Stay ahead of the curve
Get exclusive insights delivered directly to your inbox. Join thousands of tech leaders.
TL;DR
- In 2001, NMC batteries read an improvement rate of 54% a year against 21% for hydrogen fuel cells (GetFocus backtest). Batteries stayed ahead at every reading to 2026.
- Automakers had spent more than $9 billion on fuel cells by 2014 (UC Davis). In 2023, about 14,000 fuel-cell vehicles were sold against more than 10 million battery electric vehicles (MIT Technology Review).
- The arguments for hydrogen were true on the day they were made. They described performance at that moment, and the improvement rate measures how fast that performance is changing.
- The reading is directional and covers passenger cars only. It shows which option is improving faster, and the timing still depends on cost, supply and regulation.
Hydrogen cars lost to battery cars because batteries were improving faster at every reading, and in 2001 the gap was already more than double. That year lithium NMC batteries read an improvement rate of 54% a year against 21% for hydrogen fuel cells in GetFocus's backtest of automotive powertrains, and by 2023 battery cars outsold hydrogen cars by about a thousand to one (MIT Technology Review, February 2024).
We build GetFocus, which measures how fast technologies improve from the way patents cite each other. The rates in this article come from that public backtest, and every public figure links to its source. The useful question for an R&D leader is why capable engineers at large carmakers stayed with the slower technology for two decades, and what would have told them sooner.
Where do hydrogen cars stand today?
In 2023, global sales of battery electric vehicles passed 10 million, while about 14,000 fuel-cell vehicles were sold worldwide, according to MIT Technology Review. SNE Research put the year at 14,451 fuel-cell vehicles, down 30.2% on 2022 (reported by Hydrogen Central).
That outcome followed a long and well-funded effort. A 2014 study from the Institute of Transportation Studies at UC Davis reported that automakers had spent more than $9 billion on fuel-cell development by then, with governments adding about $1 billion a year worldwide for hydrogen.
What did the improvement rate say, year by year?
The improvement rate is how much a technology's performance improves per year. It works like an interest rate: each year's gain builds on the last. GetFocus reads it from how patents build on each other, and the backtest computes every reading using only the patents that were public at that date, so each number is what a team could have seen at the time.
For the two technologies competing to replace the combustion engine in a passenger car, the backtest reads:
- 2001: NMC batteries 54.2% a year, hydrogen fuel cells 21.4%.
- 2008: NMC batteries 58.8%, hydrogen fuel cells 36.2%.
- 2014: NMC batteries 52.1%, hydrogen fuel cells 32.5%.
- 2026: NMC batteries 60.7%, hydrogen fuel cells 31.8%.
NMC is the nickel-manganese-cobalt family of lithium-ion batteries used in electric cars. The combustion engine, the technology both were trying to replace, read between 11% and 13% a year across the whole period.
Hydrogen did get better faster during the 2000s: its rate rose from 21% to 36% between 2001 and 2008, which helps explain why the people working on it felt real momentum. And batteries stayed ahead at every reading the backtest reports, from the first one in 2001 to the latest in 2026.
A small gap in a compounding rate becomes a large gap in performance. As an illustration only: if 54% and 21% a year each held for ten years, the first technology would end about 75 times better than it started and the second about 7 times better. The real rates moved from year to year, as the list shows, but the direction held throughout.
Why did carmakers stay with hydrogen cars for so long?
Carmakers stayed because the arguments for hydrogen were true when they were made. The backtest page lists the main ones:
- Energy density. Hydrogen carries roughly three times the energy of gasoline per kilogram, and batteries were far behind.
- Refuelling time. A fuel-cell car fills up in minutes, and an early electric car needed a night on the charger.
- The grid. Engineers doubted that electricity networks could serve millions of electric cars.
- Everyone else. Carmakers, oil companies and governments were all funding hydrogen.
The consensus behind hydrogen had public money in it: in 2003 the United States announced a $1.2 billion Hydrogen Fuel Initiative in the State of the Union address. In 2008, hydrogen fuel cells drew 2,896 new patent families, more than the combustion engine itself, according to the backtest. In January 2015, after launching the Mirai fuel-cell sedan, Toyota opened about 5,680 fuel-cell patents royalty-free to bring other companies in.
Jard van Ingen, co-founder and CEO of GetFocus, was at a European carmaker in 2018. In a GetFocus co-founder talk recorded in 2026, he recalls that "most of the people at [the carmaker] were still convinced that it was going to be hydrogen", and that they would come with arguments like this one:
"Oh, but battery vehicles are too heavy, so it's going to destroy all the infrastructure and the bridges will collapse."
