site banner
Jump in the discussion.

No email address required.

Fossil futures seems to make several arguments.

  1. Fossil fuels help humans live a better life so we should increase their use.

  2. There is no good replacement for fossil fuels that will be available in the next 40 years.

  3. The global warming caused by 40 more years of emissions is not an existential threat.

  4. Any threat from global warming can be mitigated by increased fossil fuel use, what he calls "climate mastery".

  5. Replacements for fossil fuels should be developed.

  6. The best candidates for replacing fossil fuels are nuclear and enhanced geothermal.

I think he is being misleading or is ignorant. In the next 40 years there are obvious candidates for powering a decarbonized economy. Namely solar and batteries. New Solar is rapidly becoming cheaper than existing coal generation in large areas of the world. Battery production is scaling exponentially. While nuclear is the safest power option, it is never going to be the cheapest. Its technology iteration cycle is too slow. Solar power and batteries on the other hand both have fast iteration cycles. This makes them drop in price faster than nuclear. Even if you got rid of the unnecessarily burdensome nuclear regulations, its slow development cycle means it will not beat solar in cost per unit of electricity in the next 40 years.

https://caseyhandmer.wordpress.com/2019/06/21/is-nuclear-power-a-solution-to-climate-change/

Nuclear power plant technology is iterated roughly every 25 years, or twice in the lifetime of a plant. Many first generation plants are still operational, while few third generation plants have been commissioned, and fourth generation plants are still in the planning stage. Even if every design iteration was a factor of 10 better than the previous one, solar, iterating 50 times faster, could outdo this improvement over the same timescale with a mere 5% improvement per iteration. Since this is roughly the solar learning rate, we can now ask if each nuclear design iteration is 10x better than its immediate predecessor. Obviously not.

I am a great fan of Casey, but I think it's important to know about the assumptions that are underlying his analysis:

  1. The learning curve for solar will continue its current trajectory (aka solar keeps on getting exponentially cheaper).

  2. Some kind of solution for storage & the currently non-electrified industry emerges: probably hydrocarbons derived from CO2 direct air capture (Caseys startup).

Number 1. will allow to just ridiculously overbuild solar (so that even in deep winter it mostly generates enough for base-load) and 2. will allow to absorb the surplus by turning it into hydrocarbons (even if there are some losses, we barely care because electricity during sunshine hours is so cheap). IMHO, the logical conclusion in this scenario would be that we use hydrocarbons as storage as well (and use gas peaker plants in winter), though Casey really likes batteries for some reason [0].

About Caseys solar future: I think that Caseys assumptions aren't too far-fetched (the solar learning curve certainly has shown a lot of staying power so far), so maybe this will play out the way Casey describes it. But it really really hinges on storage working out in some way. If one of the two assumptions somehow fails, we will be in a big world of pain. In this case, we will probably get climate-change + whatever haphazard geo-engineering we can muster in an attempt to neutralize it.

About the nuclear future: On the other hand, even if we go all nuclear, something still needs to be done for the non-electrified part of the economy (e.g. ships, airplanes, steel production, fertilizer production etc). That part currently uses much more energy that the electrified part of the economy. So we'd either need great battery tech (+ lots of industrial process innovations) or hydrocarbons from CO2-captured as well if we want to become even nearly CO2-neutral.

[0] And I think that Caseys biases show most strongly on this front, e.g. the somewhat gloating tone of this article of his, which basically describes the gas peaker plants being driven out of business due to market manipulation by large batteries with fancy algorithms. This is obviously not optimal because these gas peaker plants are actually needed today since the batteries can only smooth over small variations and can't deal with longer-term shortages.

Holy shit, there are so many red flags in that screed. People take this guy seriously? He's just spewing random figures in the hopes nobody checks them.

It's just an endless stream of "here's some numbers, NEW PARADIGM! insert sneer, more numbers, SYNERGY!""

Inadequately serviced demand for short-term grid stabilization is a market inefficiency and represents an opportunity for arbitrage.

And he just lathers on the sneering so he doesn't have to explain how this arbitrage scales to 100% seasonal battery storage and "10-15x overbuilding solar". It's as if SBF claimed the Japanese bitcoin arbitrage was a secret trick to dominating the world economy, rather than a rapidly exhausted pool of free money laying on the floor, like finding coins under a vending machine.

I suppose the sneering tone can trigger some people. But I think his analysis is one of the better ones I have read. The learning curve for solar means it will be a large part of the energy future. At this point as far as I can tell the debate is about the minority of power that is baseload or long term seasonal storage. Has anyone looked at just over building batteries?

AFAIK the learning curve for batteries doesn‘t look that rosy for the near future (for grid-scale storage).