Please note: I wrote this in December 2022 and January 2023. I’m sharing it now (Aug 2023) as a way to share my thinking and so y’all can see what I was and wasn’t right about. I’d love your comments and feedback! 🙏 - Peter

<aside> 💡 I talk a lot about learning rates in this post. If you don’t know what they are, or just want a brief Econ101 recap, check out this section in the appendix. 😇

</aside>

Startups focusing on carbon dioxide removal (CDR) heard a familiar refrain in early 2022: “At high hundreds of dollars per ton, carbon is too expensive right now, but if you can produce at ‘X’ price, demand will go up exponentially.” Unfortunately, the target price of “X” fell as the year went by - keeping the elusive moment of exponential demand out of reach. Pricing data is scarce, because CDR contracts are private, but I feel comfortable saying that prices declined quickly. The going rate for DAC was quoted to cost $500 in Dec 2021, while rumors of $100 a ton carbon removal floated through the market by December 2022.

While prices fell, volume of bought and delivered CDR went up. But they didn’t go up as much as people had been hoping. According to our friends at cdr.fyi, over the past year, deliveries of durable carbon removal have a little more than doubled from 16K tons to 36K tons. Usually, as prices fall, demand goes up. But, in relative terms, it appears that prices went down much more than demand went up. Why?

If we assume the learning rate of the CDR industry similar to say, airplanes or solar, in that every of doubling of output results in a 20% reduction in costs, and we use deliveries as the basis for our learning curve, than CDR suppliers (and traditional market economics) could have reasonably expected prices (and underlying costs) to fall just 26%, from $500 in 2021 to $370 in 2022. Assuming a standard learning curve, if we wanted to achieve $200 a ton, we’d expect to see cumulative deliveries of 400,000 tons (6x more than we’re at now) or 3.5 million tons to get to $100 a ton (50+ times more than we’re at now).

Even if you use purchases of carbon, instead of deliveries, and use the same learning rate, you’d expect prices to be around $367 (excluding Airbus’s outlier agreement). You’d expect to see prices reach $200 a ton when purchases cross 2M tons (7x more than we’re at now) and $100 when purchases cross 17M (50+ times more than we’re at now).

**Check my math, please.**

If the price has come down so much, why isn’t demand going up? If demand hasn’t gone up that much, why is price coming down so fast?

CDR prices are falling comparatively faster than the cost of production.

My reasons for hypothesizing this are:

  1. Long-term corporate commitments give buyers overwhelming leverage. Now, if I’m the Chief Sustainability Officer of a public company, with serious neutrality commitments, I’ve probably got to buy some CDR. But, I was wise enough to make our promises due in 2035, so I don’t need to do anything this year. If I look at market prices over time and come to the conclusion that a ton of CDR is worth $250, or $200, or $150, then I can negotiate with CDR companies to make that a reality. If the market can’t make that work today, that’s fine, I can wait until it does, or sort it all out later. Today’s “market price” isn’t nearly important to determining sales (especially high-volume sales) as the expected future price is. Want to know why corporations (and banks) are buying lots of carbon from nature-based solutions? Because the expected future price is higher than today’s price. The opposite is true for CDR.

  2. Corporate buyers think the learning rate is very high. Thanks to strong leadership from Frontier and others, there are corporates willing to buy low volumes of CDR at VERY high prices ($1000+ per ton). Because these purchases are very early, and the companies and technologies are starting higher up the technical readiness curve, the learning rate appears to be very steep. If six months ago, carbon cost $1000, and now it costs $500, I might reasonably expect it to hit $250 in another six months.

  3. Most CDR companies are in no place to negotiate. **Why would a CDR startup burn cash closing a deal where costs > price? Well, because most CDR companies are doing expensive hard tech, and are marginal from a unit economics perspective. They need to go raise money to pay for salaries. Even in the case that unit economics are decent, CAPEX is necessary and expensive, meaning new facilities are dependent on raising VC funding. And, what every VC wants to see is product-market fit, namely increasing sales and revenue. So, if you’re a CDR company, it makes sense to boost revenue and lose a bit of money on a large volume deal, then make it back in subsequent funding at a high valuation.

    https://twitter.com/ElephantEating/status/1337671657415172097

  4. **Long delivery timelines mean CDR companies think they can drive down costs before delivery. ****Assume you’re a scientist/engineer in a DAC company’s lab. Your entire existence is predicated on the idea that you’ll make the CO2 machine work better over time. So, when the CEO says, can you figure out how to produce carbon for under $150 a ton today? You say “no.” But the answer to the same question on a five year timeline? There’s tons of stuff you can’t do tomorrow that you’re banking on being able to do in 5 years. So why not say yes to this kind of deal? And with credible major players pushing out delivery timelines to 5 years (as Climeworks does) these incentives aren’t likely to disappear.

  5. Plenty of VC in climate tech. **The dynamics above would simply slow down the growth of the CDR market, except that CDR companies are finding equity funding relatively easy to come by. Despite terrible macro economic trends, equity investing in climate tech remains a bright spot. About a quarter of all venture funding in 2022 went into climate.

    🌎 New dry powder for a new climate