Megaproject Lessons for Carbon Removal Deployment

Lessons from past megaproject successes and failures can help accelerate and derisk the massive deployment of carbon removal

How Big Things Get Done by Bent Flyvbjerg and Dan Gardner offers numerous insights that could be particularly valuable for deploying massive amounts of carbon removal. Intense planning, significant iteration, a focus on modularity, and highly experienced talent will all be necessary to help carbon removal avoid the unfortunate fate of many other massive projects and reach its needed scale.

---

Penguin Random House

---

I recently finished How Big Things Get Done by Bent Flyvbjerg and Dan Gardner. I highly recommend that everyone working in carbon removal, and climate tech more generally, go to your local bookstores and purchase a copy. For under $30, you can get advice that could easily be worth billions of dollars or more in saved megaproject costs.

Implementing gigaton-scale atmospheric carbon dioxide removal (CDR) is necessary to meet our shared climate goals, and projects moving this much mass will undoubtedly qualify as megaprojects. Rapid and low-cost execution is vital to ensure timely climate impacts and economic feasibility of large-scale removal. While CDR at this scale has not been done before, there are many general megaproject lessons shared throughout the book that can apply to huge CDR projects. I summarize a few of these here.

One of my main takeaways was the rarity of success paired with the sheer necessity of planning. According to the authors, only 0.5% of over 16,000 studied projects from around the world were delivered on time, on budget, and delivering the full extent of expected benefits. Carbon removal cannot really afford to not be in this category. The key differentiator for successful projects is slow and meticulous planning of every step. As the authors say throughout, think slow and act fast. Errors are cheap and generally fixable during planning but can be fatal during deployment. Drawn-out execution phases increase the chances of long-tail or black swan risks materializing, which risks extending the project and creating massive budget and schedule overruns.

While this lesson might seem intuitive, it is far, far too easy to fall into various psychological traps that lead to premature execution and eventual underdelivery or failure. The CDR ecosystem must continue to plan out every last aspect of deployment and mitigate as many identified risks as possible. Diverse teams are needed to discover and mitigate as many project risks as possible. In-depth plans should be made, checked, benchmarked against actual implementation timetables for comparable projects, checked again, criticized, and corrected, over and over again. This all must happen before a shovel ever goes into the ground.

Very careful planning looks a lot more active than one might think. Flyvbjerg and Gardner discuss how this can be done through a process they call Pixar planning. This is an intense iterative planning process used by creative powerhouses such as Pixar that involves making and perfecting increasingly detailed drafts until there is extremely high confidence in what will become the final product. Significant iteration, tinkering, and experiential learning are required to iron out as many issues as possible before making final investment decisions.

In carbon removal, this might look like creating hundreds or thousands of representative prototypes or doing small field trials of various system components to find and mitigate as many issues as possible. It could also look like continuously conducting deep supply chain mapping across the entire system. Some of this will happen naturally as we move from the lab scale to the pilot scale and then onward to the demonstration and industrial scales. This is where Silicon Valley banalities such as “fail fast” can actually be useful. Massive learning from trial-and-error requires lots of trials and of course lots of errors. It is far better to fail at a small, cheap, and inconsequential scale instead of a large, expensive, and reputation-damaging one. This is particularly important for hard tech companies where there are direct risks to worker and community safety. Once we work our way through this, scaling will be faster, safer, and smoother.

To help with rapid scaling and risk reduction, modularity is also highly beneficial. Flyvbjerg and Gardner discuss at length how more modular projects, such as solar power and energy transmission, are far less vulnerable to cost overruns and delays due to their modular or “Lego-like” nature. Complex, one-off projects, like nuclear power plants and the Olympic games, are in much more danger of massive cost overruns and a litany of other issues. Plenty of existing research supports this correlation between design simplicity and cost reduction, particularly as it applies to climate tech. The diagram below shows one technology typology sourced from Malhotra & Schmidt (2020) that breaks down this concept using some well-known climate technologies.

Malhotra & Schmidt (2020). This diagram shows a typology that classifies technologies based on their design complexity and need for customization. The authors demonstrate that simpler and more standardized technologies, like solar PV modules, experience faster learning and cost declines.

Na’im Merchant has written on how modularity can benefit CDR as has David Izikowitz. I have also written about learning-by-deployment for CDR and commented on its relationship to modularity. Certain types of carbon removal, such as direct air capture, already feature a significant degree of technological modularity in various system components. Pathways such as biochar have long taken place in a decentralized and more modular fashion, which has probably helped biochar be the most successful CDR method in terms of actual tons delivered based on current data from cdr.fyi.

Modularity has certainly become somewhat of a buzzword, and it is easy to claim that anything can be modular. The extent of inherent modularity for a technology might also be more of a predictor of success than a controllable variable. With this said, there are actionable insights related to these findings. Simplifying system designs, creating “scale-free” systems that operate similarly regardless of the scale of the operation, constantly updating designs based on learnings, and taking advantage of mass manufacturing of standardized components can all help expedite deployment, mitigate risk, and reduce costs.

To operationalize these insights, it helps to have experts with significant experience at the helm. Flyvbjerg and Gardner share multiple stories of the massive benefits of hiring individuals with significant relevant experience to complete megaprojects.

The trouble with CDR is that no one has done this before! There are people who have built airport terminals and skyscrapers but not megaton-scale mine tailing carbonation facilities, for example. However, we already have a sense of the types of talent that will be necessary to massively scale carbon removal.

To maximize the chances of success, CDR companies and their project developers must hire the best and most experienced engineers, scientists, CEOs, project managers, EPC firms, policy advocates, LCA/TEA analysts (😊), and others to get the job done right. We must also spend liberally to ensure these people have whatever they need to execute. This will be expensive but not nearly as expensive as overruns, delays, and failures. There is of course a benefit that comes with having some less experienced team members who can learn and provide unconventional insights—this is related to my point about diversity above—but empowering those with actual wins under their belts will help ensure projects are delivered at cost and on time.

Given my work on cost modeling, I was particularly interested in the section in Chapter 2 on strategic misrepresentation, which is when initial estimates purposely lowball expected costs to bait public or private investors into projects that will ultimately be far more expensive. Sadly, I have seen this in action multiple times, and it is one of my biggest pet peeves as it contributes to a breakdown of trust in cost modeling and the massive cost overruns we have come to expect from so many projects. Someone even said to me once, “We just need to show these [overly positive] numbers to convince investors to give us funds so we can go out, try this, and see if it actually works.” Without full risk disclosure, this is borderline fraud.

Strategic misrepresentation of costs is rampant in the carbon management world due to incomplete cost modeling and a desire to create excitement and secure scarce investment funds. If I had a removal offset for every time I have heard someone say their technology could allow for sub-$100/t direct air capture today, I could erase the emissions of a small country. To maintain the reputation of CDR and properly adjust expectations, we need to perform more accurate modeling of costs and benefits, collectively resist the temptation to oversell, and be honest and transparent with our findings. This is the only way to approach something that looks closer to truth, which is needed to build trust in removal offset markets and collectively identify and mitigate risks. Additionally, if we overpromise about cost performance today, we might not get the level of subsidization needed to help the industry scale.

If you enjoyed this post, I highly recommend reading How Big Things Get Done to learn about all the other insights and stories that could be useful to anyone working on massive carbon removal and other climate projects. If you have read this book and have your own thoughts about how its lessons might apply to CDR, please leave a comment or let me know at grant@carbonbasedconsulting.com!