There's no shortage of hydrogen project counts circulating in the market. The IEA maintains a production projects database. Consultancies publish annual pipeline reports. Conference slides cite ever-growing numbers of "announced projects." But anyone who works with this data seriously knows that the headline numbers obscure more than they reveal.
We've spent two years building what we believe is one of the most detailed hydrogen project databases available — not because we set out to count the most projects, but because we needed to structure the data properly to make it useful for actual decisions.
Here's what we've learned.
The counting problem
Most hydrogen project databases count "projects" as a single unit. But a hydrogen project is not a single thing. It's a collection of interrelated decisions: an electrolyzer procurement, a renewable energy supply agreement, a site selection, a permitting process, an offtake contract, an engineering study. Depending on how you define the boundary, the same physical development can appear as one project or five.
This isn't an academic distinction. When a client asks "how many hydrogen projects are under development in Germany?" the answer depends entirely on whether you're counting announced intentions, projects with site permits, projects with equipment orders, or projects with confirmed financing. The gap between these numbers can be a factor of ten.
Our approach is to track at the asset level and tag each with structured status indicators. A production facility is one entity. The electrolyzer procurement is a linked event. The renewable power supply agreement is another. This means our clients can filter by what actually matters to their decision — FID-stage projects only, projects with confirmed offtake, projects using specific electrolyzer technologies — rather than working from an undifferentiated list.
What we actually track
For each hydrogen project in our database, we structure data across several dimensions.
Project fundamentals include location (with geographic coordinates), announced capacity, technology type (alkaline, PEM, SOEC, or other), production color (green, blue, turquoise, pink), project phase, and timeline milestones from announcement through FID to commissioning.
Stakeholder mapping covers the full chain: developers, equity investors, debt providers, EPC contractors, technology suppliers, offtake partners. Each is linked as a canonical entity — so when you search for "projects involving Company X," you get every project where they appear in any role, not just projects they operate.
Supply chain linkages connect the project to upstream equipment (who supplies the electrolyzer, the balance-of-plant, the compression systems) and downstream consumption (industrial buyer, mobility application, ammonia synthesis, grid injection).
Source provenance means every data point links back to where we found it — a company press release, a regulatory filing, a government subsidy list, a trade publication — with a timestamp. This matters because hydrogen project data changes constantly, and knowing when something was announced versus when it was confirmed versus when it was quietly shelved is the difference between useful intelligence and noise.
The value chain gap most databases miss
Most hydrogen databases focus on production. That makes sense as a starting point, but it misses the broader picture.
The hydrogen economy is a value chain, not a single activity. Between a molecule of hydrogen being produced and it reaching an end user, there are compression, storage, transport, conversion, and distribution steps — each involving different companies, technologies, and investment decisions. A production-only database tells you where hydrogen might be made. It doesn't tell you whether there's infrastructure to move it, facilities to store it, or buyers ready to consume it.
We track across the full chain: production, logistics (pipelines, shipping, trucking), storage (salt caverns, above-ground tanks, geological formations), conversion (ammonia, methanol, synthetic fuels), and end-use (industrial consumption, mobility, power generation). The relationships between these elements are modeled as a graph, so you can trace from a production project to its expected downstream pathway.
What the pipeline actually tells us
Without publishing specific numbers that change weekly, a few structural observations stand out from our data.
The gap between announced and financed projects remains enormous. Across all geographies, the ratio of projects that have reached final investment decision to projects that have been announced but lack confirmed financing is roughly 1:8. This isn't unique to hydrogen — it's a pattern across all emerging infrastructure sectors — but it means that headline "gigawatts announced" figures dramatically overstate the near-term reality.
Equipment supply chains are becoming the binding constraint. For projects that do reach FID, the lead time from order to delivery for large-scale electrolyzers has been stretching, not shrinking. Tracking who has placed orders with which manufacturers, and when delivery is expected, has become as important as tracking the projects themselves.
Geography matters more than most analysis acknowledges. The economics of green hydrogen differ dramatically depending on renewable energy costs, water availability, existing infrastructure, regulatory frameworks, and proximity to demand. A project that works in the Arabian Peninsula may not work in Northern Europe at the same scale.
Why this matters for your work
If your organization makes decisions that depend on the state of the hydrogen market — whether you're an investor evaluating project opportunities, a manufacturer positioning equipment sales, a utility planning infrastructure, or a government agency designing subsidy frameworks — the quality of your underlying data directly determines the quality of your decisions.
We built this database because the existing options either lacked the depth we needed or presented numbers without the structure to make them actionable. Everything we track is available through the Delphi dashboard for visual exploration and the Serapis API for programmatic integration.