An extract from World Bank’s report, Green Hydrogen in Developing Countries

  • In 2019, the global hydrogen market was worth $135 billion, with over 70 million tonnes produced that year and 10 million tonnes of the demand coming from China.
  • Global electrolyzer manufacturing capacity is currently above 2 GW per annum, and it is forecast to exceed 4.5 GW on the basis of current expansion commitments.
  • Global fuel cell manufacturing capacity is currently about 1.5 GW per annum. Fuel cell demand has grown dramatically, to 1.6 GW stationary installed capacity and over 300,000 units deployed, most of which are based on proton exchange membrane (PEM) systems.
  • Similar to development of the lithium-ion battery market, PEM fuel cell manufacturing capacity and cost declines are being driven by the transport sector. However, the market focus for fuel cell in transport is moving toward larger uses—trucks, buses, trains, and ships—rather than light-duty cars.
  • In some contexts and geographies the production of green hydrogen could already be cost competive with fossil alternatives.
  • If hydrogen demand were to scale at the pace anticipated by analysts, the likely consequence would be to have a short-term increase in coal and natural gas demand for hydrogen production.
  • In the medium to long term, demand for fossil fuel–produced hydrogen will depend on whether expected electrolysis and renewable power cost declines are realized.
  • Key factors that will drive green hydrogen prices are the quality of the renewable resource, renewable and electrolyzer capital expenditures, and load factors that extend above 3,500 hours per annum.
  • Fuel cell technologies come with different costs, efficiencies, operating temperatures, and lifetimes. Therefore, capital exenditures alone will not always be the key driver of technology choice.

The global hydrogen market is valued at over $135.5 billion, with an estimated CAGR of 8 per cent until 2023. While exact figures for the volume of hydrogen produced vary, the literature suggests 55 million tonnes up to 70 million tonnes of hydrogen are produced annually. Around 96 per cent of global hydrogen production comes from fossil fuel sources, with 48 per cent from natural gas via SMR and 48 per cent from either coal gasification or other chemical processes (such as chlorine production). Only 4 per cent comes from electrolysis. Given the wide range of potential applications, determination of the potential market size for hydrogen is extremely contentious. Assessing this myriad of use cases, McKinsey & Company in a report for the Hydrogen Council (2018) determined that the future hydrogen energy market could amount to up to $2.5 trillion a year by 2050. It is worth noting though that other analysts have come up with significantly smaller figures.

Four of the most often cited growth scenarios for the hydrogen sector are from the Hydrogen Council (2017), Acil Allen (2018), IRENA (2018), and Shell’s Sky Scenario (2018). Although the studies do not compare like for like, they do illustrate the scale of hydrogen as a potential fuel source. Given that in their modeling Acil Allen, IRENA, and Shell do not appear to account for nonenergy uses of hydrogen (such as ammonia production and use as a process agent in refining), an assumption for the value of the global hydrogen market for chemical and process applications must also be made to arrive at a global hydrogen market demand in 2050. Consumption of hydrogen for energy and mobility is assumed to be below 1 per cent of total demand, so about 99 per cent of hydrogen demand today is for chemical and industrial process applications. This figure can then be subtracted from these estimates to produce an estimate of global hydrogen demand for energy applications.

Despite a wide variation in estimated annual hydrogen demand from the energy sector, it is clear that the demand for hydrogen in industrial applications will remain significantly larger than the market for hydrogen in the electric power sector. This observation not only illustrates how quickly the market could absorb additional hydrogen demand, but it also illustrates the potential scale of investment for green hydrogen to decarbonize existing hydrogen demand.

None of these scenarios provide an explicit split between what would be green hydrogen, gray hydrogen, and blue hydrogen at each milestone, so it is more difficult to provide an indication of how large the market for green hydrogen could be. However, current estimates suggest that between 0.7 million tonnes and 2.8 million tonnes of hydrogen come from electrolysis today.

The obvious questions raised by the growing interest in the potential use of hydrogen are where the hydrogen will come from and whether electrolysis companies will be able to meet the scale of demand at a cost-competitive level. Answering these questions requires obtaining clear data on the current size of the market today, which is very challenging.

