Why Small Hydrogen Markets Are Likely to Shrink

Support CleanTechnica’s work through a Substack subscription or on Stripe.

Someone recently asked me about small, distributed hydrogen use cases and whether those markets might eventually be served by imported green methanol cracked onsite to produce hydrogen. The idea is not irrational. Hydrogen is difficult to transport and store. Methanol is a liquid fuel with existing global shipping infrastructure. Catalytic methanol cracking can produce hydrogen and carbon monoxide at the point of use. If a factory needs hydrogen, perhaps importing methanol and cracking it locally would provide a simple pathway to decarbonized hydrogen supply. The question prompted me to review the literature on smaller hydrogen markets because the scale and durability of those markets determine whether such a supply chain would ever make economic sense.

My just updated projection of hydrogen demand through 2100 based on my analysis around this question provides the broader context. The chart divides hydrogen demand into several major categories. These include refinery uses such as hydrodesulfurization and hydrocracking, ammonia production for fertilizer, methanol synthesis, steel production, transportation and heating applications, hydrogenation, and a category labeled mixed other. Today the global hydrogen market is about 95 to 100 million tons per year. Roughly two-thirds of that demand sits in two sectors. Oil refining and ammonia production dominate hydrogen consumption. The other categories are much smaller. When thinking about future hydrogen demand, the key question is whether those smaller uses grow or shrink as the century progresses.

A factor shaping the future of hydrogen demand is that hydrogen production today is itself a major climate problem. Almost all hydrogen is produced from fossil fuels, primarily through steam methane reforming of natural gas and coal gasification. These processes convert hydrocarbons into hydrogen while releasing large amounts of carbon dioxide as a byproduct. Global hydrogen production is roughly 120 million tons per year when all sources are counted, including hydrogen generated within refineries and chemical plants. Producing that volume of hydrogen releases on the order of 900 million to 1 billion tons of carbon dioxide annually, a climate impact roughly comparable to the entire global aviation sector. Before hydrogen can be considered a climate solution in any new application, the existing production system has to be decarbonized. That places the priority on replacing fossil based hydrogen in large industrial uses such as refining, ammonia production, and methanol synthesis rather than expanding hydrogen consumption into new markets.

The largest structural change in hydrogen demand will occur in oil refining. Hydrogen is required to remove sulfur and other impurities from petroleum products. It is also required to crack heavy hydrocarbon molecules into lighter fuels. These processes consume large quantities of hydrogen, but the amount required depends heavily on crude oil quality. Heavy sour crudes contain higher sulfur levels and larger hydrocarbon molecules. Refining them requires significantly more hydrogen than refining lighter crude oils. Barrels from Alberta oil sands, Venezuelan heavy crude, and similar deposits can require several times more hydrogen per barrel than light crude processed in simpler refineries. In my analysis, validated with a former Schlumberger engineer, heavy crude refining requires roughly 7.7 kg of hydrogen per barrel, while lighter crudes require around 1.5 to 2 kg per barrel. This matters because the global oil market is entering a phase of declining demand as electrification reduces gasoline and diesel consumption.

When oil demand declines, the highest cost barrels disappear first. Heavy sour crude sits at the expensive end of the refining spectrum. It requires more energy, more hydrogen, and more complex refining infrastructure. Further, refineries are being pressured to decarbonize processing which will require low-carbon, and hence much more expensive hydrogen. That creates a reinforcing dynamic for hydrogen demand. As oil demand declines, heavy crude loses market share first. The barrels that require the most hydrogen are the barrels that disappear from the market earliest. If global oil demand drops by 10%, hydrogen demand in refineries can drop by more than that because the most hydrogen intensive barrels disappear first. This dynamic is already visible in refinery closures that disproportionately affect facilities designed to process heavy crude. As that trend continues over the coming decades, refinery hydrogen demand will decline sharply.

