Human Tech Tree
Unsolvedopen · Research Frontier · Today (unsolved as of Oct 2026)

Energy & Industry / Energy

Cheap Green Hydrogen

Hydrogen from renewable power still costs more than fossil hydrogen; low-emission hydrogen is under 1% of use and the 2030 project pipeline shrank.

Open in the interactive tree →

Hydrogen is a feedstock for fertiliser (Haber-Bosch), refineries and, in future, steel and chemicals. Today it comes almost entirely from natural gas and coal; green hydrogen is made by electrolysis from renewable power. It is expensive because electrolysers are costly, power must be cheap for many hours a year, and buyers shy away from the premium.

As of October 2026

According to the IEA the world used nearly 100 Mt of hydrogen in 2024, less than 1% of it low-emission; electrolysers cost $2,000-2,600/kW outside China and $600-1,200/kW in China. The IEA cut its 2030 pipeline of announced low-emission hydrogen from 49 to 37 Mt per year (electrolysis projects make up over 80% of the cut), and only 4.2 Mt per year have reached a final investment decision. Investment still rose to about $8 billion in 2025 (+80%) and installed electrolysis capacity was projected to approach 5 GW by the end of 2025, with China holding the largest share.

What is missing

  • Very cheap renewable electricity for many hours a year, so electrolysers run at high utilisation
  • Electrolysers that are cheaper, longer-lived and use less platinum and iridium
  • Committed buyers who pay the premium - quotas or CO2 prices
  • Transport and storage: pipelines, caverns, conversion to ammonia
  • Clear rules on what counts as green (for example the EU rules for renewable fuels)

Becomes possible once solved

  • Steel and chemicals without coal and natural gas
  • Fertiliser without fossil hydrogen
  • Synthetic aviation and shipping fuels
  • Seasonal storage of renewable power

Open steps

  • Iridium-free electrolyser anodes High AI leverageFind acid-stable oxygen-evolution catalysts without iridium that last tens of thousands of hours at industrial current.
  • Stack lifetime and degradation Medium AI leveragePredict and slow the degradation of membranes, catalysts and plates under start-stop operation to reach very long life.
  • Operation on variable power Medium AI leverageRun electrolysers efficiently on fluctuating wind and solar power, with sizing and control that maximize output per euro.
  • Cheap membranes and AEM stacks Medium AI leverageMake durable stacks without platinum-group metals and fluorinated membranes that compete with PEM and alkaline units.

Where AI could help

Medium AI leverage. AI speeds the search for cheaper, iridium-free catalysts, but green hydrogen cost is dominated by cheap power, utilization and buyers willing to pay.

  • Predict catalyst activity and stability to find low-iridium or iridium-free anode materials
  • Run autonomous labs that synthesize and test thousands of compositions
  • Optimize electrolyzer operation under variable renewable power using forecasts
  • Model membrane and stack degradation to extend lifetime

Shown so far

  • In September 2026 a preprint reported an AI-guided lab that synthesized and screened 2,942 catalysts and found iridium- and ruthenium-free palladium oxides, one holding overpotential below 0.5 V for 1,000 hours (not peer reviewed). source

Prerequisites

Unlocks

Sources

More in Energy · Research Frontier · Today

All 50 points in Energy →

Open in the interactive tree →