🔥 Feel the exponential

Or make a thousand bucks.

Sometimes a simple observation can be more predictive than a brick-by-brick analysis. This week was such a week. 

When I coined the phrase “Exponential View”,  I noticed that many technologies were exhibiting traits I would later describe as exponential. Getting drastically cheaper, mostly through innovation, their prices would fall on a compounded basis. They would get unfeasibly cheap over just a decade or two. As they got cheaper, they would find their way into our lives faster and faster. 

We all know about Moore’s Law and chips. An iPhone is more powerful than a Cray-2 supercomputer, yadda, yadda. Keck’s and Kryder’s laws describe similar patterns in bandwidth and hard-disk storage.

Other, very different technologies had the same surprising attributes: solar power generated by photovoltaics, for example. In my book and in dozens of essays in this newsletter, we've explained at length how and why this happens. It’s counterintuitive, perhaps, but happen it does. 

This week, Ember Climate released new data on the additions to the electrical system of new solar generation. Did the solar capacity added to the energy system increase in-line with the forecasts of the major analysts? Or did it smash past them?

Ember points out, as many have in the past, that traditional (call them linear) analysts still lowball their forecasts. They have to go back and revise them upwards, several times a year. Not because they are not smart (they are!). Not because they don’t do tons of legwork (they do!). And not because we don't love them (we do!)¹.

But because exponentials and linears don’t mix well. I know that sounds vague but it boils down to that: exponential phenomena don’t sit easily with linear thinking. You need imagination and extrapolation for one and a stubborn empiricism that would make A J Ayer proud for the other.

Flywheel go spin

I’m going to simplify the exponential process. Learning by doing and building experience brings costs down. The more manufacturers make something the smarter they get at making it. This increases demand. This in turn delivers economies of scale for suppliers who now benefit from learning and from scale. This further reduces prices, driving demand further. This new market attracts new entrants in the primary market and in ancillary services markets, creating vibrancy in distribution. This flywheel can turn very quickly. 

And what it gives us is where we are. Since the year 2000, solar capacity has been added at a non-linear rate, lots of wobbles to be sure, but it’s getting faster. Eleven years, to 2012, to add 100GW of capacity; less than three years to add a second 100GW; In 2023, we added 100 gigawatts of capacity in just three months.

The data don’t look linear. So what happens next? The greenish line? The orange one? The blue? The red one? Most analysts have historically built forecasts that look like the flattish orange one, and investors and policymakers adjust policy to match.

Linear roosters

Linear thinking still rules the roost. Esteemed linear analyst, Wood MacKenzie, pronounced that China would be producing double the amount solar panels than the world needed over the next three years. This is an alarming story: too much capacity, overextended firms, an industrial collapse. Here’s WoodMac writing in three months ago:

The total capacity of wafers and cells of the top 10 ranked companies is expected to reach 830 GW in the next three years​, which is enough to satisfy global demand twice.

Really? This suggests the firm reckons that in 2027, we’ll add 415GW of solar capacity globally. We’ll add more like 600GW in 2024, so the market will what… collapse? I’ll not have whatever they are smoking. 

Here are my predictions for solar capacity additions in 2027. I’ll use my “feel the exponential model”, which is part tea leaves, part Excel and fully proprietary². I predict that in 2027, global solar photovoltaic additions will be no less than 975GW, unless they are constrained on the supply side (which, according to WoodMcKenzie, is around 830GW.)³ 

And to make it more interesting, I’m happy to bet $975 $1,000 that I’m right. Whoever takes the opposite side stands to nab 100,000 shiny pennies. This footnote will give us more terms and conditions.⁴