For pure electrolysis solutions, current research estimates that global electrolyzer sales in 2017 reached 100 MW for the year, with 2018 sales estimated at $286 million for water-based electrolyzers and growth forecast at a CAGR of 5.67 percent during 2019–25. The most recent publicly available estimate for water-based electrolysis deployments globally is from the IEA, which estimates that the global installed-water electrolysis capacity today is over 100 MW electrical, while current orders for water electrolysis (alkaline and PEM) will see the installed capacity reach 285 MW by the end of 2020. By geography, Asia appears to be the world’s largest market for electrolyzers, with China assumed to have a market share of about 47 per cent. It is worth noting that China’s annual electrolyzer demand has grown from roughly 160 units in 2013 to 2,014 units by 2017, although no information is available on whether the average size of units has changed (GII 2018). If chlor-alkali electrolysis plants are included, then certain sources indicate global installed electrolysis capacity could be between 1.5 and 3.0 GW. Other higher-end estimates by BNEF have suggested that global installed capacity could be as high as 20 GW of electrolysis, but with only 2 GW installed since 2000.

Estimated global manufacturing capacity for electrolyzers (PEM and alkaline), 2019

Public sources estimate that 4 per cent of global hydrogen production comes from electrolysis, so green hydrogen will have to scale up significantly or CCU technology will have to grow rapidly, or both, to ensure that emissions from hydrogen production are reduced. To put this challenge in context, if the market demand for hydrogen in 2018 stood at 55 million tonnes until 2050 and current production from electrolysis was 2.2 million tonnes in 2018, then the electrolysis market would have to grow by 11 per cent annually to achieve 100 percent hydrogen generation from electrolysis by 2050. Thus, given the significant scale-up of green hydrogen required to decarbonize existing hydrogen demand, a number of analysts have become concerned that the development of a hydrogen market may perpetuate the use of natural gas reforming or coal gasification, albeit combined with CCU. The IEA, for example, notes that less than 0.4 million tonnes of hydrogen is produced with CCU and less than 0.1 million tonnes from renewables, amounts that illustrate the scale of the challenge to simply decarbonize the existing hydrogen market in the industrial sector.

Today, exact global electrolyzer manufacturing capacity figures are challenging to obtain. In early 2019 Nel, a market leader in the electrolysis market, claimed that the global market manufacturing capacity was around 90 MW per annum. Yet, this figure did not appear to reflect the considerable manufacturing capacity of electrolyzer companies in China and the existing capacity in Europe. Accordingly, ESMAP analysis suggests that current market capacity across alkaline and PEM is above 2.1 GW per annum and is scaling fast.

These figures are subject to a number of assumptions, but what is clear is that the global market is scaling rapidly. Two examples of electrolyzer manufacturing are Nel, which has committed to increasing its current manufacturing capacity 10-fold to 360 MW (with a 1 GW site identified in August 2019), and ITM Power, which also has publicly committed to a 10-fold increase in its manufacturing warehouse space. Thyssenkrupp also has publicly acknowledged it has 1 GW of alkaline electrolysis capacity, while John Cockerill also has confirmed its capacity increased to 350 MW as of Q4 2019.

Despite this growth, research conducted for this report indicates that under current practice a significant scale-up of hydrogen demand globally is still likely to increase short-term demand for fossil fuel–based hydrogen production, though the medium- to long-term outlook will be determined by the pace of growth, green hydrogen scale-up and the dynamics of hydrogen production pricing across the available technologies. Even taking a longer time horizon, current evidence suggests that to achieve the Hydrogen Council’s vision of hydrogen demand reaching 18 per cent of final energy demand by 2050 while minimizing the carbon footprint of hydrogen production, significant scale-up of green hydrogen will need to occur. The implication to policy makers considering hydrogen applications, especially in developing countries, is that it is essential to develop a clear roadmap for how the hydrogen can be sourced in a climate-sustainable manner to ensure that zero-emission sources of hydrogen production are scaled up.

Fuel Cells

Sources complied by ESMAP suggest that the global fuel cell market exceeds $2 billion per annum and that more than 2 GW of fuel cell systems have been shipped since 2000. To place this pace of deployment in context, note that the growth rate of various technologies in the five years after they reached 100 MW installed or shipped per annum indicates that fuel cells (portable, stationary, and mobility applications) have scaled faster than solar PV and onshore wind.

