In 1936 Theodore Paul Wright published a paper called “Factors Affecting the Costs of Airplanes”. His key observation is that we “learn by doing”. The more we do something, the better we get at it. It’s a simple but powerful concept that is quite good at predicting the cost curve of new technologies.

It boils down to this formula:

where Y is the cost to produce something as a function of the starting point to produce the first unit (a), the cumulative production to date (X) and the learning curve slope (b).

Here is what the graph looks like:

‍

Unlike its famous sibling, Moore’s Law, which predicts prices as a function of time, Wright’s Law predicts prices to decrease as a function of total production to date. Once we know the starting point (a) and the slope (b), we can estimate the future cost of a product by plugging in the cumulative production volume at that time.

‍

Let's look at a few examples of how well Wright's Law does at explaining the cost curves of lithium-ion batteries and renewable energy:

‍

These charts demonstrate how well Wright’s Law predicts the cost declines of these technologies. This is human ingenuity driven by competition. The more we produce something, the better we get at it.

So what are the implications for founders, investors and policy makers?

‍

1. New technologies need to be competitive from the start in at least some markets even with low production volume. Or they need to rely on subsidies and heavy investment to ramp production.

If a new technology can’t compete from the early days at least in some market (e.g. a product vertical or geography), it will never have the opportunity to get on the cost curve in order to drive down costs and unlock broader adoption. You might have a theoretical cost advantage if you compare your new technology at scale but that’s not the reality. A new technology needs to compete when its “X” (the cumulative production to date) is tiny compared to an established technology. For example, if you start a new battery company you need to have a huge advantage in “a” (your starting point) or a really steep slope “b” so you can outrun the cost curve of li-ion batteries that will continue to get cheaper as you try to scale.

That’s why starting new companies that purely compete on cost is really hard. The best approach is to find a niche market that gives you an advantage in cost structure right off the bat (e.g. solar was first deployed in space because there was no better alternative) or find buyers that are willing to pay a premium for your solution and then expand from there (e.g. Tesla’s roadster and Model S were unique enough that people were willing to pay a premium).

New technologies should always be compared on a “Wright’s Law” equivalency with incumbent technologies. To know which technology will be cheaper, we need to compare current costs and the slope of learning curves to estimate future price advantages.

‍

2. Discovering the slope of the learning curve takes time, but we can use historic data as benchmarks.

The slope of the learning curve will not be known at the start when production is starting. The best we can do is bound the slope in context of past examples.

Below are some examples of actual learning curve data of different technologies. It shows the decline in prices for every doubling in installed capacity.

• Solar: 36%
• Onshore Wind: 23%
• Lithium-Ion Batteries: 20%
• Electrolyzers: 10%

The differences between them has a massive compounding effects over time. Solar used to be a lot more expensive than wind (and all fossil fuel alternatives). But given its steep learning curve, it has overtaken all other energy producing technologies in the world.

‍

3. When investors and policy makers ignore Wright’s Law it’s a great way to set a bunch of money on fire.

The case for funding new technologies through public subsidies and private investments is to help them get on the Wright’s Law learning curve, especially when clean technologies have to compete with carbon-emitting incumbent technologies that have huge installed bases (i.e. their “X” is large).

So how do we pick the right technologies? Applying Wright’s Law, we should focus our efforts and capital on technologies that can show, even at low initial production capacity, that they can get on a steep learning curve as capacity increases. Solar is the poster child of kickstarting a technology that needed a lot of subsidies and private investments to drive costs down. It was all worth it at the end. But for every success case like PV solar, there are many other technologies that never fully took off, for example concentrating solar. And the jury is still out on technologies like green hydrogen and direct air carbon capture.

‍