Learning #2: Can we turn shipping containers into carbon capture factories?

After growing (and exiting) 2 software start-ups, I am looking at what to do next. Since Feb 21, I have been trying to learn more about the climate crisis — first by speaking with scientists and entrepreneurs in the space (See here and here), then by running real world experiments to figure out what to focus on. Here is what I learned in my second experiment.

Some context on this new experiment:

• The scientific community has known for a while that certain types of rocks — silicates, such as basalt or olivine — naturally react with the CO2 in the air and turn it permanently into a rock. This is called Mineralization. Even better, these rocks are abundant. When scientists try to compute the carbon capture potential of these rocks, numbers get crazy quickly.
• The reaction happens naturally and in ambient conditions but it takes literally hundreds of thousands of years. So the main question is: how to speed things up? There are several ways, for instance increasing the surface of rocks exposed to the air, optimising the temperature / humidity or increasing the concentration of ambient CO2.

Carbon capture factory

• Several awesome projects have been launched, each taking a spin on the concept: Project Vesta relies on sea waves to grind olivine (removing the need to use energy to grind the rocks and helping alkalinise the ocean), Heirloom grinds wollastonite and uses high heat to capture several rounds of carbon with the same batch of rocks, Future Forest grinds basalt and spread it in forests to accelerate the formation of carbonates (plants accelerates the rate of weathering), etc.
• While all these projects are really exciting bets, it really feels like they all come with big long term tradeoffs: Vesta and Future Forest spread the minerals — and therefore the carbon — in live ecosystems, making it very hard to measure precisely how much carbon was captured permanently (And in that sense sharing some of the issues of our previous experiment). Heirloom requires a huge amount of industrial heat to cycle the material, which asks big questions about where such a large amount of heat will come from and whether it can be credibly decarbonised.
• And then I came across Corey Myers’ paper, outlining a process of dry mineralization in contained environment (heat and humidity control), where the carbon is not spread in the environment (easy verification) and where there is no need for industrial heat (modest energy need). The drawback is that it takes space, a LOT OF SPACE, tens or hundreds of square kilometres. But after all, if there is one thing we have in spades, it is space.
• So, after several informal chats with (the awesome) Corey, an experiment took form: how do we prove the viability of Corey’s dry contained process and get it off the ground?

Here are the hypotheses we wanted to validate:

• There is a reasonably small size of unit where unit economics start making sense — even if we only monetise by selling high quality carbon permits
• In the current market, it is possible to get funding for an R&D-heavy hardware company on an unproven market

In order to validate these hypotheses, we started designing the process to fit inside a recycled 40ft container:

• Organising a system of shelves to maximise the packing density of material per m3 while still allowing optimal air flows.
• Set-up various sensors and a software layer to monitor and optimise inputs to drive an optimal rate of weathering.

You get the gist of it

After a few months iterating on the idea, a few things started to stand out:

• Companies specialised in retrofitting containers for industrial uses were convinced that such a unit could be produced with off-the-shelf materials for low 5-digit $.
• As a result, the financing need to build a first container and operate it for a year remains low (high 6-digit $), especially considering the kind of money being raised in venture capital markets at the moment, in particular as…
• Investors are now used to hearing about containerised projects. Folks in the vertical farming industry have taught investors the benefits of using small-size modular factories and there are established models about how it scales and when to move to custom premisses.
• The output product — a carbonate rock — can be tuned to be used as a worthy aggregate to make concrete, a bit like what Solidia is doing. This process is effectively transforming spare rocks into something valuable, capturing carbon in the process.
• More importantly, starting from first principles (i.e. physics) it looks like each individual container would have positive gross margin, albeit small, making it possible to scale things in a modular way. Using Wollastonite, each batch of material would fully weather within 5 weeks, allowing 9 batches per year per container.

As sexy as it gets

• 1st principles suggest other processes could have better unit economics, but each comes with larger risks: using copper tailings lower the cost of acquiring material, but increases uncertainty as there are few precise measurement of its rate of weathering and therefore, potential. Using a pressurised container and increasing the CO2 concentration would require to get the CO2 in the first place, turning the project into effectively a storage solution, at a time where there are solutions to do just that for less than $30 a ton (Carbon8 seems also to be doing something very similar already)
• The dry process in ambient conditions has the benefit of being pure Direct Air Capture AND to keep the carbon in one place AND to have a low energy requirement. It felt like a home run.

But not so fast. Through various rounds of iterations and validation, questions emerged:

• How much of the process is really defensible? Assuming we overcome the engineering challenge of loading and offloading, and validate the underlying carbon economics, could it all be reproduced by anyone who merely saw the unit? Carbon capture projects who get funding tend to be based on complex industrial processes with the capacity to have patents, or at the very least industrial secrets that could act as a moat. What would our moat be here?
• Along the same lines, this concept is about taking large amounts of rocks, finding the best way to grind it, tailoring the conditions to maximise weathering and bringing the output somewhere useful. It is a mining project at heart. How defensible can it be once the big mining companies realise they can make large amounts of money from their tailings? And I recently learned from primary sources that they are already quietly working on doing just that.
• The best material for weathering tend to be material that are valuable in their own right — for instance Wollastonite which costs circa $20 a ton. Worse, the most valuable rocks have small markets that would struggle supporting carbon capture at Megaton scale, let alone at Gigaton scale. Waste materials — tailings — would be a lot cheaper but have limited data on how reactive they are.
• Last but certainly not least, is the general lack of enthusiasm from geologists and engineers for this concept. Granted, I may not have spoken to a statistically significant sample, but many tech folks seem to be put off by the amount of space required to reach scale. It feels to me that a “LOW energy EASY verification and HIGH space” concept fills an interesting niche in the CDR landscape, but technical experts do not seem to agree. Not being a scientist, I am unsure why. DM me for thoughts or feedback!

Where do we go from here?

I remain very bullish on direct-air-capture mineralization and continue to believe it is a very exciting pathway to investigate. Yet, 10 months in, it is hard to escape the conclusion that this is also a science-heavy pathway with big science and engineering risks even assuming the market risk gets sorted out. Embarking on such a journey without credible technical excitement feels premature so the right next step is to find a potential tech partner with an appetite for risk, a compelling point of view on mineralization and a desire to dive into CDR. Recommendations and DM welcome!