that use free energy one quarter of the time” is one of the most important and unexplored questions in the future of energy. Make no mistake: it is hard to build these processes, because quartering your output dramatically worsens the payback period for capital intensive process plants. But whilst this is difficult to do, frankly we haven’t tried. Conventional chemical engineering wisdom optimises for expensive inputs, continuous output, and high efficiency. The design space for low utilisation, free energy, is ripe for exploration.

The fact remains: few people optimise for intermittent solar, and intermittent solar will change the world.

-

Thanks to Niko, Sean and Rob for their extremely thoughtful comments on this piece.

The origination of many ideas in this post come from Casey Handmer, who deserves great credit for popularising them. I’ve been unable to recreate numbers as optimistic as his, but we directionally agree on a lot.

Follow me on Twitter/X, and read my other writings:

• The Big Guide to Batteries
• The Big Guide to Sustainable Aviation Fuel

The world's existing fleet of planes is huge, and can only run on jet fuel.

Therefore, we need a drop-in fuel replacement that doesn’t cause warming. This is called Sustainable Aviation Fuel (SAF).

This piece is a concentrated onboarding guide to SAF. We'll walk through the different methods for making SAF, and explain what's tricky and exciting. No chemistry knowledge is assumed!

Firstly, we need to know our enemy. What is SAF trying to replace?

Jet fuel is just kerosene, which is a mix of medium-length hydrocarbons. But wtf does that even mean?

The aim of SAF is to make a fuel that’s chemically identical to kerosene. We want a drop-in replacement.

That's confusing, because it means that burning SAF emits just as much CO2 as burning regular kerosene. The difference is where the carbon in the fuel originates from.

CO2 is not the biggest problem here

There’s another twist.

Aviation’s biggest warming effect does not come from CO2. It comes from contrails.

Here’s the warming breakdown:

The crazy part is that we know very little about contrails, and the warming that they cause. It’s an active research area, but our uncertainty is massive.

One big benefit of SAF is that it causes less contrails. This is because fossil kerosene contains compounds called aromatics, which produce soot when burned. That soot causes ice to form in the air, because it provides a place for water vapour to crystallise.

SAF contains less aromatics, which means that it burns cleaner and forms less contrails [footnote 1].

How to chef up some SAF

There are two types of SAF: Bio-SAF and e-SAF. This is because there are two ways to source carbon: using a plant, or using a machine.

Unfortunately, we need a lot of plants to make Bio-SAF, or a lot of energy to make e-SAF. This is probably the most important point in the guide:

This is why only 0.2% of all jet fuel today is SAF.

Biomass supply:

The world has a limited supply of sustainable biomass. If you are in the business of producing Bio-SAF, your quest is roaming the earth, desperately searching for enough plants/oils/trees to feed the loud sucking sound emanating from your bio-SAF production plant.

If we replaced all of today’s jet fuel with Bio-SAF, we’d need to use all of our energy crops and agricultural wastes to make it. And that biomass supply is also needed for other things.

Energy supply:

Making e-SAF requires an extraordinarily large amount of energy. It’s like if your mum was an amount of energy.

If we replaced all of today’s jet fuel with e-SAF, we’d need the entire energy supply of a global superpower.

To be clear: e-SAF only works if there is a wartime deployment of clean electricity generation, obliterating any speed of energy infrastructure deployment the west has seen over the past couple of decades.

Without that, SAF displaces green electricity from better climate solutions like electrification, which save more CO2 per kWh. That is greenwashing, and it is extremely cringe.

Meet the pathways

There are numerous ways to produce SAF. Here are the main production pathways:

We’re going to walk through each method one by one. First: Bio-SAF.

Most Bio-SAF today is made using the HEFA process. HEFA takes waste cooking oil and turns it into fuels like kerosene and diesel.

Ngl that’s kinda epic? We literally collect used oil from McDonalds fryers, cart it over to refineries, then turn it into SAF.

The critical inputs to the HEFA process are oils/fats, and hydrogen. Yes, bio-SAF needs hydrogen too!

HEFA uses traditional refining and chemical engineering processes, so it’s well suited to oil companies and existing refineries.

The problem with HEFA

Unfortunately, HEFA has a mahoosive problem: we don’t have enough waste cooking oil or waste fats.

In Europe, we import 80% of our waste cooking oil from abroad. Most of it comes from China, who will soon start using that oil to make its own SAF, leaving little for export.

What about growing new oil?

Instead of relying on waste cooking oil, what if we grew new vegetable oil just for SAF?

This is a non-starter. To produce enough rapeseed oil to fuel all UK aviation, we’d need to dedicate 68% of the UK’s agricultural land to growing it.

Fortunately, HEFA is not the only way to make Bio-SAF.

Alcohol-to-Jet

Alcohol-to-Jet is a newer pathway for making Bio-SAF, as well as a British stag party tradition. It's a simple idea:

1. Make alcohols (for example, ethanol) from plants.
2. Turn your ethanol into SAF.

There are two big advantages of Alcohol-to-Jet over HEFA.

#1. Alcohol-to-Jet can use more types of biomass. Anything that can be converted into an alcohol (which is most plants, if you try hard enough) can make SAF.

#2. Alcohol-to-jet can be set up to produce a high proportion of jet fuel, with very few other by-products. In the business, this is called having a high kerosene selectivity, and it's important to understand.

All SAF pathways produce a range of hydrocarbons. SAF production with high selectivity will produce chains that are mostly of the right length for jet fuel.

Alright - let’s move on to our second type of SAF!

e-SAF is pretty magical when you stop for thunkin’. We use air, water, and energy to produce jet fuel on demand.

It sounds like science fiction, but it’s entirely possible. The reason that we don’t make e-SAF commercially today is that it uses so much goddamn energy.

e-SAF is literally the law.

Policymakers realise that planes powered entirely by biofuels would be a disaster for land use and food production. So in addition to mandating minimum blends of SAF in jet fuel, they are mandating minimum blends of e-SAF too.

The Fischer-Tropsch process

The most mature way to make e-SAF is the Fischer-Tropsch process.

Fischer-Tropsch has been around for ages, but not because of climate change.

Fischer-Tropsch works by making artificial crude oil from something called syngas.

Syngas

Syngas is made of hydrogen and carbon monoxide.

Traditionally it was produced from coal (cringe) or gas (cringe). For SAF, we can make syngas from hydrogen and CO2. We do this with the Reverse Water Gas Shift reaction, which incredibly is not part of a car, but a nifty way to re-shuffle hydrogen and carbon atoms.

Alright, let's look at some numbers.

Ingredient-economics of e-SAF.

You only need two ingredients to make e-SAF: hydrogen and CO2.

But making hydrogen and capturing CO2 both need extraordinary amounts of energy.

You need about 50 kWh to make a kilo of green hydrogen, and about 2.5 kWh to capture a kilo of CO2 from the air. (I wrote about the energy requirements of hydrogen here, and the energy requirements of Direct Air Capture here)

That means,

Let’s be clear, this is a lot of energy.

It totals out around 100 kWh per gallon of SAF, and we use around 100 billion gallons of jet fuel per year. If we were to replace all jet fuel with e-SAF, we’d need 10,000 TWh to make it.

Unit costs

Let’s break down the operating costs of making these e-SAF ingredients.

Let's say that we can pull CO2 out of the sky for $200 / ton, and build solar for $15 / MWh.

I’ll cover how we can minimise the cost of SAF in the “what’s exciting” section below.

Methanol-to-Jet

Methanol-to-Jet is a newer pathway that is a contender to Fischer-Tropsch. It needs the same ingredients as Fischer-Tropsch (CO2 and hydrogen), so the ingredient-economics that we covered above still apply here.

It has a few subtle advantages, which I'll leave as an optional dropdown:

Which SAF technologies are exciting?

The answer to this question is actually very simple.

SAF is bottlenecked by the availability of biomass and cheap energy. Therefore,

For e-SAF, that means abundant, off-grid, energy.

One thing is clear: using electricity from the grid to make e-SAF is too expensive.

Additionally,

1. If you’re taking electricity from the grid to make SAF, you’re probably stealing electricity from more potent ways to reduce emissions (ie, electrification).
2. The necessary grid upgrades would be ludicrous. We’re talking about taking the entire energy consumption of dragging humans through the sky, and forcing it to pass through pylons beforehand. Pylons that are already under strain, and extraordinarily costly and slow to build.

Cool, so we’ll build off-grid solar just for SAF, or use SAF to mop up excess power when the grid is ‘full’. That’s great, but it means we can only run our SAF plant 20-40% of the time. Synfuels plants are expensive, so it sucks to have them dormant 70% of the day.

Thus, the critical advancements to make e-SAF are:

1. Build cheap, abundant, off-grid renewables (already on track)
2. Make synfuels plants really really cheap - even at the expense of efficiency.

For Bio-SAF, we need abundant biomass.

The good news is that we have plenty of plants that grow very fast, like seaweed and bamboo.

The bad news is that growing them from scratch and converting them to SAF is expensive and requires using the land or ocean (& oceans require permits). Nonetheless, I’ve yet to see a good reason why we can’t grow new biomass in a much cheaper and automated way.

It's worth noting that in the past decade, most oil majors took a swing at growing algae for biofuels, but all their programmes ran into difficulties and eventually shut down.

I’m still optimistic, because our recent advancements in biotech are astounding. We have capabilities today that we could only dream of in 2010. Whether it’s cheap cellulase enzymes that make cellulosic ethanol economical, or simply plants-that-grow-fast, we should not ignore the potential of biotech to help us out here.

Conclusions

SAF is the only realistic way to decarbonise long-haul flying, and producing it is entirely possible.

But without abundant energy or biomass, making SAF will actively increase emissions. It will steal energy or land from more effective climate solutions.