Ammonia production is the other major hydrogen consumer. Hydrogen is combined with nitrogen through the Haber Bosch process to produce ammonia for fertilizer. Around 30 million tons of hydrogen are consumed globally in ammonia synthesis each year. Many decarbonization scenarios assume that hydrogen demand will grow because ammonia production will shift from natural gas derived hydrogen to electrolytic hydrogen produced from renewable electricity. That assumption deserves scrutiny. Agricultural productivity is improving through several parallel developments. Precision agriculture allows farmers to apply nutrients more accurately. Crop genetics and microbial soil treatments improve biological nitrogen fixation. Nutrient recovery from biomethane digestate and other waste streams adds nitrogen, phosphorus, and potassium back into agricultural systems. These developments reduce the amount of synthetic fertilizer required per hectare of farmland. Even if ammonia production decarbonizes, total demand may plateau or decline as agricultural efficiency improves. Green ammonia represents decarbonization of an existing hydrogen market rather than guaranteed expansion of it. Combined with peak population already having arrived in North America, Europe and China, with any population growth due to immigration, and global peak population likely between 2050 and 2070, demand curves for food aren’t going up exponentially as they were for the 20th Century.

Steel production is often cited as another major future hydrogen market. Hydrogen-based direct reduced iron processes are promoted as a pathway to green steel. However hydrogen direct reduction is only one option among several competing technologies. Electric arc furnaces using scrap steel already produce about 30% of global steel. That share can grow as recycling systems improve. Boston Metal’s molten oxide electrolysis, Fortescue’s electrochemical ironmaking processes, flash ironmaking, biochar blast furnace ironmaking, and biomethane-fueled direct reduction offer alternative pathways that do not rely on hydrogen as a primary reducing agent. Steel production will decarbonize through a mix of technologies rather than a single dominant approach. A rule of thumb I have is that if there are alternatives to hydrogen, they will be cheaper, simpler and more dominant in the market place, especially as the price of hydrogen inevitably increases with its decarbonization.

Once refining declines and growth assumptions around ammonia and steel are moderated, the remaining hydrogen demand outside of large chemical industries falls into smaller categories. In my projection these appear primarily as hydrogenation and mixed other. Hydrogenation includes chemical processes where hydrogen is added to organic molecules, such as in food processing, pharmaceuticals, and biofuel upgrading. Mixed other includes a wide range of industrial uses. These include metal annealing atmospheres, specialty metallurgy, electronics manufacturing, and various small scale chemical reactions. These uses are often geographically distributed and operate at relatively small scale. They are the markets that the methanol cracking proposal would theoretically serve.

Reviewing the literature reveals several consistent characteristics of these small hydrogen markets. The first is scale. Many facilities consume hydrogen in quantities measured in kilograms or tens of kilograms per day. Occasionally consumption reaches hundreds of kilograms per day. These facilities rarely produce hydrogen themselves. Instead they purchase hydrogen from industrial gas suppliers. Companies such as Air Liquide, Linde, and Air Products operate networks of hydrogen production plants and deliver compressed hydrogen by tube trailer to customers. A tube trailer typically carries a few hundred kilograms of hydrogen. Deliveries arrive periodically and hydrogen is stored in onsite cylinders or small tanks. This system has relatively high operating cost but very low capital cost for the user.

The second characteristic is cost sensitivity. Hydrogen is not the product of these facilities. It is simply an input to their processes. When the price of hydrogen rises, plant managers look for alternatives. Delivered hydrogen from industrial gas suppliers already costs significantly more than hydrogen produced on site in large industrial plants. If low carbon hydrogen produced through electrolysis or carbon capture pathways costs three to five times more than conventional hydrogen from natural gas, the price signal reaches small users quickly. These facilities cannot spread the cost of hydrogen across millions of tons of production. For many of them hydrogen is a convenience rather than a necessity.

