Hydrogen for Energy Sucks
TL;DR
- Hydrogen is not (practically) an energy source, only an energy carrier, like a rechargeable battery.
- A carrier can only add losses. Turning electricity into hydrogen and back into electricity returns only 24–48% of what you put in. That makes hydrogen a poor choice for cars and energy storage.
- For sectors that can’t electrify, blue hydrogen (made from natural gas, with the CO₂ captured and stored) could work as a temporary bridge, but there is too little of it, it still emits, and it is too leaky for current infrastructure.
Free molecular hydrogen (H₂) is almost absent from Earth’s surface [1], so it has to be produced. For this reason, hydrogen works like a rechargeable battery: you can extract energy from it only after putting that same energy in. And an energy carrier can only add losses: it’s always more efficient to charge your phone straight from the socket than to charge a power bank first and then use that to charge the phone [4].
In particular, the round-trip efficiency of electricity → hydrogen → electricity is 24–48% [5]. This has consequences for applications that rely on that conversion chain:
- Cars. Hydrogen cars running on green hydrogen (hydrogen produced from renewable energy) go through exactly this cycle, so they are worse than using that electricity directly, as electric cars do. For the same distance travelled, hydrogen cars use roughly 2× the electricity [8].
- Energy Storage. If you store energy as hydrogen, you get back only 24–48% of what you put in.
What about blue hydrogen for non-electrical uses?
Most hydrogen today is made from natural gas [11], which is mostly methane, by reacting it with steam (CH₄ + 2H₂O → 4H₂ + CO₂). If the CO₂ is released, it’s called “grey” hydrogen. If it’s captured and stored, it’s “blue” hydrogen.
Why would you want that? Some sectors, like aviation, shipping and high-temperature industrial heat, still burn fossil fuels, and capturing CO₂ from every engine or chimney is impractical or impossible. Blue hydrogen could move the capture to a single plant, where the CO₂ comes out concentrated and is easier to capture. The hydrogen is then shipped to the users, who burn it and emit only water.
This could serve as a medium-term solution for the energy transition. The good news is that converting methane to hydrogen is fairly efficient (76–78%), even with carbon capture [12]. But there are three problems:
- Production is too low. The most optimistic scenario reaches 10 Mt/year by 2030, which would cover only 2% of the energy currently needed for aviation, shipping, steel and cement [13].
- Lingering emissions. Carbon capture is not 100% efficient, and methane leaks while it is transported to the hydrogen plant. Even with a high CO₂ capture rate (93%, commercially demonstrated) and a low methane leakage rate (0.2%), blue hydrogen would still cause 25–28% of the emissions of burning methane directly [14]. The reason is that leaked methane is a powerful greenhouse gas [15]. Real supply chains do worse: the average European natural gas supply had a methane emission rate of about 1.3% (2019) [16][17].
- Infrastructure. Hydrogen is highly flammable, like natural gas. Being the smallest molecule, it also diffuses through some metals and makes them crack [18], and it leaks easily. Leaks matter for the climate too: hydrogen is not a greenhouse gas itself, but it indirectly warms the planet by extending methane’s lifetime and increasing ozone and stratospheric water vapor [19]. A 1.5% H₂ leakage rate would contribute as much warming as a 0.2% methane leakage rate [20]. Building infrastructure that can handle all this is costly and slow.
N.B. This article exclusively discusses hydrogen for energy applications. In traditional ones (refining; ammonia, methanol, and iron production), this molecule continues to play a crucial role.
Notes
[1]Natural hydrogen does exist underground and is being actively explored, but only one site, in Mali, is producing, at village scale. Elsewhere, drilling has found high concentrations but no demonstrated commercial flow, so for now it remains a promising unknown, not an energy supply. [2][3]
[2]Zgonnik, V. (2020). The occurrence and geoscience of natural hydrogen: a comprehensive review. Earth-Science Reviews 203:103140. doi:10.1016/j.earscirev.2020.103140
[3]Royal Society (2025). Natural hydrogen report: royalsociety.org/news/2025/06/natural-hydrogen
[4]For typical lithium-ion battery power banks, around 85–95% of the initial energy is retrieved at the end. A. A. Kebede et al. (2022). A comprehensive review of stationary energy storage devices for large scale renewable energy sources grid integration. https://doi.org/10.1016/j.rser.2022.112213
[5]The conversion electricity → hydrogen is achieved by electrolysis of water with an efficiency of 60–80% (considering commercially available technologies) [6]. The opposite conversion, hydrogen → electricity, is realized by fuel cells with an efficiency of 40–60% [7]. The product of these two values gives the round-trip efficiency reported in the text.