The question "will SAF work?" is just a wrapper for the question "how fast do you believe we can build things?". Can we hurl our infrastructure into the future fast enough to catch ourselves?

To bet on SAF is to bet on bold & targeted energy policy, a revival of industrial dynamism, and the fastest buildout of energy infrastructure in living memory.

Would you take that bet? I'll leave you to decide.

Thanks!

Thanks to Finlay Asher, Dusita de Hoop, Valentin Batteiger and Oliver Booth for their time / help with this piece. Opinions and mistakes in this piece are entirely mine, not theirs.

Footnotes

Ethanol and methanol always sounded confusing and boring to me. But they're actually pretty interesting.

This post will explain everything you need to know, from scratch.

Ethanol and methanol are used for very different things.

We make ethanol and methanol in roughly equal quantities. As a rule of thumb: every year we make around:

• 100 million tonnes of hydrogen
• 100 million tonnes of ethanol
• 100 million tonnes of methanol

The basics

Most ethanol today is used as biofuel for cars. When people talk about ethanol for climate, they usually just mean making more biofuel. This would be used for cars (burned directly) or planes (as an ingredient in SAF).

Most methanol today is used for chemicals, though roughly a third is used in cars. When people talk about methanol for decarbonisation, they usually mean using it as a fuel. This would be for ships (used directly) or planes (as an ingredient in SAF).

Hol' up, these guys still make CO2 when u burn them?

Yeah. Ethanol and methanol both contain carbon, so they emit CO2 when burned. Ethanol produces slightly less CO2 than gasoline, and methanol produces slightly less CO2 than shipping fuel. But not by much.

So what makes them any better than fossil fuels? The difference is where the carbon comes from. If the carbon in your fuel comes from the air, not the ground, you’re not adding extra CO2 to the atmosphere.

In theory, this means that fossil fuels add new CO2 to the atmosphere, whilst biofuels “recycle” what’s already there. In practise, the lifecycle emissions of biofuels are more dubious, but we’ll cover that below.

Ethanol for cars, methanol for ships.

Both methanol and ethanol have lower energy densities than fossil fuels, but methanol has the worst.

This is one natural reason why ethanol suits cars and methanol suits ships.

Additionally:

• Methanol is nastier. It’s a boring level of toxicity - nothing dramatic, but not friendly to humans.
• Methanol storage is already common in ports, because of its extensive use in the chemical industry
• Methanol has historically been a tad cheaper than ethanol, though not by much.

Ok, now let’s dive into them one by one.

Ethanol

Ethanol is the most widely used biofuel in the world. Production and usage are dominated by two countries:

The US and Brazil are raging ethanol junkies. They use almost all for cars and motorbikes.

Brazil uses so much ethanol that no Brazilian cars run on pure gasoline. There is a minimum blend of 27% ethanol in all fuel, and 80% of light vehicles are special “Flexfuel” varieties, which can run on up to 100% ethanol.

Ethanol production

We make ethanol via fermentation. It’s literally the same thing as making beer. “Alcohol” (the drink, not the chemistry term), actually just means ethanol.

Fermentation is pretty simple. Just add sugar to some lukewarm, yeasty water.

Producing ethanol also produces CO2. But notably, the CO2 that gets emitted is 100% pure. That’s great news, because it means that it’s easy to capture. By contrast, most industrial processes emit CO2 that’s much more highly diluted.

In fact, capturing CO2 from ethanol production is so easy that ethanol production is the biggest source of purified CO2 in the US, and the second biggest source in Europe. This is where we source the CO2 that we use in fizzy drinks, food production, fire extinguishers, etc.

Here’s an ethanol plant in Minnesota:

First-generation ethanol

The US makes its ethanol from corn, whilst Brazil makes its ethanol from sugarcane. Both are simple processes that are mature and well-established. Corn ethanol and sugarcane ethanol are known as first-generation biofuels.

Corn and sugar are easy to turn into ethanol, but they’re also… edible. That means that we’re using agricultural land that could otherwise be used for producing food for humans. It would be great if we could make ethanol without competing with food production…

Second-generation ethanol

An alternative source of ethanol is cellulose, the stringy and fibrous bits of a plant. You can get cellulose from grass, trees, and loads of different types of plants. It’s everywhere, but it’s harder to make ethanol from.

Second-generation biofuels mean biofuels made from non-food crops.

Cellulosic ethanol is the largest potential source of biofuel in the future, because about half of all plant biomass is cellulose. But whilst sugar and corn are easy to break down into ethanol, cellulose is a lot more complicated. It’s not yet clear that we’ll be able to do it cost effectively.

What about ethanol from waste?

People love to talk about making ethanol from food and ag waste. That’s great, but we don’t do much of it at the moment. In the US, more than 95% of ethanol comes from virgin energy crops.

We should do more, but there's only so much food waste.

Ethanol in planes

One exciting future use for ethanol is in jet fuel.

Scaling ethanol to replace gasoline is kinda silly, because we already have an easy way to decarbonise cars: EVs. By contrast, we don’t have an easy way to decarbonise planes.

To be clear, we won’t fuel planes with ethanol directly. Instead, we’ll use an Alcohol-to-Jet process that turns ethanol into synthetic kerosene - a drop-in replacement for fossil jet fuel. More on that in my next piece on SAF.

Ethanol has some haters.

There are two controversial things about ethanol. They mostly pertain to its use today.

Emissions reduction

Most ethanol today is used to displace gasoline, and the emissions savings from doing this are highly contentious. This is because there are plenty of GHG emissions in the life cycle of producing ethanol, for example: tractors, fertilisers, and uncaptured CO2 during fermentation.

Ethanol’s lifecycle emissions are famously hard to quantify, and some people even suggest that ethanol is worse than gasoline. Studies vary a lot, but most that I’ve found suggest US corn ethanol reduces emissions by roughly 40-50% versus gasoline.

Land use

Today, most ethanol production consumes crops that use vast areas of land. Land that could otherwise be used for food.

Corn is by far the biggest crop produced in the US, and a hefty chunk of it goes to ethanol production.

Ethanol production is also highly subsidised in the US. From an energy lens, the US ethanol programme is bloated and unfit for purpose. It sticks around because it is primarily a political decision to subsidise US farmers.

Alright! Time to switch gears and dive into methanol.

Methanol

Today, most methanol is used in the chemicals industry. Methanol is one of the four basic building blocks that is used to make most other complex chemicals (the other ones are ethylene, propylene and ammonia).

Most talk of methanol for decarbonisation focuses on methanol as a fuel - usually for shipping and aviation.

Methanol for shipping

Methanol is considered one of the main contenders to decarbonise shipping. Other approaches include ammonia, carbon-capture-on-ships, and (surprisingly) big, heavy batteries.

Methanol’s primary advantage is that it doesn’t require huge ship modifications. Because it’s already an internationally shipped commodity, ports have existing methanol storage (over 100 major ports), and are used to handling it.

That doesn’t mean that methanol for shipping is all roses. We’ll cover the energy requirements below.

Can methanol do planes too?

You betcha - methanol-to-jet is a hot topic. BUT methanol is specifically excluded from the list of approved SAF production pathways. That means it’s nearly impossible to sell commercially atm.

More on that in my next piece on SAF.

Methanol production

Methanol’s chemical symbol is CH3OH. The important part is that it contains a “C” for carbon. Ie, to make methanol, we need a source of carbon.

Today, that carbon comes from fossil fuels. In China, it’s coal; in the rest of the world, natural gas. Before we think about making more methanol, we first need to decarbonise the production of our existing methanol.

There are two ways to make clean methanol - and both have different sources of carbon:

• Biomethanol uses carbon from biomass (eg plants or household waste).
• E-methanol uses carbon from CO2 captured from industrial processes, or pulled straight from the atmosphere (Direct Air Capture).

The fossil fuel route and bio route both use something called “syngas”. What’s that?

Sourcing carbon for methanol

The bio route sources carbon from biomass that originally absorbed CO2 from the atmosphere. This includes:

• Agricultural waste (twigs, stalks, grass, etc)
• Leftovers from making pulp and paper
• Municipal waste
• Sewage
• Biogas (which I explain here), though this requires reformation, not gasification to make syngas.

The electro route sources carbon from CO2 streams. This includes:

• CO2 captured from industrial plants using fossil fuels. This is bad, because the CO2 didn’t come from the atmosphere originally.
• CO2 captured from industrial plants using biomass. For example, plants that make ethanol! This is better, because the CO2 came from the atmosphere originally - however, sourcing non-bullshit sustainable biomass is tricky.
• CO2 removed directly from the atmosphere, using Direct Air Capture. This is the best source of CO2, since it unequivocally came from the atmosphere originally. However, DAC has its own problems!

E-methanol is a bit needy 😇

Let’s revisit our production diagram:

Going the electro route requires the critical ingredient of green hydrogen. And remember, green hydrogen needs a lot of electricity.

As a rule of thumb, you need about 10 MWh of electricity to produce 1 ton of e-methanol. Almost all of this is for hydrogen, and this doesn’t include energy for powering DAC.

To replace just our current methanol production (100 Mt / yr) with e-methanol, we would need around 1000 TWh of electricity. That’s roughly the electricity consumption of the UK, Germany, and the Netherlands combined. Crucially, that’s before we even start making any more methanol to use as a fuel.

This means e-methanol is pretty costly, especially when you factor in the costs of DAC.

Conclusion

That’s all for ethanol and methanol!

Here’s a cheat sheet of the important questions to ask about production projects.

It’s bad when you suffocate decomposing biomass, because it makes methane. Examples of this are:

• Landfills
• Rice paddies (biomass covered with water)
• Big piles of biomass, where the center isn’t aerated (eg, big piles of animal manure)
• The reservoirs of dams. The biomass in the valley is covered by water, so dams can emit a lot of methane.

The point of the biogas industry is to capture methane from things-decomposing-without-oxygen.