Each of those arguments describes a technology on the day the argument is made. Weight, range, charging time and cost are all snapshots. The improvement rate measures how quickly the snapshot is changing, and a battery improving at more than 50% a year had an expiry date on every one of those objections. We wrote more about this pattern in how expert bias shapes R&D decisions.
What did the hydrogen bet cost?
The documented part of the bill is more than $9 billion of automaker spending on fuel-cell development by 2014, alongside about $1 billion a year of public funding worldwide for hydrogen of all kinds, cars included (UC Davis, August 2014). Fuel-cell programmes at Toyota, Hyundai, Honda, GM and Daimler ran for another decade after that, as the backtest page notes. We found no public source that totals the spending since 2014, so this article stays with the documented figure.
The larger cost is the head start. The backtest page puts it this way: a carmaker reading these rates in 2001 would have been a decade ahead of Tesla and BYD on the technology that won. A powertrain platform is a long commitment, so a late start is expensive to recover. Our article on when to move away from a core technology covers that timing question.
How does the improvement rate see what patent volume misses?
Patent volume shows how much effort an industry is putting into a technology. In 2026, hydrogen fuel cells still draw 2,841 new patent families a year against 504 for NMC batteries, according to the backtest. Read as a forecast, that volume pointed at hydrogen for 25 years.
The improvement rate looks at something else: how inventions build on each other. GetFocus forecasts it from two citation measures.
- Cycle time is the number of years between one generation of inventions and the next. In the backtest, NMC batteries ran at four to five years and the combustion engine at twelve to thirteen.
- Knowledge flow is how much later work builds on each invention. NMC's knowledge flow climbed to double the combustion engine's.
Think of two research groups. One is large and publishes a great deal. The other is smaller, and each of its results is picked up and built on within a few years. The second group's field moves faster, and that speed shows in the citations long before it shows in a product. Our guide to backtesting a forecasting method explains how to check a claim like this for yourself, and our technology forecasting software page shows how the same reading is produced for other fields.
What are the limits of this reading?
- It is directional. The improvement rate is an estimate from citation patterns and carries measurement error. A gap of 54% against 21% is a clear signal. Treat readings that sit close together with more caution.
- It gives a direction and leaves the date open. The 2001 reading favoured batteries, and battery cars approached a tenth of global new-car sales in 2021. Twenty years is a long time to hold a position, and factors such as cost, supply chains and regulation shape the timing.
- It covers one contest. This backtest compares technologies for passenger-car powertrains. Hydrogen in trucks, steelmaking or long-duration storage competes with different alternatives, and each case needs its own reading.
- The decision stays with people. A company can have sound business reasons to wait on a faster-improving technology. The rate tells it what it is waiting against.
How do you find the hydrogen in your own portfolio?
- Take one technical problem and list every technology competing to solve it, including the ones your team has already dismissed.
- Read the improvement rate of each, and look at how the readings have moved over the past ten years.
- Write down the arguments your organisation uses for its current choice, and mark which ones describe today's performance and which ones describe speed.
- Agree in advance on the reading that would make you reopen the decision, and check it on a schedule.
Your industry has a hydrogen somewhere. Send us one live technology decision and get a free 48-hour Decision Brief: a first-pass radar of the competing options, their improvement rates and suggested triggers.
FAQ
Are hydrogen cars the future?
For passenger cars, the evidence points to batteries. In 2023 about 14,000 fuel-cell vehicles were sold worldwide against more than 10 million battery electric vehicles (MIT Technology Review), and in GetFocus's backtest batteries have improved faster than fuel cells at every reading from 2001 to 2026.
How many hydrogen cars are sold each year?
About 14,000 fuel-cell vehicles were sold worldwide in 2023, according to MIT Technology Review. SNE Research counted 14,451, down 30.2% from 2022.
Why did carmakers invest in hydrogen cars?
Hydrogen stores far more energy per kilogram than a battery and a fuel-cell car refuels in minutes. Governments backed it too: the United States announced a $1.2 billion Hydrogen Fuel Initiative in 2003. Those arguments described the technologies as they stood, and batteries were improving faster at every reading from 2001.
Does this mean hydrogen has lost in trucks, steel or energy storage?
This backtest covers passenger-car powertrains only. Each other use has its own set of competing technologies, and each needs its own improvement-rate reading before anyone draws a conclusion.