[6]S. Shiva Kumar, V. Himabindu (2019). Hydrogen production by PEM water electrolysis — A review. (Table 1). https://doi.org/10.1016/j.mset.2019.03.002
[7]B. Tanç et al. (2019). Overview of the next quarter century vision of hydrogen fuel cell electric vehicles. (Table 1). https://doi.org/10.1016/j.ijhydene.2018.10.112
[8]For 100 miles, electric cars consume 24–45 kWh [9], while hydrogen cars consume 58–100 kWh [10].
[9]2024 Fuel Economy Guide. (pp. 41–42). https://www.fueleconomy.gov/feg/pdfs/guides/FEG2024.pdf
[10]Hydrogen cars consume between 1.4 and 1.8 kg of H₂ for 100 miles (www.fueleconomy.gov, 2023–2024 Fuel Cells). The electricity required to generate this H₂ is its energy content, 33 kWh/kg H₂ (lower heating value), divided by the efficiency of water electrolysis mentioned in note [5]: 60–80%.
[11]IEA (2023). Global Hydrogen Review 2023. https://www.iea.org/reports/global-hydrogen-review-2023
[12]C. Antonini et al. (2022). Hydrogen production from natural gas and biomethane with carbon capture and storage — A techno-environmental analysis. (Fig. 3). https://doi.org/10.1039/D0SE00222D
[13]The energy density of hydrogen is 120 MJ/kg (lower heating value). Therefore, 10 Mt of hydrogen correspond to 1.2 EJ (EJ = 10¹⁸ joules of energy). Iron and steel production requires 26 EJ (IEA, steel); shipping and aviation, 21 EJ (IEA, aviation and shipping); and cement, 9.5 EJ (in 2022, 4200 Mt of cement were produced, the clinker-to-cement ratio is 0.71, and the fossil fuel energy used per ton of clinker is 3.2 GJ; IEA, cement). Therefore, 10 Mt of hydrogen would only cover 1.2 EJ / 56.5 EJ ≈ 2% of the total energy consumed by these sectors.
[14]C. Bauer et al. (2022). On the climate impacts of blue hydrogen production. (Fig. 1). https://doi.org/10.1039/D1SE01508G
[15]The global warming potential (GWP) of a gas measures the intensity of its greenhouse effect relative to CO₂, whose GWP is set to 1 by definition. Methane-GWP is much higher: its value ranges between 84–87 when considering a 20-year timeframe and between 28–36 when considering a 100-year timeframe. The reason why it decreases with longer timeframe is that methane has a much shorter atmospheric lifetime than CO₂: 12 years against centuries. IEA, Methane Tracker 2021
[16]M. Bussa et al. (2021). Life cycle inventories for long-distance transport and distribution of natural gas.
[17]C. Meili et al. (2021). Life cycle inventories of crude oil and natural gas extraction.
[18]Hydrogen embrittlement (Wikipedia).
[19]Hydrogen’s global warming potential (see [15]) is approximately 12 over a 100-year period and 35–40 over a 20-year period. Maria Sand et al. (2023). A Multi-Model Assessment of the Global Warming Potential of Hydrogen. https://doi.org/10.1038/s43247-023-00857-8
[20]The energy content per kg of hydrogen / methane is 120 / 50 MJ (lower heating value). This tells us that 120 / (50 × 0.77) ≈ 3 kg of CH₄ are required to produce 1 kg of H₂. The 0.77 factor derives from taking into account the energy efficiency of the conversion methane → hydrogen, which is 76–78% [12]. Therefore, while an n% hydrogen-leakage rate causes a greenhouse effect proportional to (H₂-GWP) × n% per kg of H₂, an n% methane-leakage rate causes an effect of (CH₄-GWP) × 3n% for the same kg of H₂. Since, from notes [15][19], CH₄-GWP ≈ 2.5 × (H₂-GWP), a 1.5% hydrogen-leakage rate would contribute to emissions as much as a 0.2% methane-leakage rate.