The third characteristic is the availability of substitutes. Several industrial processes historically using hydrogen now operate with alternative technologies. Metal annealing provides a clear example. Hydrogen atmospheres prevent oxidation of metals during heat treatment, but they are not the only option. Vacuum annealing eliminates oxygen exposure without requiring hydrogen. Nitrogen and argon atmospheres provide inert environments for many metals. Modern heat treatment furnaces often operate without hydrogen entirely. Chemical hydrogenation also offers alternatives in some sectors. In food processing the hydrogenation of vegetable oils has declined because of health concerns related to trans fats. Transfer hydrogenation reactions using donor molecules such as isopropanol or formic acid can perform similar chemistry without requiring gaseous hydrogen. Pharmaceutical and specialty chemical producers frequently use these methods because they simplify safety requirements and reduce the need for high pressure hydrogen systems.

Hydrogen leakage introduces another factor influencing these markets. Hydrogen is now understood to be an indirect greenhouse gas. When released into the atmosphere hydrogen reacts with hydroxyl radicals, reducing the atmosphere’s ability to break down methane. This interaction increases methane lifetime and contributes to warming. Several recent studies estimate hydrogen’s global warming potential over a twenty-year timeframe at roughly 37 relative to carbon dioxide. That number is smaller over longer time horizons, but still a dozen times as potent as CO2 over 100 years. As regulators incorporate hydrogen leakage into lifecycle emissions accounting, hydrogen supply chains may face monitoring requirements similar to those applied to methane systems. For large centralized hydrogen plants such compliance will be manageable. For small distributed users it adds complexity and cost.

Distributed hydrogen supply chains also contain more potential leakage points than centralized systems. Hydrogen produced in a central plant may pass through compressors, storage tanks, tube trailers, pressure regulators, and distribution piping before reaching the end use process. Each stage introduces the possibility of small losses. Large industrial hydrogen users operate integrated systems with fewer handling stages and continuous monitoring. Regulators focusing on hydrogen leakage may therefore apply stricter oversight to distributed delivery systems. That additional regulatory burden reinforces the economic incentive to reduce hydrogen consumption where alternatives exist.

The green methanol cracking proposal must be evaluated in this context. Methanol can be produced from biomass and transported globally using existing shipping infrastructure. Catalytic reactors can decompose methanol into hydrogen and carbon monoxide. Installing such reactors at industrial facilities could in theory provide on site hydrogen supply without relying on compressed hydrogen deliveries. However the economics depend on scale. A facility consuming 20 kg of hydrogen per day would require a relatively small reactor but would still face capital costs, maintenance requirements, and operational oversight. For many facilities the simplest solution remains purchasing more expensive low-carbon hydrogen from industrial gas suppliers or avoiding hydrogen entirely.

What most firms will actually do is straightforward. Some will substitute alternative processes that eliminate hydrogen use. Some will reduce hydrogen consumption through recycling or process optimization. Others will continue purchasing hydrogen from industrial gas suppliers and allow those suppliers to manage regulatory compliance and infrastructure. Installing methanol cracking equipment will make sense in some cases, particularly where hydrogen consumption is larger and steady. In many other cases the complexity outweighs the benefits.

Biofuel processing deserves a brief proviso because it represents the only segment of hydrogenation that I expect to plausibly grow. Hydrotreated vegetable oil and sustainable aviation fuel pathways consume hydrogen during hydrodeoxygenation reactions that remove oxygen from bio based feedstocks. These processes can require tens of kilograms of hydrogen per ton of feedstock. Transfer hydrogenation methods cannot easily replace these reactions at industrial scale because the hydrogen donor molecules would need to be regenerated, reintroducing the hydrogen supply requirement. Other biofuel pathways such as alcohol to jet or methanol to jet may require less hydrogen, but hydrogen demand in biofuel upgrading will not disappear entirely. For that reason hydrogenation in biofuel processing remains a possible growth area, although its ultimate scale depends on which sustainable aviation fuel technologies dominate.

