Electrolyzer Learning Rates Overstate Hydrogen Cost Declines
A new analysis argues that conventional experience curves, which forecast hydrogen cost declines based on cumulative installed capacity, are misleading.
Hydrogen cost forecasts that rely on conventional experience curves are likely overstating future price declines. This is the conclusion of a 2025 European study of electrolyzer project data, which found that raw learning rates conflate multiple cost-reduction mechanisms with genuine manufacturing experience.
According to the analysis published by TFIE Strategy Briefing and covered by CleanTechnica, the raw data from projects dating back to 2005 show impressive cost declines. For each doubling of cumulative installed capacity, costs fell by 23.3% across all projects, with rates of 32.1% for PEM electrolysis and 22.9% for alkaline electrolysis. These figures appear comparable to historical trends for solar and batteries. However, when researchers adjusted the data to account for economies of project scale, the learning rates dropped significantly.
Scale Economies Distort the Curve
After normalizing for project-size effects, the adjusted learning rates fell to 13.3% 17.6% for PEM, and just 7.3% for alkaline technology. The remaining cost relationship for alkaline electrolysis was no longer statistically significant. The study argues that the initial, steeper curves were attributing several different mechanisms to a single concept of "learning." Chemical-plant scale economies are substantial but front-loaded. A 100 MW hydrogen plant does not require one hundred times the equipment or engineering of a 1 MW plant. Larger projects share equipment and spread fixed costs across more output, leading to dramatic cost reductions as the industry moves from demonstration projects to industrial-scale facilities. Once key process equipment approaches a practical scale, however, further capacity gains increasingly come from replicating optimized designs, which yields smaller incremental savings.
The Problem of Stack Size Growth
The denominator of cumulative gigawatts becomes even more problematic when the physical size of electrolyzer stacks increases. If cumulative capacity grows tenfold but the average stack size also increases fivefold, the number of completed stacks only doubles. Factories gain manufacturing experience, but not at the rate implied by the growth in installed capacity. "During rapid equipment upscaling, installed megawatts can materially overstate the repetitions occurring at several important manufacturing levels," the analysis states. An experience curve based solely on cumulative gigawatts hides this changing physical architecture.
The System Boundary Constraint
A further constraint comes from the system boundary of a complete hydrogen plant. Based on an IEA 2025 electrolyzer cost breakdown, the stack itself represents only about 15-20% of the total installed capital cost. Balance-of-plant equipment like power electronics, compressors, and gas treatment accounts for 25-30%, while engineering, procurement, construction, and contingency can make up more than half. Consequently, even a significant cost reduction in the stack component has a muted effect on the total project cost. A 20% reduction in a component representing one-fifth of total capital expenditure removes only about 4% from the overall installed cost. Other project elements have their own, separate scale and productivity trajectories.
| Cost Component | Share of Total Installed CAPEX |
|---|---|
| Electrolyzer Stack | 15-20% |
| Balance-of-Plant Equipment | 25-30% |
| Engineering, Procurement, Construction & Contingency | >50% |
Implications for Hydrogen Forecasts
The analysis does not suggest electrolyzer projects will remain as expensive as today's first-generation installations. Manufacturers can improve stack designs, contractors can standardize layouts, and procurement can become more efficient. The mistake lies in collapsing all these mechanisms into a single historical learning rate per installed-capacity doubling and compounding it mechanically for decades. Several of the largest early cost reductions occur because the industry is making a one-time transition from small demonstrations to properly scaled industrial facilities.
Electrolyzer capital expenditure is only one part of the final hydrogen price. Electricity dominates the variable cost, and manufacturing learning cannot eliminate this expense. Very cheap but intermittent wind and solar power creates a tension with capital equipment that benefits from high utilization. Subsequent compression, storage, and distribution have their own infrastructure economics. Better equipment will reduce costs, but it cannot turn every part of the production chain into a mass-manufactured module with a solar-style experience curve. A defensible hydrogen forecast must separate manufacturing improvement, electrochemical performance, scale effects, and construction from electricity costs, utilization, financing, and logistics, modeling each on its own terms.