In this piece we’ll walk through biogas, landfills, anaerobic digesters, and whether biogas is any good. Let’s go.

Biogas and landfills

A big ‘ol pile of biomass, like a landfill site, will generate methane and CO2 as it decomposes. This mixture is called biogas, and it’s usually about half methane, half CO2.

Places like the UK & US capture most of this gas. But worldwide, most landfill gas is not captured, so the emissions from landfill sites are colossal. Landfill gas is the third biggest source of methane (after livestock/ag and fossil fuel extraction).

Luckily, it’s easy to capture the gas. Just seal the landfill with a plastic liner and poke some straws in.

Free electricity baybeeee (kinda)

The captured biogas is burned to make heat or electricity. Burning things sounds bad for the climate, but in this case, it’s actually good. By burning the methane in the biogas, we’re turning it into less-warming CO2.

We “upgrade” a small amount of biogas (about 17%), by removing the CO2 to leave pure methane, called biomethane. We can then inject this into the gas grid and call it Renewable Natural Gas (RNG), because it’s a non-fossil-fuel source of methane.

Here’s the summary so far.

Note that there are two main ways of producing biogas: landfills and anaerobic digesters.

Anaerobic Digesters

An anaerobic digester is just an accelerated landfill. It’s a big, airtight tank where you stir your waste to turn it into biogas quickly - a month instead of a lifetime. Anaerobic just means “without oxygen”.

Anaerobic digesters really do just look like big tanks irl.

Is this stuff good for the climate?

It’s worth stepping back for a moment to evaluate the flow of GHGs here.

Remember: because we’re dealing with biomass, the CO2 that it contains originally came from the atmosphere.

Capturing and burning biogas from landfills is good, because those emissions are unavoidable. Without capturing it, we’d just release methane.

But what about anaerobic digesters? Aren’t they just accelerating the process of making methane?

There are two lenses we can take here: the methane lens and the feedstock lens.

The Methane Lens

If you are using biomethane from ADs to displace natural gas, that is objectively good for emissions.

• The CO2 from burning fossil natural gas came originally from the ground.
• The CO2 from burning biomethane came originally from the atmosphere.

It seems that most biomethane today does displace fossil fuels, so this is good!

The Feedstock Lens

Here we consider the counterfactual: what would have happened to our feedstock if we didn’t anaerobically digest it?

• If our biomass would have decomposed to make uncaptured methane -> AD is good.
• If our biomass would have decomposed to make uncaptured CO2 -> AD is a wash, and quite possibly worse than nothing. Not only is the end product still CO2, but you’re making intermediary methane, which could leak.

Most of the things that we put into anaerobic digesters today fall into the first category, which is GOOD! Two examples:

1. Animal manure. Many ADs in the US are 100% animal manure, that otherwise would have decomposed into unabated methane.
2. Food waste that otherwise would have gone into landfill. Even if the landfill has methane capture, anaerobically digesting waste will have a higher capture rate.

Can biogas be scaled?

Today, biogas is predominantly a German story. Europe produces more biogas than the US and China combined, and two thirds of European capacity is in Germany.

Across the whole of Europe, we produce 191 TWh of biogas and upgrade 17% into biomethane. That’s not nothing, but it doesn’t make much of a dent in our gas supply.

Worldwide, we have enough sustainable biomass to replace 20% of natural gas usage with biomethane, today. It’s just that natgas is cheap, and making biomethane is expensive.

Natural gas is currently about $30 / MWh in Europe, and about $7 / MWh in the US (fracking!). Today biomethane will run you about $100 / MWh. Most of the world’s biomethane potential is estimated to sit between $35 and $70 / MWh.

Conclusion

The world is full of decomposing plants, food waste and manure. Right now, a lot of this is spewing uncaptured biogas. Capturing that is a good thing.

Theoretically, we could tell people to aerate their waste biomass, so that it just makes CO2, and not biogas. But they don’t get much in return for doing this. By making biogas - either from anaerobic digesters or landfill capture - they get a fuel that they can use or sell. Not only that, it can displace fossil methane.

The capacity for biogas to scale is limited. We can’t produce enough biomethane to replace all of today’s gas demand. Luckily, gas demand will decrease due to electrification. Denmark is aiming for 100% of gas usage to be biomethane from 2030, because of high heat pump adoption and increased biogas production. So biogas could make up an increasing proportion of a shrinking quantity.

I’ll start with the spoilers.

• Ammonia is the biggest CO2 emitter in the chemical industry.
• Ammonia makes a terrible fuel. It’s just a more expensive form of hydrogen. And it’s toxic.
• Using ammonia for electricity generation is bullshit, in Europe and the US.
• But ammonia is easier to transport by ship than hydrogen. Because of this, Japan, Korea, and other energy-poor countries plan to import ammonia.

Let’s begin with the basics.

30 second ammonia crash course

Ammonia is made of hydrogen and nitrogen.

Today, we use ammonia for its nitrogen (we make nitrogen fertiliser).

In the future, we might use ammonia for its hydrogen (for energy).

In the context of energy, ammonia is just fancy hydrogen.

In 2024 almost all ammonia is used for making fertiliser. Without ammonia, 8 billion people could not be alive at once. But producing it is dirty - ammonia is 45% of emissions from the chemical sector.

In the future, we’ll clean up our existing ammonia production. We might also use it for new things.

This blog will walk through these three categories.

#1 - Fertiliser

Plants grow better when you add nitrogen.

In the 1800s, Europeans’ eyes were popping out their sockets at how fast their crops were growing after they added guano (bird poop, which contains nitrogen). They found some islands near Peru with loads of guano and cheerfully mined the heck out of them.

But by the early 1900s, the guano was running out. Two German chemists invented the Haber-Bosch process - the first decent industrial process for producing ammonia. Ammonia contains nitrogen, so this meant that for the first time ever, humanity could produce nitrogen fertiliser on demand.

The Haber-Bosch process is probably humanity’s most important invention, because it feeds most of the global population today. Roughly half of the nitrogen inside your body came from the Haber-Bosch process.

Unfortunately Haber-Bosch plants today are mostly fed by natural gas, which is a total bummer.

Ammonia production today

Here’s how we make fertiliser today.

Let’s zoom in on the Haber-Bosch process.

We make three types of nitrogen fertiliser from ammonia: nitrates, urea, and just raw ammonia. They each have different CO2 and NOx emissions in production and on the field.

But all types of nitrogen fertiliser emit a lot of CO2 in production, because they use dirty hydrogen made from fossil fuels.

Clean ammonia

So how do we clean up the process shown above?

There are two sources of emissions here.

1. CO2 released from grey hydrogen production (see my guide to hydrogen for more on this)
2. CO2 released from burning natural gas to heat up our Haber-Bosch plant.

Emissions from hydrogen (#1) are the most important. They make up between 75% and 90% of the CO2 emitted in ammonia production. That means you can clean up most of ammonia’s emissions by simply substituting grey hydrogen for green hydrogen. Right now, green hydrogen is much more expensive than grey hydrogen - but that might change soon.

Decarbonising #2 is hard. The race for the best solution is being run as we speak. The contenders are:

1. Keep doing Haber-Bosch, but heat it with electricity or hydrogen. Haber-Bosch runs around 450°C, so replacing natural gas with electric or hydrogen heating is pricey (24/7 electricity typically costs around 3x more than gas). Note: the plant heats itself once started. Currently cold starts are rare, but intermittently-made electricity or hydrogen might change this in future.
2. Electrochemical nitrogen fixation. This replaces the Haber-Bosch process entirely. Using two plates and some electricity, we can turn nitrogen and hydrogen into ammonia. I’m excited about this - particularly with AI-powered materials/catalyst discovery - but the electricity requirements are still eye-watering. Also, most current electrochemical processes for this require a lot of lithium.
3. Bacteria. Again, this replaces Haber-Bosch. We can use bacteria to turn nitrogen into ammonia, but doing it at scale is tricky.

Cleaning up ammonia production is important, not just because of emissions. It also means that fertiliser and the price of food are more insulated from fossil fuel price spikes and the accompanying geopolitics.

Ammonia infrastructure

Because we already make fertiliser, some places already have infrastructure and laws for handling ammonia.

In the UK and Europe, we primarily use nitrate fertilisers. Nitrate fertilisers are made from ammonia, but shipped to farms as coarse powders. In the US, ammonia is applied directly to the soil. It’s stored as a pressurised liquid in tanks on the farm, and there are ammonia pipelines to transport it from gas-rich Texas and Louisiana up to the corn belt.

#2 - Ammonia as a hydrogen carrier

Here’s something that blows my mind: liquid ammonia contains 50% more hydrogen by volume than liquid hydrogen. You can transport the same amount of hydrogen, using less space, if you transport it as ammonia. Pure hydrogen is of course much lighter, but ammonia takes up less space.

Ammonia also turns into a liquid much easier than hydrogen. To turn hydrogen into a liquid, you have to squish it to 300x atmospheric pressure, or cool it to -253°C. For ammonia, it’s 10x atmospheric pressure, or -33°C.

That means that ammonia could be a convenient way to transport hydrogen over long distances. The idea goes like this:

Upon reaching its destination, ammonia can be “cracked” back into hydrogen, or used directly as a fuel (which we’ll cover in section 3). Cracking just means splitting ammonia back into hydrogen and nitrogen. You heat it up to about 800°C over a nickel catalyst, and it pops right apart. Note: cracking takes a lot of energy!

This whole scheme sounds pretty neat on paper, but the real world is harsh.

The problem: ammonia is just more-expensive-hydrogen

Every conversion step in the picture above loses energy. You can expect to waste at least 70% of your energy - meaning that your electricity-to-electricity efficiency is 30% on a good day, and probably much lower.

