SDAI Insights
Hydrogen's bottleneck is shifting from supply ambition to bankable demand
Low hydrogen production cost is not enough to make a project bankable.
Event date
Thesis
Low hydrogen production cost is not enough to make a project bankable. The stronger question is whether a real customer can use the hydrogen, at the required reliability and delivered price, under terms strong enough to support investment.
What changed
The IEA's Global Hydrogen Review 2026 reports global hydrogen demand above 100 Mt in 2025, while low-emissions hydrogen remained a small share of the market. At the same time, the low-emissions project pipeline contracted, final investment momentum slowed and only a minority of newly announced offtake volumes were backed by firm commitments. Electrolyser deployment continued to grow, and low-emissions hydrogen production was still expected to reach a new high in 2026. Taken together, those trends point to a familiar infrastructure problem: supply capability is advancing faster than committed demand.
Analysis
This is a useful correction to supply-side hydrogen TEA. Many early models ask: How cheaply can we produce a kilogram of hydrogen? That is necessary, but it is not the commercial question. A project developer ultimately needs to know: At what delivered hydrogen price will a defined customer switch, and can the project reliably supply hydrogen at or below that threshold? That brings the end use into the model. For a refinery, ammonia plant or industrial heat application, hydrogen competes against an incumbent route with its own fuel price, conversion efficiency, reliability requirements, capital needs and policy exposure. The customer does not buy an LCOH; it buys a production outcome.
Engineering implications
A decision model should connect hydrogen production with: - compression, storage and transport, - delivery reliability, - the incumbent fuel or feedstock, - end-use conversion efficiency, - carbon costs or incentives, - co-product value, - customer switching costs, - the strength and duration of offtake. LCOH then becomes one input to a broader project comparison rather than the final answer. A project with a slightly higher plant-gate LCOH may still be more viable if it is closer to demand, easier to deliver, more reliable or better matched to the customer's operating profile.
What to watch next
The existing PEM TEA work can be extended into an end-use-constrained hydrogen model for refining, ammonia and selected industrial applications. Instead of optimizing only USD/kg H2, the model should compare the delivered economics and emissions of the hydrogen route against the incumbent production route. That would put demand and customer economics inside the same decision model as electrolyser performance.
Primary source
IEA project-pipeline, offtake and production figures depend on the report's definitions and cut-off date. Announced capacity should not be treated as committed production, and announced offtake should not be treated as bankable demand unless the underlying commitment is firm.
IEA, Global Hydrogen Review 2026Related work
See the research and case studies behind SDAI's technical analysis.