Li-ion Batteries – Doing Great on Cost but Not on Energy Density

I was reading a fascinating book Speed & Scale by John Doerr (of Kleiner Perkins ), in which he consolidates impactful insights on climate action from people who have been there and done that. More about the excellent book at the end of the post.

While reading the inputs from Ryan Popple, a brilliant electric mobility pioneer who sadly passed away just after the book was published in 2021, I came to know about the Wright's Law, a law much less spoken about compared to Moore's Law, but one that has perhaps an equal amount of importance in the context of industrial economics.

Wright's Law says that costs fall by a predictable percentage every time the total amount of items produced doubles. Thought of by Theodore Wright in 1936, it uses experience as the key driver to enhance efficiency and reduce costs.

Of course, any MBA student worth half his degree will recognise parallels in the concept of "economies of scale" but these two are quite different.

While they both describe how production economies improve and costs drop, they provide different reasons for these to happen. In Economies of Scale, costs drop because we make more items at the same time, resulting in lower unit fixed expenses, better input prices etc.

Wright's Law is more about the power of experience, and refers to the cost of a manufactured item dropping as the total production increases over time, and has more to do with the positive results from learning curves. (Ok, it does have something in common with Moore's Law, though the Moore one takes time as the metric for efficiency/productivity growth).

Wright's Law, as Ryan says in the book, has indeed worked quite well for Li-ion batteries. From a high of about $6000/kWh in 1996, the price of Li-ion battery has dropped to about $100/kWh in 2026 (it ranges between $100-200/kWh depending on the country and application - in India, for large stationary applications, I have seen quotations at $75/kWh!).

Wow, that's a 98% drop.

If I take 2010, the benchmark price was $1000/kWh. Between then and now, it still is an awesome 90% drop.

While the dramatic reduction in prices is good news for many segments that will use Li-ion batteries on large scales (especially the stationary use or BESS segment as it is better known), another important parameter - in addition to price - for the use of LiB batteries in the mobility segment especially is the mass energy density (amount of energy/Kg or kWh/Kg).

On energy density, the picture is quite different. In 1996, the best energy density one had for Li-ion battery was perhaps around 150 Wh/Kg; fast forward 2026, the best commercialized battery energy density is around 300 Wh/Kg.

Just a doubling in 30 long years. Not really a growth story to write home about.

That is, while costs have fallen by 98%, actually over 99% if I take into account the higher number of cycles that today's batteries have, energy density's growth has been rather muted - and is still only a bit more than 2% of that of fossil liquid fuels such as petrol or diesel. That is, a Kg of petrol or diesel has 50 times as much energy as a Kg of Li-ion battery.

I have indicated the contrasting performances in the table below.

(All estimates in the table are approx, as I've tried to make them international averages...)

Why have Li-ion battery energy densities failed to scale significantly, just a 2-fold growth, while a related law, Moore's Law, has worked so well that transistors on a chip have grown over 5000-fold in the past 30 years?

Well, we are talking about quite different things when we talk about energy density and the number of transitors on a chip.

In the case of computer chips, an analogy could be printing a book. By using smaller and smaller fonts, you can fit many times the number of words on the same page without buying a larger piece of paper. Similarly, by using advanced lithography, chip makers pack billions of tiny switches onto a single square of silicon - in essence, engineers have been able to make transistors smaller and smaller over time. When switches cannot get any smaller, engineers stack transistors on top of each other, and increase the number per unit area. (While Moore's law has undoubtedly served the world magnificently so far, it is now approaching its material limits as well.)

In the case of Li-ion batteries, its density is limited by chemistry and physics. The base materials have maximum theoretical limits for energy storage - about 500 Wh/Kg. So, this is very different from the case of transistors and chips, where it is a question of making the unit smaller, at least until it reaches a size when size starts affecting fundamental material properties.

So, Li-ion batteries have done exceptionally well on economics but still do poorly on energy density. And the future doesn't hold great promise for energy density with Li-ion chemistry alone. Solid-state batteries and other chemistries could take it higher, but let's get honest - these are at least a couple of decades away from large-scale commercialization, given their state of development, and even more important, given the incumbent stakeholders' strong affinity to protect their hundreds of billions poured into the Li-ion battery ecosystem.

The poor energy density of Li-ion battery has implications:

1. Battery adoption will accelerate significantly in domains where weight is not a big issue (stationary applications, small and medium vehicles)

2. Big challenge for heavy vehicles (trucks, ships, airplanes)