It simply costs money to do this. Just converting hydrogen to ammonia adds roughly an extra $1 per kg of hydrogen. Remember from part 1 that we’re competing with grey hydrogen, which you can pick up for $1.50 per kg in the US.

So let’s be clear - it makes no sense to ship ammonia to places that can make their own hydrogen. No one sensible is proposing it.

However, for countries without the resources to produce hydrogen, importing ammonia could be useful. A good example is Japan.

Japan and ammonia

Japan drew the geographical short straw for energy production. It has a terrain of rugged mountains and forests, so solar capacity is limited. It has a deep seabed close to the coast, so conventional offshore wind is tricky too. After Fukushima, there is an understandable resistance to nuclear. Interconnecting with China is politically risky. And limited oil & gas fields mean that there is limited capacity for sequestering captured carbon.

Japan currently imports a lot of LNG (Liquefied Natural Gas) on big ships. They can’t use dirty natural gas forever - but they will still need to import some energy. Thus, they are placing bets that they will be able to import hydrogen and ammonia in the future. They've made some strong policy decisions to this effect. Korea and other smaller Asian countries aren't far behind.

Ammonia is shipped

It’s worth noting that ammonia is already shipped around a lot. Port infrastructure is already mature because of Big Fertiliser.

Not only does infrastructure already exist, but regulations do too. They’re important when you’re shipping a nasty chemical like ammonia.

It’s useful that the cost of ammonia transport is relatively small, because the spreads between clean ammonia prices in different countries will be big. Here’s the example of Singapore producing vs importing ammonia.

We know for a fact that green hydrogen production will cost wildly different amounts in different parts of the world. That’s because green hydrogen is dependent on renewable electricity, so places with lots of sun and wind will have a big price advantage.

Yes, shipping hydrogen via ammonia will waste a lot of energy. But if the green hydrogen is sufficiently cheaper at the source than the destination, that might not matter.

#3 - Fuel

Finally, let’s talk about using ammonia itself as a fuel. The two most-discussed applications are shipping and electricity production.

Why not cars, road vehicles and boilers/furnaces too? For starters, ammonia is even more inefficient than hydrogen. You will waste a lot of electricity making green hydrogen to make clean ammonia. You’re much  better off using electricity directly, in an electric vehicle or heat pump.

Ammonia vs hydrogen

It’s worth emphasising this one:

In Europe and the US, no-one talks about ammonia as a fuel very much. That’s because we can make green hydrogen. If you have access to hydrogen as a fuel, you will use it directly instead of making ammonia. It’s only in places with limited hydrogen production (like Japan or Korea) where ammonia as a fuel is interesting.

Also, ammonia hates you.

Ammonia and human bodies are not pals. Ammonia burns and destroys human tissue, including eyes, lungs and probably most of your top 10 fave limbs. Inhaling ammonia can make you blind, or dead.

When ammonia is handled by industry, it’s centralised and contained, so it’s easier to keep safe. But using ammonia as a consumer fuel would be silly. In the US, where ammonia is used on farms, severe accidents are frequent.

Boom-boom vs zippity-zappity

There are two ways that you can use ammonia as a fuel. You can burn it, or put it through a fuel cell - just like hydrogen.

A fuel cell takes in ammonia, and outputs electricity. You can get 40-65% efficiency from an ammonia fuel cell, depending on how much you spend.

Meanwhile, burning ammonia produces heat, and as always, turning heat into useful energy is inefficient. You might get 35% efficiency in an internal combustion engine (heat to motion), or 55% in a combined cycle power plant (heat to electricity).

One more thing - if you don’t combust ammonia completely (which is quite easy to do), the nitrogen in ammonia can produce nitrous oxide (NOx). This is extremely cringe and a massive self-own as NOx is a potent greenhouse gas (about ~300x the warming potential of CO2).

Ammonia for shipping.

Ammonia and methanol are the two big contenders for alternative maritime fuels.

We’ll have biodiesel, batteries, and maybe even carbon capture on some ships, but there remains a need for a renewable, clean, long-distance fuel.

Methanol is a hydrocarbon. That means it’s more similar to current fuels, so ships don’t need heavy modification. That’s an advantage, but also its achilles heel. The carbon that it contains has to be sourced from somewhere, and it’s expensive. We’ll cover methanol in depth in part 4 of this series!

Big ship-makers like Maersk are informally adopting the strategy of “methanol today, ammonia tomorrow”.

Ammonia for power generation

It’s possible to burn ammonia in a modified gas turbine or coal plant to produce electricity.

By now you should know where this is going: ammonia costs more than green hydrogen, and green hydrogen costs more than gas. Because ammonia is hard to ignite, it would likely be co-fired with fossil fuels.

Whilst it might be a necessary option for somewhere like Japan, that doesn’t mean it’s any cheaper.

To be honest, I haven’t quite decided what I think about Japan's strategy here, because two extremely strong forces are colliding.

#1 - You can’t operate an industrial economy with expensive energy

Japan has a large manufacturing and industrial base that relies on cheap energy to be a competitive exporter.

Importing energy in the form of hydrogen, via ammonia, means that Japan’s energy costs will be far higher than most other countries. Ammonia is simply far more expensive than fossil fuels - at least for the medium term.

Switching to ammonia anytime soon would be like a self-imposed russian gas crisis. It would put a huge drag on energy prices, industry, and living costs - and make it difficult to be a competitive global exporter.

#2 - Cheap solar

Solar is getting stupidly cheap, year after year. This is unquestionably the most important fact for future energy.

If you take the PV price curve seriously and extrapolate it a decade or two, then we will have cheap green hydrogen. It will become cost-effective to build crappy electrolysers that can run on limited capacity factor solar, because energy is nearly free.

Since green hydrogen is the main cost centre for clean ammonia, cheap solar means cheap clean ammonia.

Summary

Ammonia is hard.

• It’s even more expensive than green hydrogen.
• It’s even less efficient than green hydrogen.
• It’s toxic, and hard to burn.

But it has some useful features.

• It’s more compact to transport than hydrogen.
• Unlike methanol, it doesn’t contain carbon. It doesn’t emit carbon when burned, and it doesn’t need to be made with expensive carbon from the atmosphere.

Huge thanks to Teja, Sharon and Lisa giving their thoughts on this piece 🙏

Hydrogen is also the dominant cost in other clean fuels. It is not possible to produce cheap green ammonia or Sustainable Aviation Fuel without cheap hydrogen.

Steel, shipping, aviation, and chemicals will need a lot of hydrogen to decarbonise. In most cases, they won’t use hydrogen directly, but will use chemicals built from hydrogen.

Caution! Sometimes hydrogen is shit!

Here’s an example: we could use hydrogen for cars, but we don’t. EVs eat hydrogen cars for breakfast.

Here’s a simple guide to where we should use hydrogen to decarbonise.

For a more detailed version, check out Michael Liebreich’s hydrogen ladder. He does an excellent 20 minute walkthrough.

Now let’s get into the interesting bit. How do we produce hydrogen?

Today, hydrogen is a climate problem, not a solution.

We already make a lot of hydrogen today - about 100 million tons a year. It’s mainly used to:

1. Make ammonia for fertiliser, and
2. Refine oil and produce petrochemicals.

This means that producing hydrogen is responsible for 2% of total emissions, and 6% of global natural gas usage.

Today, we mostly make hydrogen with a process called Steam Methane Reforming (SMR). We call this type of hydrogen grey hydrogen, because it uses fossil fuels.

About 70% of global hydrogen is made using methane (natural gas). In China, coal is often used, which emits about 2x the CO2. China is coal-rich and gas-poor, so it's a logical choice.

Clean hydrogen

Fossil fuels are cringe, so we need find a way to produce hydrogen cleanly. There are two main methods.

Blue hydrogen

Blue hydrogen simply takes grey hydrogen and slaps carbon capture on top.

Blue hydrogen is controversial, cos carbon capture has a long history of... not working. Carbon capture tech is pretty simple, but the incentives for companies to make it work are poor. You are adding cost & energy requirement (all downside) for no financial gain.

Additionally, blue hydrogen still uses methane (natural gas) - which leaks during production and transport. These methane leaks can cause even more warming than the CO2 emitted from SMR (depending on where you get your methane).

Green hydrogen

Now we're hitting the cool stuff. Green hydrogen is the most important production method to understand.

Making green hydrogen is simple: it’s just two electricity-conducting sticks in some water. Passing electricity through H2O splits it into H and O.

(PS - if you are a looking for a climate tech activity for a dinner or an EXTREMELY ROMANTIC DATE, make green hydrogen with a 9-volt battery, two pencils, and a glass of salty water)

Green hydrogen uses a lot of electricity

It’s important to understand how green hydrogen is made, because it will have a huge impact on our global energy system.

The IEA predicts that by 2050, making green hydrogen will use 20% of global electricity.

Let’s learn a bit more about the machines that will slurp up all this electricity.

Splitting water to make hydrogen is electrolysis, and the machines that do this are called electrolysers.

There are three types of electrolysers. You can impress any hydrogen nerd by asking what type of electrolyser they’re using.

PEM is well suited to running off of renewable power, because you can temporarily overload it. It deals well with spiky electricity.

There’s also a fourth type of electrolyser called AEM - which is a cross between Alkaline and PEM. In theory, you receive the benefits of PEM (flexibility), without the high cost of the rare materials. It’s early days, but AEM is being commercialised by Enapter.

Over the past five years, we’ve installed roughly twice as much alkaline as PEM.

Unit economics of hydrogen

Let’s go through some rules of thumb for the cost and energy requirements of green hydrogen.

Heads up: people usually measure hydrogen in kilograms (kg).

Energy (kWh per kg)

To make green hydrogen, there is a minimum energy requirement. It is the minimum energy needed to split water, and it is 40 kWh per kg of hydrogen.

