Since 2023, I've been assessing deeptech startups for Team for the Planet — a citizen-owned fund that invests in high-potential emission-reduction innovations, and that relies on the collective intelligence of its shareholders to sort the pipeline. In practice, dozens of assessors read the same projects, score independently, and the decision emerges from where the opinions converge. I go deeper on that role in my Team for the Planet case study.

My personal checklist sharpened file after file. It's not exhaustive, and it is not what decides in the end — collective intelligence decides. But it's my filter before I post a score. I'm sharing it because the same mistakes keep coming back, and a bit of clarity helps everyone, founders and assessors alike.

1. Technical feasibility: where are they really?

First question, and the worst-handled: does the science behind the project actually scale? Not in theory, not in a 2014 paper, but under industrial conditions today.

The tool I use constantly is the TRL (Technology Readiness Level) from 1 to 9:

  • TRL 1–3 — fundamental research, lab proof of concept. Investable, but on a 10+ year horizon.
  • TRL 4–6 — prototype validated in a representative environment. This is where 80% of climate deeptech risk plays out.
  • TRL 7–9 — pilot demonstration, first commercial unit, deployment. Much less technically risky, but often already expensive.

My trick question: "show me the data from your latest run in representative conditions, not in the lab". If the answer is fuzzy, you are not where the pitch claims you are. An honest startup will say "TRL 4, aiming for 5 within 18 months". A struggling one will say "we're almost industrial-ready".

2. Real impact: how many gigatonnes, on what horizon, and would it happen anyway?

A climate startup gets judged on three impact questions, not one:

  1. How much CO2 avoided or captured? In gigatonnes / year at maturity. If the answer is below 0.1 Gt CO2eq / year, this isn't a systemic lever — it can still be a good business, but it isn't the global climate stake.
  2. On what time horizon? A tonne avoided in 2027 is worth far more than a tonne avoided in 2045. Carbon is a cumulative problem. Be wary of curves that only take off after 2040.
  3. Additionality. Would this reduction happen without them? If the same outcome is coming via EU regulation in 2028, the startup's marginal contribution is thin. This criterion knocks out a lot of "me-too" projects that only accelerate a trajectory already in motion.

"Gigatonne thinking" is the opposite of the humblebrag about "tonnes of CO2 equivalent saved by offsetting a flight". If we're not playing at gigatonne scale, we're playing a different game.

3. A business model that holds up without carbon credits

Recurring red flag: a startup whose P&L only works at 100 €/t carbon. The carbon market is unstable, politically dependent, and historically disappointing. The best climate companies have unit economics that work even without carbon credits — credits are a bonus, not the spine.

My three business-model questions:

  • Who pays, and why would they pay a lot?
  • What is the gross margin excluding subsidies and carbon credits?
  • Is the final product price competitive with the fossil alternative today, or only with a 150 € carbon tax?

4. Red flags

Signals that almost automatically lower a score for me:

  • "We use AI to optimize…" with no specifics on what, how, or with what data. AI is not a climate thesis. It's a tool.
  • No scientific co-founder on a deeptech project. If the subject is chemistry or biology and the pitch is carried by two business profiles, technical execution is at maximum risk.
  • A "solution" to a problem nobody pays for today. Eco-consciousness doesn't pay B2B invoices. The customer needs an economic or regulatory reason to buy.
  • No measurable impact figure, or top-down numbers like "if we capture 1% of the global cement market, we avoid 8 Gt". Without a bottom-up build, that's projection.
  • Hand-wavy patents, sponsored articles without peer review, no publications at all.

5. Green flags

On the flip side, what raises a score for me:

  • Recent peer-reviewed publications from the founders or their scientific board. Nature, Science, Joule, ACS — any of them, as long as it's real.
  • An identifiable Scientific Advisory Board, with named researchers you can reach, not a wall of university logos.
  • An MVP with a paying pilot customer, even a small one. Early commercial validation beats ten TAM slides.
  • A clear regulatory roadmap: they know which authorizations they need, in what order, and who they've already met on the regulator side.
  • Honesty about risk. The best pitch I've ever read opened with "here are the three reasons this company might fail". I gave it the highest score.

6. The checklist, tested against reality

A grid is worth what it produces. Here are a few innovations Team for the Planet has funded — public files, public numbers — re-read through the criteria above. That's the best test of a method: does it light up the projects that, two or three years later, are actually moving?

  • Beyond the Sea (~€2.5M) — kite-type traction sails for cargo ships, targeting −20 to −40% fuel. A textbook case of the "systemic impact" criterion: maritime shipping is nearly 3% of global emissions, so the lever is massive by construction. Where they are: a 50 m² sail in dynamic flight tests in the Arcachon Basin — a concrete step from prototype to demonstration (TRL 5-6 → 7).
  • Leviathan Dynamics (~€2.34M) — industrial cooling produced with water, replacing ultra-warming refrigerants. A textbook case of "works without carbon credits": the customer buys cheaper cooling and sheds a regulatory risk on fluorinated gases — the climate impact is a bonus, not the spine. Where they are: first machines shipped, ~2,000 t CO₂ avoided over the past year, moving into industrialization.
  • Cool Roof France — a reflective paint that cuts a building's cooling needs by about 40%. A textbook case of "feasibility": mature tech (TRL 8-9), immediately deployable, no 10-year science bet. Where they are: hundreds of thousands of m² of roofs repainted across France, in full growth.
  • Seaturns (~€0.8M, wave energy) and Monomeris (near-infinite plastic recycling) — two earlier-stage bets, closer to TRL 4-6: technically riskier, but with huge tipping potential if they clear the industrial step. Exactly the "high risk / high impact" profile the citizen-owned model can afford where a classic fund would flinch.

Across ~30,000 innovations assessed and 14 deployed, what strikes me is the consistency: the projects moving fastest are precisely the ones that ticked "systemic impact", "viable without carbon credits" and "honest about TRL". The grid isn't magic — but it ages well.

To close

This checklist is my filter — not the final algorithm. What makes the Team for the Planet model strong is precisely that we do not rely on a single assessor. I can be wrong about a file; the collective wisdom of dozens of independent assessors corrects my biases. That mechanism, more than my own score, is what makes the system robust.

If you're building a climate company and want to stress-test your file before submitting — or you're another assessor and want to compare methods — email me at contact@thanaelfontaine.eu or book a slot from the contact page. My filter is the entry door; collective wisdom does the rest.