Batteries Are Advancing According to Their Own Little-Known Moore’s Law
Batteries, say deeply frustrated developers, follow no Moore’s Law. Unlike semiconductors, whose performance has doubled roughly every two years for at least a half-century, just as Intel co-founder Gordon Moore posited, batteries have merely trundled along, eking out advances in crawling fits and starts, seemingly lost to any governing master force. All there is to do is keep working on the chemistry, and hope for the best.
But, as slowly cheapening lithium-ion batteries finally seem close to taking electric vehicles down to the same sticker price as combustion, a couple of industry bulls are taking issue with the industry orthodoxy: The cost decline, they say, has been foreseeable. The lens is an industrial principle just as ironclad as Moore’s Law.
Go on Google, punch in T.P. Wright, and you will find a multitude of links to Wright’s Law. Starting in the 1920s, Wright was an aerospace engineer and later an official in the FDR and Truman presidencies. In 1936, he wrote a seven-page paper in the Journal of the Aeronautical Sciences in which he sought answers to a couple of questions: What impact precisely does scale have on the price of an airplane? And, can that impact be quantified? The result was his eponymous law, which states that when the cumulative production of a given thing doubles, its cost declines by a fixed amount.
Today, Wright’s brainchild, not widely known outside of engineers and consultancies like the Boston Consulting Group, is especially important for EV and battery developers because it explains how they can take on and beat combustion. The law is also a bright oncoming locomotive bearing down on the gasoline-powered vehicle industry. As long as cost parity — expected in 2023 or 2024 — leads to a nascent mass EV market, the embryonic industry will rapidly go on to further price declines as its cumulative production doubles and doubles again. As for the already-humongous conventional auto industry, at current sales rates, it will have to wait about 29 years to double the fleet, according to Sam Korus, an analyst with ARK Invest.
Korus and ARK are among a few outlier bulls who have long forecast very fast cost declines and the onset of a robust EV industry. ARK’s main bet on batteries has been Tesla: It has been perhaps Wall Street’s most aggressive Tesla bull, keeping the EV company its number one investment through thick and thin. To get to this conviction, ARK, again among very few in EVs and batteries, have made Wright’s Law a core factor in their analysis of what comes next across technologies. “When we are looking at disruptive technologies, we are looking at technologies experiencing a significant cost decline, that cut across industries and sectors, and that can serve for further innovation,” Korus told me last evening. “Batteries fit that, and Wright’s Law has been the key element to finding the cost decline.”
Analysts fix the Wright cost decline at different percentages depending on the industry. James Frith is head battery analyst at BloombergNEF (BNEF), a clean energy research firm, which is the other major advocate of Wright’s Law. Relying on the principle, BNEF has long maintained among the industry’s most bullish forecasts for the drop of battery costs and the growth of EV sales.
Frith’s group puts the battery cost decline at 18% per production doubling at the pack level. As an example, from 2010 to 2015, lithium-ion battery capacity doubled seven times, from 0.48 gigawatt-hours to 62 GWh. The average price of batteries at the pack level plunged from $1,194 per kilowatt hour to $384. Strictly speaking, the 18% rule should have taken prices down to about $261 ($384 is about what Wright’s Law calls for with six doublings). But it was still a roughly two-thirds plummet.
From 2015 to 2020, battery capacity grew 2.7 times and the price again plunged by two-thirds, to an average of $137/kWh. That overshot the 18% rule, by which the price should have dropped only to about $213/kWh. But if you had strictly tracked the 18% rule from 2010 to 2020, the price ended up right around where it should have been.
Now, BNEF is using the rule to project what happens next. Following Wright, battery prices should drop to $84 per kWh in 2025, well below the holy grail milestone of $100/kWh. In 2030, they should be an astonishing $58/kWh, and $45 in 2035. The price movements in the 2030s seem dramatic, but they are less so than the prior two decades, namely because it will be harder and harder to double total capacity, Frith said.
Frith said his model assumes that, for the declines after 2025, batteries incorporate advances like silicon and lithium metal anodes. In addition, moving from the currently popular NMC cathode to one with just 0.5% of both cobalt and manganese, and 90% nickel, should reduce raw material costs by about a fifth while increasing energy density by almost a quarter, he said. Among the battery-makers, SK innovation plans to commercialize such a cell starting next year. Wright’s Law, Frith said, “is a great top-down method and tells you where prices are going but not how you get there.”
It is perhaps predictable that Moore and Wright would end up in a competition, even though Wright died in 1970. A number of articles (like this one, in the prestigious IEEE Spectrum) say Wright’s Law is a sounder explanation of what is going on with semiconductors. ARK’s Korus says Wright’s utility seems not to stop. “As we use it, more and more and more interesting things are illuminated with Wright’s Law,” he said. “It’s not just the technology itself, but every process underlying it.”