Of course, electrolysers are not 100% efficient, so most today achieve around 50 kWh / kg on a good day.

Cost ($ per kg)

Here, we’re competing with fossil fuels.

• In Europe (where gas is expensive), grey hydrogen costs about $2.50 / kg
• In the US (where gas is cheap) grey hydrogen costs about $1.50 / kg

The dream is for green hydrogen to one day cost $1 / kg or less. At that level it would blow grey hydrogen out of the water. But is it realistic?

Firstly, let’s figure out how much the electricity for green hydrogen will cost us.

Grid electricity

Let’s initially assume that we power our electrolyser from the electricity grid.

Remember, we need 50 kWh of electricity to make a kg of hydrogen.

If each kWh costs $0.10 (roughly the price of US grid electricity), then the electricity will cost $5 per kg of hydrogen.

That’s not a good start. Even with the colossal IRA subsidy of $3 / kg for green hydrogen, we’re still far from our target price of $1 / kg. And we haven’t even spent any money to build an electrolyser yet.

Off-grid electricity

If electricity at $0.10 per kWh is too expensive, where can we find some that’s cheaper?

The answer is renewables. When the sun is shining and the wind is blowing, electricity is the cheapest it’s been in history.

We could either:

• Build dedicated solar or wind farms, just for powering electrolysers. Or,
• Use the excess energy from existing renewables, when they generate too much energy.

(That second one might sound silly - how can we have too much renewable energy? But it’s actually a big problem. In the UK we spend £billions per year turning off wind turbines).

The problem with both of these approaches is that you build an expensive electrolyser that just doesn’t get used very often - because renewables are intermittent.

The best way to get an intuitive grasp of the CAPEX/OPEX tradeoff here is to play around with a hydrogen calculator. I like Electric Hydrogen’s one.

This means that the hydrogen tech we should be building is the opposite to what many assume (and what most acadamic research is focused on). It’s not about making efficient electrolysers that are cheap to run, it’s about making rudimentary ones that are cheap to build. Reducing CAPEX, not OPEX, is the key to building an electrolyser that can pay off its costs.

Transport

Hydrogen is a pain to handle. Today, most hydrogen is not transported at all. It’s used right next to where it was produced.

Hydrogen is tricksy because (1) it’s not very dense, (2) it’s a very small molecule, so it leaks a lot, and (3) hydrogen is very flammable - The Hindenburg was not cool.

You can move hydrogen by pipeline, truck, rail, or boat.

Pipeline

Sometimes, natural gas pipelines can be repurposed to carry hydrogen - but there are some hefty caveats.

Three things to consider:

1. Leaks. Hydrogen is the smallest molecule in the world. It will find its way out of seals, valves and cracks that natural gas would not.
2. Embrittlement. Local gas distribution pipes (made from soft steels and plastic pipes) can handle hydrogen just fine. But putting hydrogen through large national gas pipelines - made from hard steels - can “embrittle” (corrode and crack) the steel.
3. Compression. To deliver the same amount of energy using hydrogen as with natural gas, you must use 3x the energy compressing it (h/t Paul Martin). That’s why today we pump around natural gas instead, and turn it into grey hydrogen where required.

Truck

Hydrogen as a gas at room temperature takes up a lot of space. Roughly 3x more space than natural gas, and 300x more than diesel.

So moving it in its uncompressed form isn’t an option: we have to squish it.

To put hydrogen on the road, you can either: (A) cool it a lot and transport it as a liquid in a tanker, or (B) pressurise it a lot and transport it as a gas in a tube trailer.

Both of these options are a bit pants. Liquifying hydrogen uses up about 30% of its energy, and hydrogen as a gas just isn’t very dense. You’d need >10 tube trailers to transport the same energy as a diesel tanker truck.

Ship

Fun fact: there is one hydrogen tanker in the world, and it is the Suiso Frontier - made in 2019.

We liquefy hydrogen to ship it, and so we encounter the same 30% loss from liquefying it - plus around 0.2-1% of the hydrogen will “boil off” for every day of the voyage.

Shipping pure hydrogen is rubbish. A more attractive way to ship hydrogen is in the form of ammonia. People sometimes refer to ammonia as a hydrogen carrier. This is pretty interesting, and we’ll cover it in part two of this series, which is all about ammonia!

Storage

Lastly, we have storage.

We can store hydrogen as a dense liquid for a short time, but because we have to keep it at -253°C, it would be too energy intensive to do this for long periods.

That means we have to store it as a gas. Storing it in big gas tanks is tricky, because (1) they'd have to be huge (2) you're storing a a very flammable gas, and (3) the hydrogen will just escape. Yep, hydrogen will literally just diffuse through a steel wall, which is sooo annoying. As the smallest atom, it's hard for other atoms to keep it in check.

Luckily some geographies are blessed with the ability to store hydrogen underground. There are three main places we could store it: gas fields, aquifers, and salt caverns. These are the same places that we store natural gas. If you want to learn more, the wikipedia page on geological natural gas storage is a great place to start.

Conclusion

We will need a lot of green hydrogen - even if it's not as much as the natural gas industry wishes.

That means we'll need an eye-popping amount of electricity to make it. This simple fact produces some of the biggest costs & headaches in industrial decarb. If you want to make a clean steel plant that uses green hydrogen, the easy part is building the green steel plant. The hard part is sourcing the green hydrogen and necessary TWhs of electricity.

Huge thanks to Gniewomir Flis for giving his thoughts on this piece. Opinions and mistakes are all mine.

Next up: ammonia.

This was #1 in my series on climate molecules. Keep reading here!

Firstly, it’s much easier to artificially dim the sun than most people imagine.

100 planes injecting sulfur particles into the stratosphere would dim the sun by about 1%, and cool the earth by about 1°C.

For context, our current sulfur dioxide emissions (from fossil fuels) are >10x bigger than the 1.2m tons shown above. The difference is the height of injection. We’d be putting sulfur into the stratosphere, which is higher up than our normal sulfur emissions in the troposphere.

Spraying a form of sulfur from a plane is incredibly cheap. A full programme would cost less than $20b per year. That’s much cheaper than carbon removal ($600b per year, to remove just 10% of annual emissions @ $100 / tCO2).

Modifying the earth like this is called geoengineering, and blocking out the sun with particles is called Solar Radiation Management (SRM).

SRM is awakening in 2024

Interest in SRM spiked in 2024, after years of being too controversial to discuss. Here are a few reasons why.

#1 - We don’t know why 2023 was so hot

The climate data from 2023 is scary, because we do not understand why 2023 was so warm. El Niño and low-sulfur fuels had an impact, but their effects were anticipated, and don’t fully explain the temperatures that we saw.

The actual warming in 2023 fell far outside of scientists’ predictions. This is what hitting a climate feedback loop looks like.

#2 - Billionaires are getting interested

It only takes one person rich enough to start a geoengineering programme. In 2023, some powerful and wealthy people started talking publicly about geoengineering.

#3 - Institutions are awakening to geoengineering

• In 2023, both the US and EU commissioned research on solar geoengineering. In the past, state governments have refused to fund anything related.
• Climate scientists are starting to seriously talk about solar geoengineering. Traditionally, they’ve self-censored on this topic, but this is changing. I highly recommend Robinson Meyer’s fascinating account of this.
• The internet noise is rising from a stutter to a low hum. In 2023/24, people are talking about SRM. Casey Handmer, Keep Cool, The Economist, and the BBC, to name just a few. I personally feel that 2023 was the year that SRM broke a critical threshold in the broader climate consciousness.

Ok, let’s rewind for a sec.

You’re telling me that sulfur reflects the sun?

If reflecting the sun with sulfur sounds far fetched - well, we’re already doing it.

The two best examples:

1. In 1991, Mount Pinatubo erupted and ejected millions of tons of sulfur dioxide. It cooled the earth by around 0.5°C, and lasted for around two years.
2. We already put sulfur into the atmosphere by burning fossil fuels. The sulfur that we’ve emitted so far is already reflecting sunlight, and is currently “masking” about 0.5°C of warming. Without these particles, we’d already be at 1.8°C warming, not 1.3°C.

In 2020, the International Maritime Organisation introduced limits on the sulfur content of shipping fuel. It reduced sulfur dioxide emissions by about 8 million tons in its first year - about 10% of the global total.

This reduction in sulfur emissions has “un-masked” warming that was already there. It’s one of the reasons for the jump in warming in 2023 (though it doesn’t fully explain the abnormality).

Instant effect

Because SRM acts almost instantly, we’d need to phase it in gradually. If we began blasting maximum sulfur tomorrow, we’d cause a sharp fall in temperature that would be a nasty shock to the climate.

A better plan is to use sulfur to pause our warming at its current level (about 1.3 °C) whilst we slash emissions. Eventually SRM would become redundant as methane’s warming falls off, and carbon removal takes effect.

Sulfur disappears from the atmosphere quickly - it rains out after about a year. This means that once we’ve started SRM, it’s dangerous to suddenly stop. We need to keep spraying particles, all the time. If we suddenly stopped, the warming would spring back rapidly, causing a bad temperature shock. The correct way to stop is a gradual phase out.

The rub

Unfortunately, Solar Radiation Management (SRM) has some fairly gigantic problems.

1. It doesn’t fix the root cause. Cooling the planet does not remove the CO2 which has accumulated in our atmosphere. It doesn’t stop that CO2 from acidifying the oceans and irreversibly destroying marine biodiversity.
2. SRM could make us complacent about dramatically cutting emissions.
3. Sulfur increases acid rain (harmful to many life forms) and will likely harm the ozone.
4. If SRM is poorly implemented, it could dramatically change weather and rainfall patterns. For example, if sulfur is not injected near the equator, it will not evenly mix into the stratosphere, causing uneven cooling and heating.

