By: Johannes Fiegenbaum on 7/29/25, 12:02 PM · Last updated October 5, 2026
Most climate tech lists rank niches by funding. That is useful to an investor and useless to a company that has to buy the technology. In the EU the buying moment is set by a rule with a date on it, and a market that was optional turns into procurement. This page sorts ten niches under €1 billion by the rule behind them.
A niche below €1 billion is too small for a dedicated fund and too specific for a general software vendor. What it has instead is a legal trigger: someone must comply by a date, and until then the demand is theoretical.
That changes the screening question. Instead of asking how fast a market grows, ask who is legally obliged to buy and when the obligation first bites. Narrow is fine. A niche of a few hundred million euros with a hard deadline converts better than a billion-euro market that runs on goodwill.
EU money follows that logic rather than setting it. The Innovation Fund, Horizon Europe and the national efficiency programmes pay for the supply side once a rule has created the demand side. Funding is the accelerator, so a funding chart is a poor map of where climate technologies sell.
My position: the date in the Official Journal forecasts better than any funding round. Where a date slipped in 2025 and 2026, the niche slipped with it.
One row per niche: the instrument that creates the obligation, when it first bites, and who sits on the buying side. Market sizes are deliberately absent, because published figures for niches this small are vendor estimates.
| Niche | EU instrument | First compliance wave | Who has to buy |
|---|---|---|---|
| CO₂ tracking for SMEs | CSRD as amended by the 2026 Omnibus directive | Financial years from 2027 | In-scope companies and their suppliers |
| Supply-chain ESG data | CSDDD, amended thresholds | Applies from 26 July 2029 (single date after Omnibus I) | Groups above the due diligence thresholds |
| Climate risk analytics | ESRS E1 resilience analysis | With the first report | Reporting companies and their lenders |
| Circular economy for building materials | Ecodesign regulation, Digital Product Passport | Delegated acts per product group | Manufacturers and importers |
| Energy efficiency in existing buildings | EPBD recast | Transposition, then minimum standards | Owners of the worst-performing stock |
| Low-carbon process heat | EED recast, audit and energy management duties | By consumption threshold | Industrial users above the threshold |
| Climate-friendly refrigerants | F-gas regulation | Stepwise bans | Operators replacing cooling equipment |
| Local hydrogen for industrial clusters | RED III industry sub-target | 2030 target year | Industrial hydrogen users |
| Biogenic CO₂ sinks and carbon farming | Carbon Removals Certification Framework | Once methodologies are adopted | Buyers of certified removals |
| Smart grids for EV charging | AFIR corridor targets | Staged to 2030 | Charge point and grid operators |
Four questions to screen any niche:
Live data: solar, heat pump and clean-tech price trajectories are in the Fiegenbaum Atlas dashboard.
Three of the ten niches sell the same thing in different packaging: a number that somebody else asked for. The obligation rarely sits with the buyer.
One pattern from my advisory work: a mechanical engineering supplier with roughly 250 employees in southern Germany, with no reporting duty of its own. Two customers in the first reporting wave asked for product-level emissions data with a deadline attached. It bought a carbon accounting tool, then found the tool was the cheap part and the supplier data it had never held was the expensive part. Its buying trigger was not the directive that applies to it, but the directive that applies to its customer. The VSME modules exist for exactly that position in the chain.
A second pattern: a logistics operator with about 600 employees. After a climate risk analysis along ISO 14091, it bought no analytics platform at all. It bought one flood assessment for two sites and took the result into a financing conversation. The lender was the actual customer of that analysis.
Adaptation demand works the same way, except the trigger is a measured figure rather than a date. For a vehicle dealer group in southern Germany with several sites, a site-level exposure model built on ERA5 reanalysis and EU-CORDEX projections showed substantial total exposure per year, with hail the leading risk at several of the sites. Nobody there bought an adaptation product because adaptation was fashionable. They bought roof and canopy work at the most exposed locations, in the order the model ranked them.
My position: the SME reporting niche is not a software market, it is a supplier-response market. Whoever wins here sells the answer, not the dashboard. The carbon border adjustment mechanism pulls the same figures out of suppliers outside the EU, which is why climate questions now reach IT procurement too.
Four niches share a different mechanism: the obligation attaches to a physical asset that already exists, so the market is bounded by the existing stock.
Under the recast buildings directive, member states must move the worst-performing stock first. That makes the addressable market countable: it is a percentile of an existing register, not a forecast. Energy efficiency retrofits sit there, and so does low-carbon process heat, driven by the audit and energy management duties in the recast energy efficiency directive, which catch sites above a consumption threshold rather than above an employee count.
Refrigerants are the cleanest case on the list. The F-gas regulation bans specific substances in specific equipment on specific dates. Every operator replacing a cooling system before those dates either buys a compliant system or buys it twice.
Building materials are the slowest of the four, because the Digital Product Passport arrives product group by product group through delegated acts. Reuse platforms have a real market only once the passport makes a second-hand component documentable. Until the delegated act for a group exists, that niche is a pilot business.
Three niches lost credibility between 2025 and 2026, and a niche list written before the deregulation debate is no longer honest.
What I stopped believing: that local hydrogen for industrial clusters would produce a supplier market this decade. The industry sub-target survives, final investment decisions do not, and a cluster project without an offtaker is a grant application. Carbon farming is the second. The certification framework is in force, but a removal without an adopted methodology is not a tradable product, and buyers have noticed. Smart grids for EV charging is the mildest case: the corridor targets hold, the vehicle numbers underlying the business cases did not.
None of the three is dead. All three are now bets on a date rather than on a technology, which needs different capital. I review the driver behind each of the ten rows every January and change the table when a date moves. Last reviewed September 2026, after the Commission's Omnibus package moved several of these dates.
Count the obliged entities in the first wave from the threshold in the legal text, then multiply by a realistic one-off spend. That bottom-up number is usually smaller than the published forecast, and it is the one you can defend.
Demand does not disappear, it loses its deadline, and pipelines stretch from months to years. Niches that survive a postponement have a second driver, usually a customer requirement or a financing condition.
The data and reporting niches need a supplier, not capital. Process heat, hydrogen and cooling are capital-intensive, with long payback and hardware risk. Building retrofit sits between them, because the capital belongs to the asset owner.
ESG and sustainability consultant based in Hamburg, specialised in VSME reporting, carbon accounting and climate risk analysis. Works with mid-sized companies, investors and financial institutions, and built the VSME reporting software VSEasy.
More aboutA green SaaS tool is a small piece of software that answers one climate question with live environmental data, and nothing else. Most sustainability software fails because it asks ...
Read more →