Technology deployment curves for fuel cells versus wind, and solar photovoltaic

The big drivers of this change have been the rapid improvements in the cost of fuel cell units, combined with improvements in efficiencies and in the operational lifetimes of fuel cell stacks. For example, the previous SOFC lifetime record was under 20,000 hours in 2005. Today, Bloom Energy, which produces SOFC technology, estimates that its systems can operate for 40,000 hours before the stacks may need replacement. Meanwhile, Ballard, which specializes in PEM fuel cells, has provided a warranty that its latest stationary fuel cell systems will perform for over 34,000 hours, while Doosan fuel cells have reported over 70,000 hours of operational lifetime for their PAFC units.

Increasing system efficiencies also has played an important role in improving the economics of fuel cell applications. These efficiencies include not only general improvements to the chemistry but also commercial innovations such as adapting units to provide CHP solutions in stationary contexts and recycling heat in vehicles to improve efficiency.

These incremental gains in efficiency and rapidly declining costs have therefore led some industry analysts to ask whether the market has sufficient manufacturing capacity to meet the rapid increases in demand. This is an important but complex question. Finding publicly available sources for global fuel cell manufacturing capacity is extremely challenging, especially because many companies are privately held and few public market reports provide granular information. Notwithstanding these challenges, by using publicly available sources and making reasonable assumptions, one can establish a “baseline” production capacity. Analysis conducted for this report suggests that significant capacity for PEM fuel cell manufacturing already exists and that plans are in place for a significant expansion. Other fuel cell technologies seem to be considerably more constrained, and that situation may lead to supply issues if these technologies experience a shortterm surge in demand.

Average fuel cell electrical efficiencies between 2005 and 2019

The ability to provide PEM fuel cell solutions to the mobility sector is a key driver of scaling capabilities and, in turn, will likely drive costs down and increase uptake, despite the potential for other fuel cell technologies to provide solutions with higher efficiencies and longer stack lifetimes. Further, because the MCFC, SOFC, and PAFC markets are dominated by one or two companies, the success of these applications will likely be far more closely tied to the companies’ own decisions than to the fate of the broader fuel cell industry. This may create another driver for consumers to focus on PEM technology, where they are more likely to be able to find alternative suppliers.


Hydrogen and Electrolyzers

The actual cost of hydrogen production and the price paid by most end consumers are difficult to determine because of the lack of publicly available data. Reports frequently state that hydrogen from electrolysis is more expensive than from steam methane reformation and from coal gasification. Yet that observation is far too simplistic. As noted by the EU’s Fuel Cells and Hydrogen Joint Undertaking (FCH JU in Fraile and others 2015), hydrogen prices can vary from EUR 1.5 per kg to EUR 60 per kg in the EU market, a range that remains true today, according to ESMAP’s discussions with current market participants. This range is also noted in other large hydrogen markets outside the EU.

The most commonly quoted figure for hydrogen pricing is derived from steam methane reforming technologies, which can produce hydrogen from natural gas at around $1.00–$1.50 per kg. But this price is typically linked to large-scale SMR units, with access to low-cost natural gas such as those in the United States and Northern Europe. Further, these costs often refer to already existing assets and do not reflect the costs of capital expenditure in their pricing. As shown in table 3.3, sources suggest that hydrogen production costs from SMR are closely correlated to moves in natural gas prices, with a recent study by WEC Netherlands observing that natural gas costs correspond to 70–80 percent of the total hydrogen production cost. According to this source, a EUR 6 per gigajoule price increase in natural gas corresponds to a hydrogen production cost increase of EUR 1 per kilogram.

Estimated global manufacturing capacity for fuel cells across all technologies, 2019

Nevertheless, these figures are in many senses limiting because they almost exclusively provide the cost only for consumers who are colocated and who represent the primary off-taker for the SMR operator. The reality for smaller-scale consumers is that hydrogen costs vary considerably, and calculating what the pricing should be of hydrogen delivered to customers is challenging. The simple reason is that transportation and storage of hydrogen remain expensive and cost varies depending on the volume of hydrogen demanded by customers and their distance from the hydrogen production site. How broad the spread can be is illustrated by the publicly available Hydrogen Demand and Resource Analysis tool (HyDRA) from the National Renewable Energy Laboratory (NREL).

