The world is on the brink of what may turn out to be its most important energy experiment.
If proposals to build a new industry producing so-called green hydrogen succeed, we may have the final piece of the puzzle to prevent devastating climate change. If they fail, we may be about to spend hundreds of billions of dollars on a white elephant.
That's why there's both excitement and trepidation around the run of dramatic hydrogen announcements from Europe, Australia and Chile in recent months. The European Union alone envisages spending as much as 470 billion euros ($558 billion) on green hydrogen by 2050. To shift the whole world in the same direction would cost at least twice as much.
A viable green-hydrogen industry could power production of steel, cement and fertilizers; fuel trucks, trains, ships and aircraft; and balance wind- and solar-based power grids — and in the process eliminate roughly a quarter of the world's carbon dioxide emissions. Such a prospect would help decarbonize parts of the economy that wind and solar aren't well-placed to reach. It would also provide a potent new source of demand for the zero-carbon electricity that powers electrolyzer cells, splitting water into oxygen and green hydrogen.
The future, however, is uncertain. At present, such electrolytic hydrogen is barely more than a cottage industry. Most water-splitters are manufactured by hand, and 99% of the world's industrial hydrogen is not green but "gray," produced from gas or coal with the carbon emissions to match. The biggest producer of electrolyzers, Norway's Nel ASA, can make a modest 80 megawatts per year. To put the world on a path to zero emissions, we'll need to install two million megawatts or more.
The fact that such an expansive vision is seen as remotely viable is a tribute to the way renewables and lithium-ion batteries have transformed the energy industry over the past decade.
In the mid-2000s, even an advocate of climate action like British economist Nicholas Stern didn't think wind and solar could compete economically with fossil fuels until the 2030s. Things turned out very differently. Since 2009, the cost of unsubsidized solar power in the US has fallen 90% and wind is down 70%, notes Lazard Ltd. Battery prices have slumped 87% over a similar period, according to BloombergNEF. Coal is already in retreat from the power sector, and many of the world's biggest independent oil companies think petroleum demand is at or near its peak.
If green hydrogen can achieve renewable power-style cost declines from its current pricing of around $3 to $8 a kilogram, it stands a good chance of competing with gray hydrogen, which costs as little as $1. The risk, though, is that the forecast reductions aren't achieved. If a botched deployment or technical problems result in more modest economies of scale, the world will be left with a legacy of uneconomic hydrogen-production plants. On top of that, billions that could have been spent on other decarbonization technologies will have been wasted.
Which of those two futures we face will be determined by Wright's Law, a hypothesis about manufacturing dating from the early years of the aircraft industry. It states that with every doubling of cumulative output, the cost of technology tends to fall by a constant percentage. Factories get better at finding efficiencies; increased demand drives economies of scale; and larger volumes encourage suppliers to produce raw materials more cheaply. (The better-known Moore's Law, which predicted drastic declines in the cost of computing power, is best understood as a special case of Wright's Law.)
The cost-decline percentage, known as the learning rate, seems to explain why nascent renewable technologies can get so cheap so quickly. The learning rate for solar modules is a blistering 28.5%, according to BloombergNEF, meaning that an eightfold increase in installations will reduce costs by nearly two-thirds. That fall in prices then triggers more demand, encouraging further solar installations and reducing costs again in a virtuous circle. Fossil technologies can't compete with that advantage, because their largest cost is typically the fuel itself, where prices show no long-lasting downward trend.
Will green hydrogen follow the same path as solar, wind and batteries? There's good reason to think so. For one thing, nearly half of the cost of a green hydrogen plant would come from the renewable generators and batteries that provide power to the electrolyzer, and we already have plenty of data about learning rates there. Most of the other half is the electrolyzer itself, and BloombergNEF estimates these show improvements in the 18% to 20% range, comparable to lithium-ion batteries.
There are potential problems, though. The learning-rate hypothesis is treated as certain now, but just 15 years ago it was viewed with more doubt. The prices of raw materials can derail cost reductions for long periods — as we saw during the 2000s, when solar prices stood still for a decade thanks to a shortage of the polysilicon needed to make photovoltaic wafers. Comparable shortages of cobalt may yet derail projected price reductions for lithium-ion batteries. Current designs of PEM electrolyzers — one of the most promising technologies for producing green hydrogen — are highly exposed to the prices of platinum-group metals and Nafion, a synthetic membrane made by Chemours Co.
The larger issue may be that cost declines are highly dependent on the accuracy of estimates about learning rates and cumulative capacity, and there's still a dearth of solid data to produce that analysis.
Estimates of electrolyzer installations range from around 170 megawatts to 20,000 megawatts. In the former case, installing 100,000 megawatts of splitters over the coming decade would involve ten doublings in capacity. At an 18% learning rate, that should reduce costs by nearly 90%, making electrolyzers competitive with any fossil-fuel based alternative. In the latter case, we'll only see two or three doublings, which is unlikely to be enough. Learning rates themselves show substantial error bars, too. A technology with a 24% learning rate, at the upper end of electrolyzer estimates, will need to double capacity only four or five times to cut prices by half. With a 12% learning rate, at the lower end, you need to increase installations 50-fold.
That uncertainty could drive wildly different outcomes. If existing capacity is low and learning rates are high, green hydrogen may revolutionize energy as dramatically as wind, solar and batteries have done. If existing capacity is high and learning rates are low, investors might give up on the technology long before it's able to scale up.
Over the coming days, we're going to look at the potential shape of the new hydrogen economy, from the government policies that may direct it to the ways in which the existing fossil-fuel sector may hope to benefit. Energy investors, and anyone hoping that the world can avoid catastrophic climate change, had best hope that green hydrogen lives up to its promise.
Disclaimer: This article first appeared on bloomberg.com, and is published by special syndication arrangement.