Shifted weather and rainfall patterns are one of the scariest possibilities of SRM. Even if deployed “correctly”, it’s unclear how SRM would affect rainfall and monsoons. Our current research suggests that when deployed in the correct locations and volumes, SRM significantly moderates climate hazards everywhere, and significantly exacerbates them nowhere. But we will not know until we try.

Crucially, the biggest problems with SRM are probably not yet known. The side effects of putting sulfur into the stratosphere could be some of the most consequential unknowns in human history. Clearly, that’s a huge reason to do more research. But still - no matter how much research is done, when humanity tries to bend nature to its will, we can be sure that unintended consequences won’t be far behind.

SRM will be deployed without your consent

Geoengineering will not happen by international consensus. I find it impossible to imagine a UN mechanism approving something so universally contentious.

Rather, someone will probably just do it.

A country with no choice

Some nations will soon have a simple choice: deploy geoengineering or cease to exist.

People underestimate how controversial it can become to not do SRM. Imagine you are the leader of a country close to the equator. Crop failures, extreme heat, and city-destroying cyclones mean that your people are without drinkable water, have nowhere to sleep, and cannot feed their children. Mass social unrest and physical violence become normal for your country. SRM is the only action that you can take to turn off the disasters, and prevent your government being overthrown.

Because SRM is so cheap, it’s completely feasible that one country could do this alone. (The Ministry for the Future and Termination Shock imagine interest from India, Singapore, Venice, Texas and The Netherlands)

Ultimately, the decision to turn on the SRM machines will not be made by climate scientists, or carefully calculated risks. It will be made on the basis of nations rising or falling - by starving populations, revolutionaries, and leaders with their back against a wall.

The conditions only have to align for one country, at one time, to begin geoengineering.

A billionaire

There are three important things to know about billionaires:

1. The business interests of many billionaires face total decimation at the hand of climate-induced conflict and destabilisation.
2. Many billionaires attribute their success to being hyper-logical, no-BS, getting shit done, and acting upon contrarian beliefs (sometimes this is even correct). The appeal of SRM to this mindset needs no explanation 🤠
3. Some billionaires want to use their money for good. The great challenge of the billionaire is how to convert a fortune into a timeless legacy.

It doesn’t matter if 99.9% of billionaires think that geoengineering is a terrible idea. It only takes one person rich enough to implement it.

Every year, SRM inches towards “the logical choice”.

Deploying SRM in 2024 is probably not the logical choice. But it’s not long before that changes.

Taking climate catastrophe off the table

The further that we depart from our current temperature, the more we risk runaway changes that we can't control. We don’t know much about when tipping points will occur - we just know that they are out there.

Why risk the arctic permafrost melting? Why risk losing the Greenland ice sheet? Why risk the tipping points that we don’t even know about?

These risks increase every year that our emissions are above zero - including every year that we are dramatically reducing emissions. SRM gives us the ability to lop off the most catastrophic outcomes from the end of the probability distribution.

Geoengineering is no replacement for getting our shit together. But there would be no honour in allowing the deaths of hundreds of millions of people, simply because they could have theoretically been avoided through more mitigation.

Let’s check in on global emissions

Here’s the actual forecast from the US Energy Information Administration (for energy-related CO2 only).

Whilst the EIA is usually wrong, it’s true that things are going to get bad before they get better. The maths no longer checks out to hit 1.5 degrees, and even succeeding in holding 2 degrees would be a disastrous outcome.

I don’t point this out because emissions cuts are hopeless. On the contrary - we will electrify everything, solar & batteries will eat the world, and emissions will plummet. But it still might take a decade or two too long. It would be a shame to die in the meantime.

To be clear: there is no valid argument that SRM should replace deep emissions cuts. But deploying it alongside deep emissions cuts could reduce risk and deaths from climate change.

Climate controversy repeats itself

SRM is the last of the four climate solutions to become widely discussed. It seems to me that their entry into the zeitgeist is ordered like this:

Take a look at this flow of climate effects.

The best blend of future solutions to minimise human suffering probably includes some SRM. This is because:

• Mitigation can’t go fast enough to do 100% by 2050.
• Carbon removal is too expensive to do more than a few percent.
• Adaptation doesn’t actually stop temperatures rising.

Final thoughts

So how would we actually put sulfur into the stratosphere?

The most widely discussed method uses customised planes to fly missions into the stratosphere and release sulfur dioxide. People also talk about cannons that launch sulfur rockets, and (more practically) high-altitude balloons filled with sulfur and hydrogen/helium.

It’s easy to imagine a world where amateur enthusiasts launch large sulfur balloons from their backyard, just like hobbyists launch weather balloons today. Scientifically, this would be a terrible idea (remember, we only want to inject sulfur near the equator, in a controlled way). But it’s easy to see the individual appeal of offsetting the warming of hundreds of thousands of people with just a few balloons.

Is it only sulfur that will work?

No. In fact, sulfur is unideal for a lot of reasons (it causes acid rain, could harm the ozone, and in general is not friendly to humans). But it mimics the “natural” effects of volcanoes, and our human sulfur emissions already do a lot of cooling, so we know some things about its effects on the atmosphere.

One promising alternative is calcium carbonate (chalk). Unlike sulfur, it is alkaline, not acidic - which means that it would accelerate the healing of the ozone hole as a co-benefit. Here’s a long blog post on the details of calcite geoengineering.

Harvard has been trying to run a small experiment called SCoPEx to investigate calcium geoengineering. Despite being a fairly basic experiment (injecting less than a kg of calcium into the atmosphere from a balloon), the project has been stuck in approval committees for several years, because of the controversy of researching geoengineering.

Conclusion

SRM might not make sense in your mind (it certainly doesn’t in mine). But do you view the world in the same way as a military dictator, “benevolent” billionaire, or leader of a starving country?

It's 2024, and for better or for worse, people are starting to take geoengineering seriously.

The end

Thanks to Ed Maclean (writes Breakfast of Challengers), Laura Lock, and Art Lapinsch (writes Delphi Zero) for their thoughtful comments on this one. All opinions are mine, not theirs (in fact they all disagree with at least one thing I’ve written above).

📫 If you enjoyed reading this and want to get my future climate blogs in your inbox, you can drop your email here.

Other writing

If you enjoyed reading this, you might like some of my other posts:

• The Big Guide to Batteries
• The Energy Fundamentals of Carbon Removal
• Three Inputs = All Decarbonisation Tech

This piece is here to walk you through them all. It's intended as a concentrated onboarding guide that will orient you within the battery world of 2024 🧭. There's no chemistry, and no assumed knowledge. (By the way, this post is exclusively about electrical batteries - not thermal batteries or other types.)

In the future, our battery usage will be dominated by two things: electric vehicles and grid storage. In the first part of this guide, we’ll learn about batteries for EVs (batteries that are light and powerful). Then, we’ll move onto batteries for grid storage (batteries that can be heavier and cheaper).

Part 1: The EV battery battle

Lithium ion

I used to think that all lithium batteries were the same. I knew that these things called “lithium ion batteries” were in our phones, laptops, EVs, and rechargeable nose hair trimmers. I figured that apart from some slight tweaks, all lithium ion batteries were pretty similar.

I was very wrong! Different types of lithium ion batteries are actually made of completely different materials. Lithium is the only big one that they have in common. In fact, there’s a fascinating battle unfolding between different types of lithium ion battery. It’s not just a story about the best battery chemistry, but one of cost, patents, and jostling superpowers.

On top of this, there’s a new kid in town. Until recently, most people thought that all EV batteries would be lithium ion. But the sodium ion battery had a breakout year in 2023. There’s a lot to learn - so let’s dive in.

Here’s a picture of how fast things are moving in EV battery land. My promise to you is that by the end of this section you’ll be able to explain all the trends happening in this graphic.

We’re going to learn about the two dominant types of lithium ion batteries: NMC and LFP. First, we need to arm ourselves with some basic terminology.

Parts of a battery

It’s useful to know about the main parts of a lithium ion battery. Basically, it’s a cheese sandwich.

Either side of the separator (cheese) are the cathode (bread 1) and the anode (bread 2). The cathode and the anode are the two large bits of the battery where the energy is stored. A lot of people are working on making bread that is lighter, smaller or cheaper, because it allows our sandwich to store more energy for less weight/space/money.

Let’s briefly go through the story of the lithium ion battery because, unbelievably, it starts with Exxon.

The birth of lithium ion in two paragraphs

In the 1970s, OPEC had put an oil embargo on the US, and people were getting nervous. Exxon decided to do some research on electric vehicle batteries, and made some progress understanding this funky thing called “intercalation”. Intercalation is a way of storing atoms in separate “host” material. If you make the atoms charged (that’s called an “ion”), you can store charge in the host material. Congratulations, you just invented the “cathode” from the picture above 🤝. Intercalation is also the namesake of the premier newsletter on batteries.

Some early lithium batteries got rapidly commercialised by a Canadian startup - but they had a tendency to catch fire. Luckily, in the early 1980s a dude named John Goodenough lived up to his name and tried using a material called lithium cobalt oxide (LCO - our first acronym 🤘) for the cathode. It worked great. Once we figured out how to make a good anode (spoiler, it’s also intercalation), Sony commercialised the first rechargeable lithium ion battery in 1991.

The crazy part is that LCO batteries are still used in smartphones and laptops today. The fundamental chemistry hasn’t changed.

A phone-powered EV?

Great, we have a stable and rechargeable lithium ion battery that’s very energy-dense. We can just make some more LCO phone batteries for EVs, right?

So we started replacing cobalt with a mixture of cobalt, nickel and manganese. These “diluted cobalt” cathodes are called NMC (Nickel, Manganese and Cobalt). Btw, the lithium is still in there, even if there’s not an L in the acronym. Don’t ask me why.