When revisiting the hydrogen demand projection with these considerations in mind, the implications for the hydrogenation and mixed other categories become clearer. Earlier versions of the projection assumed that hydrogenation would decline modestly while mixed other would remain roughly stable over time. The literature review suggests that assumption was optimistic. Rising hydrogen costs, substitution options, regulatory scrutiny related to leakage, and operational simplicity all push in the same direction. Many of the smaller hydrogen uses that exist today are likely to shrink gradually over the coming decades. Hydrogenation linked to biofuel processing may remain an area of demand, but the broader category of mixed other industrial uses will likely decline as industries adopt alternative processes. My scenario testing for price sensitivity at different price points for green hydrogen suggests 50% is a reasonable decline, with much of that front end loaded. My projection for hydrogen demand for biofuels hydrogenation remains the same, but as I dig through biomass pathways and demand projections for my updated maritime shipping and aviation scenarios, it may drop as well.

Across my analysis over the past several years I have consistently found that hydrogen is poorly suited to become a broad energy carrier because it performs badly when compared with direct electrification and biological fuels on efficiency, cost, and system complexity. Converting electricity into hydrogen through electrolysis, compressing or liquefying it, transporting it, storing it, and then converting it back into electricity or mechanical work wastes a large share of the original energy. Battery electric systems typically deliver 70% to 90% of input electricity to useful work, while hydrogen pathways often deliver closer to 30% to 40% once production, compression, distribution, and fuel cell conversion losses are included. Hydrogen infrastructure also requires expensive compression equipment, specialized pipelines, storage systems, and strict safety management. In most sectors direct grid connections, batteries, and biofuels provide simpler and more efficient solutions.

Hydrogen is also a poor solution for seasonal energy storage when compared with simpler options that already exist. Grid reliability emerges from layering solutions rather than relying on a single technology. Batteries handle seconds to hours of variability. Pumped hydro and flow batteries manage hours to a day. Interconnection, demand flexibility, and modest renewable overbuild smooth multi day fluctuations. Seasonal challenges are better addressed through two practical mechanisms. The first is seasonal thermal storage using aquifer thermal energy storage, district heating reservoirs, or large insulated hot water systems that store summer heat for winter use. The second is maintaining existing gas generation assets as strategic reserves fueled by biomethane stored in the natural gas infrastructure we already have. Biomethane can sit in underground gas storage for months or years and be burned in existing gas turbines during rare dunkelflaute events when wind and solar output fall for extended periods. This approach avoids building entirely new hydrogen infrastructure while using assets that already exist, providing reliability at far lower cost and complexity.

Another area where hydrogen is frequently proposed but performs poorly is commercial and industrial heating. Industrial heat accounts for roughly one quarter of global final energy demand and spans a wide temperature range from under 100°C for food processing and textiles to several hundred degrees for chemicals and manufacturing. In my work I have consistently argued that most of this demand can be electrified directly rather than replaced with hydrogen combustion. Electric boilers, resistance heating, induction heating, and industrial heat pumps already provide reliable solutions for a large share of industrial heat requirements. Industrial heat pumps in particular can deliver process heat up to around 200°C with coefficients of performance between three and five, meaning each kWh of electricity can deliver three to five kWh of heat. Combined with mid-depth aquifer thermal upgrading and thermal storage for load-shifting, this is an area rich with alternatives as well.

Electrification avoids the energy losses associated with producing hydrogen from electricity and then burning it for heat, a pathway that discards a large fraction of the original energy. When factories can connect directly to the grid and generate heat with electricity, hydrogen offers little advantage while introducing additional fuel handling costs, safety concerns, and infrastructure requirements. Studies and industrial pilots increasingly point in the same direction. Electrification technologies already exist for most commercial and industrial heating applications, leaving little room for hydrogen outside a narrow set of specific chemical processes that like flames, and even there biomethane is a cheaper alternative.

That’s why my projection has no hydrogen for transportation, energy storage or heating applications. Where an alternative to hydrogen exists, it is usually cheaper and simpler. Hydrogen handling requires specialized equipment, safety protocols, and regulatory oversight. That is why hydrogen has not become a widespread energy carrier despite decades of experimentation. It remains valuable where its chemistry is required, but many applications historically using hydrogen have already found alternatives. As hydrogen production decarbonizes and costs rise, that substitution trend will continue.