Production cost estimates of hydrogen from steam methane reforming and coal gasification (excluding transport and storage costs)

The cost of delivering hydrogen to customers is considerably more than the production cost; thus there remains a significant cost advantage to being able to produce hydrogen on-site. This dynamic has underpinned much of the growing interest in electrolyzers, and it is worth noting that as early as 2015 the IEA, citing the US DOE, argued that the cost of distributed hydrogen production via electrolysis using off-peak electricity could be $3.90 per kg (IEA 2017). On this basis, hydrogen created from an electrolyzer on-site could be cheaper than the price of commercial forecourt hydrogen from SMR in all the locations shown in figure 3.8 except in Alaska.

Spread in United States hydrogen prices from HyDRA, April 2019

The question then is what drives the cost of hydrogen from electrolysis, and can these solutions consistently deliver green hydrogen at prices that are below those prices currently accessible to existing hydrogen consumers, excluding the largest captive producers (for whom there is no need to transport hydrogen because it is produced and consumed on-site). Historical analysis of the cost of hydrogen from electrolyzers has typically considered capex and electricity costs to be around 50:50 with regard to their impact on the cost of hydrogen. That is especially true for units below 1 MW. Yet, as the market has expanded and electrolyzer costs have fallen, companies have begun to argue that electricity costs now account for around 75 percent of hydrogen’s cost. Nevertheless, capex clearly is still an important consideration, especially at the smaller, decentralized scale. Thus, achieving load factors above 3,500 hours per annum is essential to securing low hydrogen prices, even when the cost of power might be low (or free when curtailed).

To illustrate some of the current green hydrogen cost estimates, table below consolidates assessments from a wide array of sources, under differing sets of assumptions. As the table shows, there is considerable variation in hydrogen pricing points. That variation reflects two sources of uncertainty: (a) considerable variation in electrolyzer cost assumptions and efficiency values and (b) considerations around the LCOE of electricity used and load profile/ capacity factor of the electrolyzer, reflecting the use of the asset. One source of discrepancy on cost assumptions is whether the capex figure given includes only the electrolyzer, or the balance of plant, installation, or both. Yet, even with these considerations, the pricing range remains large.

Cost estimates of hydrogen generated via water electrolysis

Fuel Cells

One of the best illustrations of the historic decline in the cost of fuel cell technologies comes from observing the cost declines for PEM fuel cell systems. That is because PEM systems have been developed since the 1950s and remain the most commonly procured fuel cell technology in aggregate across sectors (largely driven by mobility applications). It is important to note that costs for PEM fuel cells in mobility have always been lower than for stationary applications, owing to different stack lifetime requirements. Therefore, there are PEM fuel cells available today that suppliers will quote for below $2,000 per kW. Typically, the higher stack lifetime requirements for all stationary fuel cells lead to higher costs, part of the reason that stationary systems have a higher capex than mobility solutions.

Sample of electrolyzer capital expenditure estimates

PEM fuel cell suppliers are not the only ones seeing significant cost declines. Publicly available data also show that manufacturers across other fuel cell technologies are reporting significant cost declines as orders scale up. These companies are often well-established businesses, a fact that indicates not only the time it has taken to develop products that are commercially available to go to market, but also the importance of achieving scale to drive down costs and improve the economics of fuel cell projects.

But not all fuel cell system costs are starting from the same position. Fuel cell technologies have different applications, stack lifetimes, electrical efficiencies, and fuel sources that drive their overall cost structures.

Reported equipment cost decline curves from leading fuel cell suppliers

Assessing fuel cell costs, therefore, requires more than a simple analysis of capital expenditures. For example, PEM fuel cells rely on high-purity hydrogen, which is more expensive and more complicated to transport and store than other fuel alternatives such as ammonia, methanol, or natural gas. Further, it is important to consider the application. Where fuel costs are cheap or maintenance is a concern, there may be an incentive to switch toward longer-duration and lower-cost systems, such as PAFC or MCFC units instead of SOFCs. For customers who may value a CHP application, high-temperature fuel cells can provide a more compelling solution and a much higher system efficiency than lower-temperature fuel cells. For example, Transport for London installed a CHP (PAFC) unit in its Palestra building to help provide power as well as heating and cooling.

In short, choosing the most appropriate fuel cell solution requires companies to make a variety of assessments and judgments when scoping the technology solution and the application for the end customer. Capex alone is not always the key driver of technology choice.