A star is born

It turns out that NMC lithium ion batteries are great for EVs. In fact, NMC batteries make up the majority of batteries in EVs today. The main reason for this is that they are fantastically energy dense - they carry a lot of energy for not much weight.

Up until 2015, NMC batteries had equal proportions of nickel, manganese and cobalt. This chemistry was referred to as NMC333 (meaning ~30% of each). But over time, we managed to reduce the cobalt content even more, to make chemistries like NMC532 (50% Nickel, 30% Manganese, 20% Cobalt), and nowadays, NMC811.

Whilst we’re here, you should know that NCA (Nickel Cobalt Aluminium) is a relative of NMC that uses aluminium instead of manganese. It’s a slightly lighter type of battery, but very similar to NMC. It’s been widely used by Tesla in the past, but its use is now declining.

Let’s take a look at this chart again:

We’ve got one last major type of chemistry to talk about, and that’s LFP. LFP is an old technology, and up until 2019 was dying out. However, recently it has staged a meteoric comeback that is only accelerating. This is where stuff gets really interesting.

Hi, I’m LFP and I’m a worse battery, but you’ll use me anyway 💅

LFP is lithium iron phosphate (LiFePO4, hence the acronym). LFP’s party trick is that it contains no cobalt, and no nickel - which are both expensive and geopolitically sensitive. They’re replaced by iron and phosphorus, which are abundant (and much cheaper). Iron doesn’t even make it onto the US’s chart of critical minerals.

On top of this, LFP has a much longer cycle life than NMC (the batteries degrade less over extended use). They’re also much safer!

There’s just one problem: LFP batteries are not very energy dense. To store the same amount of energy, they just weigh more. And as a battery for an EV, that’s kind of your one job.

For years, everyone had assumed that LFP’s low energy density would be a dealbreaker for EVs. In an industry obsessed with ever-increasing vehicle range, it was thought that LFP just wouldn’t be able to compete.

But in the last few years, four huge factors swung in LFP’s favour.

LFP-IS-COOL REASON 1: Safer cells = bigger cells

EV batteries are conventionally made of small cells that get assembled into modules, which get made into big complete battery packs.

Cells -> Modules -> Packs.

But because LFP is a much safer and stabler chemistry than NMC, it’s possible to produce cells that are waay bigger. If you make your cells big enough, you can skip the modules, and assemble your cells directly into a pack.

This means that you can do away with module packaging and other miscellaneous annoying shit. Your battery gets denser.

Doing away with the modules is called Cell To Pack (CTP or C2P). And although this is now happening for NMC too, LFP benefits much more because it's possible to make really huge cells.

So the punchline is this. NMC cells are way more energy dense than LFP, but by the time you’ve put them in a pack, the advantage is a lot smaller.

LFP-IS-COOL REASON 2: Patents

Previously, 95% of LFP production was in China. But in 2022, a key patent controlling LFP production outside China expired.

In early 2023, Ford announced a new $3.5bn battery plant in Michigan - to make LFP cells. Once complete it will mark the first ever large-scale US production of LFP EV batteries.

LFP-IS-COOL REASON 3: Nickel prices

Reminder: NMC is Nickel, Manganese and Cobalt. And it’s not just cobalt that’s an expensive part of NMC batteries - nickel prices have shot up too.

LFP-IS-COOL REASON 4: About that range...

In the west we love to obsess about high-performance EVs with bigger and bigger ranges. But for many people who live in the city, a short range does the job just fine.

A lot of Chinese EVs have used LFP for a while, because the Chinese city commuter car market (huge) doesn’t care so much about range.

The future of the lithium battery

The future is bright for the lithium battery. There are numerous highly plausible routes to improve its performance even further. I'm leaving these here in optional dropdowns, as it gets a bit nerdy 🤓.

Wow - we talked so much about lithium ion that we're two thirds of the way through the entire piece. Now let's talk about the contenders.

Hi, I’m sodium and I’m an even worse battery, but you’ll use me anyway 💅

Ok, let’s back up for a mo. LFP is eating NMC because it’s cheaper and made from abundant materials. Even if it’s a “worse” battery (heavier for the same amount of energy stored).

What if we took this idea even further? Well, then we arrive at the sodium ion battery, sometimes called Na-ion after sodium’s chemical symbol (Na).

Why sodium? Are we just picking random elements now? It actually turns out that there are fundamental physical reasons that some elements make great batteries. Lithium is unreasonably good at being a battery material. This is because:

1. Lithium is really light (it’s actually the lightest metal)

2. Lithium really likes to donate electrons.

It just so turns out that the second best battery material is sodium, which comes in quite a bit below lithium on paper, but not too far away in practise.

And whilst you can get rid of the cobalt and nickel from lithium batteries, you’ll, err, never get rid of the lithium.

Sodium ion has been under development for a while, but 2023 served up a healthy dose of this-is-actually-happening announcements. Eg

• BYD announced that its Seagull EV will have a sodium-powered version, sold for $11,600. BYD recently overtook Tesla as the biggest EV manufacturer in the world.
• Swedish battery giant Northvolt recently announced a 160 Wh / kg sodium ion battery that they will soon be mass-producing
• Natron Energy recently opened the first commercial scale sodium ion factory in the US.

Sodium: the numbers

CATL’s sodium battery is expected to cost 30% less per kWh than an LFP battery. That’s huge - but the battery is too. Here’s how much bigger & heavier our sodium battery packs would be than today’s LFP systems.

Basically, we are trading off mineral abundance (and therefore, cost) with energy density (and therefore, range).

One of the most interesting things I read about sodium ion was Huiling Zhou’s post for Powerhouse. She introduces it like this:

It strikes me that just a few years later, the rise of LFP may be repeating itself with sodium. The US & the west, obsessed with needlessly high-range EVs, plough ahead with high performance cells whilst China commercialises what is practical & low-cost. First LFP, and now sodium.

The last interesting thing to note about sodium is that it lives and dies by lithium prices. If lithium prices fall (from more supply coming online), then sodium’s trump card is significantly weakened. But if there’s a lithium supply crunch (like many are predicting), then sodium could really shine. In fact, it’s not difficult to use lithium kilns to process sodium. So if the lithium dries up, there might be a lot of repurposed infrastructure.

Part 2: grid storage

In part two of this guide, we’re leaving EVs behind and looking at batteries for grid storage. Don’t worry, part 2 is a lot shorter 🤠

Uhh, why can’t we just use lithium batteries for grid storage?

We can, and we do! Most grid-scale battery storage today is lithium-ion. We’re starting to see the very early stages of sodium-ion deployment here too.

But lithium batteries were originally designed for portable devices, not energy storage. They happen to be great at short-duration energy storage, because they can charge & discharge very rapidly. They’re not too expensive, and there’s also the potential to recycle old EV batteries into grid storage applications. In other words: it is very unlikely that lithium-ion will be displaced for short-duration energy storage (let’s say, up to 4 hours - but who knows where lithium’s economic cutoff will be in future).

But we also have a great need for long-duration energy storage. We need batteries that we can cycle less frequently, whilst remaining cost effective. These batteries don’t need to discharge rapidly, and they don’t need to be light. They just need to be cheap.

This is where our last two types of battery come in.

Iron-air batteries

Iron-air batteries are just one example of “metal-air” batteries - but they are of the most commercial interest.

We’ve known about iron-air batteries for ages, but no-one really bothered to build & scale them. Why? Because everyone was looking for lightweight & fast batteries.

That changed a few years ago, when Mateo Jaramillo co-founded a company called Form Energy, who set out to build a new battery specifically for multi-day energy storage. Mateo (who was previously head of stationary energy storage at Tesla) says this:

“I knew that nobody had ever tried to get to [low cost] by trading off those things...

In other words, I’m not sure anybody had ever said: this is the combination of things I’m looking for, what solves that problem?”

According to Form Energy, after a lot of experimentation, the answer to this question is iron-air batteries. When starting with the objective of low cost, the most important part is choosing abundant materials. This made iron (the cheapest metal) an obvious choice.

So here it is - I’m showing you how it works because it’s an unambiguously cool process. By rusting and “unrusting” a piece of iron, we can store electricity 🤯.

For now, the story of the iron-air battery is primarily the story of Form Energy. After years of R&D, they’ve begun construction of their first factory (it’s big), and have contracts to deliver gigawatt-hour scale projects. Incredibly, the fact that Form chose to build its factory in West Virginia may have been a key reason that the Inflation Reduction Act passed. Wild.

Form’s battery will be optimised for a duration of 100 hours, and they think that they can build it for less than $20 / kWh (lithium ion packs are yet to fall below $100 / kWh). This cost level seems to be feasible.

I’m excited to follow their progress 👀

Redox flow batteries

Out of all the batteries listed here, flow batteries have perhaps the most uncertain future. But they come up a lot, and they’re interesting because they take the concept of a battery in an entirely different direction.

A conventional battery is kind of an “all in one”. Energy is converted and stored within the same package. But a flow battery splits up the key parts. It has (1) a “storage” part of the battery, and (2) a “reactor” part of the battery.

Energy is stored in the tanks, in chemical form. When the battery needs to be charged or discharged, the chemicals are pumped to the reactor, where they interact to consume or generate electricity. One chemical loses an electron (oxidation), whilst the other gains an electron (reduction). This is why they are called redox flow batteries, short for reduction-oxidation.

The killer feature of flow batteries is that to make your battery last longer, you can just increase the size of your tanks & chemicals - you don’t have to increase the size of your reactor. To get technical, you can scale energy and power separately.

This makes flow batteries attractive for long duration energy storage, because scaling the capacity of your battery doesn’t mean scaling the whole cost. Unfortunately, flow batteries are quite expensive to start with.

The leading type of flow battery uses an element called vanadium to store energy in the tanks.

