Why CDR startups should deliver certifiable tonnes embarrassingly early
After 30-plus investments in CDR startups, one pattern keeps showing up. The companies that look most future-proof to me are not the ones with the biggest plant on the roadmap, the biggest financing rounds or the deepest stack of patents. They are the ones that got to their first certified CDR tonne early. Embarrassingly early. Often just a handful of tonnes.
The number did not matter. What mattered was that the team had been through the whole loop once: capture, measure, document, get audited, issue, sell. Everything after that is repetition with bigger numbers.
Two ways to forget the CO₂
The climate crisis puts a strange pressure on climate-purpose-driven founders (strategically and mentally): everyone knows gigatons are needed, and soon. Teams react in one of two ways.
The first group scales too early. The prototype works, the deck says gigatons, so the next step is the 10,000-tonne plant or project. Certification will get sorted out "once we are at scale".
The second group never leaves the prototype. One more design iteration, one more lab run, one more parameter to optimise. Four years in: a lot of knowledge, zero tonnes.
Both groups are not collecting any CO₂ that counts.
The software-startup-world (that I have spent mot of my life in) has an old rule for this: make your first 100 customers excited fans (not just customers), learn from them how to do it right, before you think about 100,000. In CDR, your product is a certified tonne, your first customer is whoever bought it, and the registry that issued it is the toughest reviewer you will ever have.
What a certificate teaches you
You cannot get a certificate by being roughly right. That is what makes it useful. How hard it is depends on the pathway. Mineralisation produces carbonate you can weigh; the open question is storage stability. Direct air capture produces CO₂ in a pipe, meterable to the kilogram; the hard part is the stable store at the other end. Biochar is measurable too. The open-system pathways are different: in enhanced weathering or marine CDR the CO₂ kind of disappears into soil water, rivers and the ocean, and the methods to count it are still being developed, sometimes contested. ERW certificates are being issued today, but still require a lot of buffer to solve the remaining uncertainties.
That buffer is the point. Every discount a registry applies tells you what to work on next. Build credible claims with a registry from day one and you learn within months where your measurement chain is weak. Scale first and you learn it years later, with a plant designed around assumptions nobody checked. "We will measure it later somehow" is not a plan.
The usual objections
The same four objections always come up.
"Our pilot emits more than it removes." Probably. Buy somebody else's certificates for your own emissions and draw the scope of your first tonnes after that. They will be few, but clean.
"Registry and auditor fees are absurd per tonne at this size." At this stage everything is absurd per tonne. Investors and pushing-the-limits-philanthropists pay for that; registry fees are part of the tuition.
"There is no accepted methodology for our pathway." Well, then you know what to actually work on and why I write this post. Where the methodology is still being written, being in the room is the strongest position a startup can have. It may even be the main reason for your existence.
"This distracts from the core R&D." This is the one I do not understand. Research that does not lead to a certifiable tonne may be good science. It is not a CDR company.
The north star
If I could put one sentence on the wall of every CDR startup, it would be this:
Make a plan, and build your systems in a way that will enable you to get even just a few certified tonnes of CDR as early as anyhow possible.
This is not "scale fast". It is almost the opposite: find the smallest version of your system that a registry will certify and a buyer will pay for, and build that one first. A minimum certifiable unit. It produces revenue, small but from the actual product; experience with the actual process, which no prototype gives you; and credibility with registries, buyers and the next investor. If the small unit works, it may even be a product in itself, and every certified module is both income and a field trial.
One side note: the certificate should be the north star, not the only income. Wherever possible, have other revenue as well, be it co-products, heat, building materials or services, whatever the process throws off. Only then can a startup sit out…. political rapids and delays in regulations.
In our portfolio, 55% of companies issue verified certificates. In July we called that our most important number: it counts teams that have closed the loop, not tonnes. It is also Prove & Learn, seen from the founder's side. Proving means proving to a registry, with a buyer attached.
Gigatons can wait. The first tonne cannot.
Gigatons are a scaling problem. The first tonne is a credibility problem. In CDR, the second comes first.
Founders: how early did you get your first certified tonne, and what did it teach you?