There are two things keeping people optimistic about flow batteries:

1. A lithium battery for grid storage will probably last less than 10 years, but a vanadium flow battery battery can run for 20-30 years, and sometimes longer.
2. Vanadium is not the only type of flow battery, and other electrolytes (chemicals in the tanks) are under development. There are countless combinations to explore.

Because of their simplicity, flow batteries are relatively easy to build as large scale experiments. There are iron flow batteries in Oregon, zinc-bromine batteries in California, and many more experiments worldwide. But with a cheap supply of vanadium, China leads the world with the largest installed flow batteries.

Against the backdrop of falling prices for lithium, sodium, and iron-air batteries, the future of flow batteries is less than certain. But in the wild west of batteries, nothing is off the table.

Thanks to Nicholas Yiu and Lisa Wang for reviewing this monster piece and giving their thoughts 🙏

Other writing

If you enjoyed reading this, you might like some of my other posts:

• The Energy Fundamentals of Carbon Removal
• All Decarbonisation Tech Has Just Three Inputs
  

But first, how do we determine the “start of decarbonisation”? We could measure the ambition of national targets, corporate commitments, or public sentiment. But these are indirect metrics. What really matters is how much clean shit we’re building. The end goal of policy, R&D and activism is to replace fossil fuel infrastructure with clean infrastructure.

The 1990s, 2000s, and 2010s laid important foundations (politically & technologically) for the climate movement. But there was little large-scale build out of clean infrastructure. Where the west did "reduce" emissions, it was often through transformations that were going to happen anyway (coal -> gas, efficiency improvements, exporting manufacturing). We developed and tested new climate tech, but we didn't deploy much of it.

In 2020, things started to change. We started deploying climate tech at a rate that surprised the most ambitious analysts. Because of this, between 2020 and 2022, the US, EU and China all announced net zero pledges. Meaningful or not, they were announced because for the first time, a zero-emissions economy seemed possible. (In 2010, when wholesale solar & battery prices were >5x what they are today, a net zero pledge would have been a sure-fire way to get voted out.)

Today, we’re still really off track. But since 2020, we've entered a new era where the rate of change is explosive. Let's call 2020 year zero in the climate calendar. BC = Before Climate. AD = Active Decarbonisation 😎.

Here are a few graphs showing that 2020 was the year decarbonisation started.

Charts

Decarbonisation starts with cheap renewables, because they are upstream of most other climate solutions. Even though electricity generation is only ~27% of our emissions, it will provide the energy to decarbonise most of the rest of them, through electrification. There is no path to zero emissions without wartime-style deployment of renewables.

Whilst most of these additions are in China, the US and the EU, the rest of the world is accelerating fast as well.

After remaining stagnant for years, only in 2020 did investment in fossil fuels and clean energy significantly diverge. Fossil fuel investment is still too high, but a threshold has been crossed.

Cheap renewables get firmed by cheap batteries.

EV sales have tripled since 2020.

In China, a quarter of new car sales are now EVs.

Heat pumps are having a moment too. Here’s heat pump sales in Europe:

and the US:

Green hydrogen is becoming investable.

Since 2020, oil and gas companies have quadrupled their spending on clean energy (even if it’s still a woefully tiny proportion of their investments). I hesitated about including this, but a tiny proportion of a large amount is still quite big (check the axis of the chart).

Zooming out, the overall investment in the energy transition has nearly doubled since 2020.

Onwards

This isn't a victory lap - we're still way off target. But electrification is a self-reinforcing spiral, which means that we're standing at the bottom of a giant exponential curve. Change happens slowly, then all at once.

(A) Clean electricity

(B) Biomass

(C) Carbon capture or carbon removal.

Whilst there are a plethora of different technologies for decarbonisation, when you boil them down, they are all dependent on one of these three ingredients.

Take hydrogen, for example. To produce clean hydrogen, you must make green hydrogen from renewable electricity (A), or blue hydrogen using carbon capture (C).

Or take the shipping industry. You might decide to fuel your ship using methanol. But to make the methanol, you’ll need clean electricity (A) and captured CO2 (C). Alternatively, you could make the methanol from biomass (B). Another option might be to keep burning bunker fuel, and to put carbon capture on your ship (C).

It's a bit like cooking with only three ingredients. You can make a lot of different meals with different combinations and skilful cooking. But whatever you make will always be comprised of the base ingredients. No matter which technology you choose, you will find yourself critically dependent on one of these three resources.

This is one of my favourite thinking lenses for analysing decarbonisation tech. At the end of this post I'll share how you can use this framework to ask great questions of any clean technology.

Are there really only three?

Here are some proposed climate tech solutions, with their dependencies mapped to our three ingredients.

At this point, you may be trying to think of exceptions to the rule. What about smart thermostats, or tech that reduces food waste? These fall under the category of demand reduction. Tech in this category is usually awesome low-hanging fruit with great financial incentives.

But once we’ve done all the demand reduction that we can, we will need to decarbonise everything that’s left. For everything that we want to continue doing, we’ll need to use one of the three ingredients.

OK - so decarbonising our existing processes depends on these three resources. Why is that important?

It's important because all three of our critical ingredients are limited.

To illustrate this, I’ve plotted three future Net Zero scenarios modelled by different organisations. Since we're measuring the use of three resources, they're plotted in 3D. The points represents how much of each resource is used annually in the year 2050. Compared to today, most future Net Zero scenarios assume that we use a lot more of these inputs.

Note that today, our total carbon capture capacity is 0.043 Gigatons / year, which doesn’t even take us off the axis.

Allocation matters

To return to our cooking analogy, it's not just that we only have three ingredients to cook with, it's that we have a limited amount of them in the cupboard.

You could use all your eggs (bioenergy) to make an omlette (jet fuel), but then you wouldn’t have any left over as ingredients for other things (chemicals, steel, etc).

Therefore, we have to carefully choose how we allocate the resources we have.

Limitations to biomass

Most energy experts immediately wince when biomass is mentioned as a solution, simply because our supply of sustainable biomass is so constrained.

Food production is the leading cause of deforestation, so dedicated biomass production that competes with food supply should be meticulously avoided. There is a small amount of waste biomass that we can use, but it will be under intense demand from a variety of sectors. Using it to decarbonise the chemicals industry alone would highly stress our supplies.

Limitations to carbon capture and carbon removal

Carbon capture and carbon removal are two separate and complex topics that are difficult to do justice to in one paragraph.

On the one hand, a certain amount of carbon capture and carbon removal will be essential to Net Zero, not least to provide the CO2 for sustainable chemicals, fuels, and plastics. Market commitments (both public and private) are setting the stage for tremendous growth in both.

On the other hand, the limitations of both carbon capture and carbon removal are infamous, and for good reason. Whilst plopping carbon capture on a cement plant conceptually sounds quite easy, in practise it is expensive and temperamental. Carbon capture has been a leading talking point since the very first COP in 1995, but has made very little progress since.

People love to talk about new announcements of CCS projects. But our actual global capacity is miniscule, and has barely grown in the last decade.

Regardless of your optimism on CCS and/or CDR, it's safe to assume that neither will contribute more than 5% of emissions reduction before 2040.

Limitations to clean electricity ..?

So, we are left with clean electricity.

Out of our three resources, clean electricity is the least limited one. It is also the only one currently undergoing exponential growth.

We will deploy solar, batteries and (to a lesser extent) wind, like nobody's business. In 2024, we are deploying a megawatt of solar per minute, and the plummeting cost of PV means that it will eat the world.

But energy infrastructure is deeply distributed in the real world. It is subject to planning, permitting, grid constraints, and everything that afflicts big projects.

PV and batteries will scale stupidly fast, but so will our demand. We will be moving towards abundant renewables by 2035, but they will not be unlimited in the short term.

Asking good questions

My favourite thing about the three-ingredient thinking lens is that you can use it to ask really good questions about decarbonisation tech.

If someone tells you that they’re making clean cement, clean jet fuel or clean plastics, you can ask them how much renewable electricity, biomass, or carbon capture they need. Usually, it’s quite a lot of one of the three.

It’s easy to get confused or distracted when someone explains their complex chemical or biological process for turning X into Y. But you can cut through a lot of complexity by asking questions like these:

If their solution depends on clean electricity (A):

• How many tonnes of CO2 do you save per kWh of clean electricity used?
• How does this compare to alternative uses for clean electricity? For example, displacing gas/coal generation?

If their solution depends on biomass (B):

• Where the heck are you getting your biomass from?
• Is it competing with food supply? Is it interrupting any natural carbon cycles?
• Is it competing with biomass supply for other climate solutions? If so, how many tonnes of CO2 can the other climate solutions save for the same amount of biomass?

If their solution uses carbon capture / carbon removal (C)

• For Direct Air Capture: how are you powering it? See questions from (A).
• For bio-based solutions: see questions from (B).
• For point-source capture & utilisation: how additional is this emissions reduction from utilisation? Isn't this emissions source going to be redundant or sequestered soon anyway?
• For point-source capture & sequestration: how reliable is your storage, what is your capture rate, & what is the uptime of your capture equipment?

Of course these questions are overly generalised, but they can be a good place to start.

In summary

The best climate tech does one of three things:

(1) Increases the abundance of one of these three resources

(2) Reduces consumption that would otherwise have needed to be decarbonised using one of these three resources

(3) Uses one of these three resources to decarbonise something really efficiently.

Conclusions

Let me know if you found this interesting, if you disagree, or if you just want to say hi 👋.

Lastly, if you enjoyed reading this, I'll be posting more soon. You can leave your email to get notified about new posts 💌.

Thanks to Prof. Julian Allwood for originally introducing me to this idea. And thanks to Archy de Berker for thoughtfully emphasising how important the tech is that increases the abundance of these three resources.

Sources are linked here.

Caveats

The broad exceptions to this rule are renewable sources of heat. Solar thermal for domestic use or industrial process heat is a sizable exception. So would be geothermal (heat only - not electricity).