[adgjlsfhk1]

12 hours of storage is likely more than needed. With a 20% nuclear, 40% solar, 40% wind generation mix (for a very simplified example), you will have solar producing solid power for 10 hours, with wind and nuclear keeping up overnight.

However lets say that it is 12 hours/30TWh. In 2023, the world produced ~1.1 TWH of batteries. In 2014, the world produced 0.05 TWH of batteries (with steady growth year over year while prices fell by 10x). If you give grid scale batteries a 5 year lifespan (before recycling), that means we need 6TWh/year of grid scale battery production, which at current rates of increase in battery production, we are 5-7 years away from.

For comparison, 5-7 years is roughly the time it takes to build a single nuclear reactor.

[cupcakecommons]

This is under current assumptions which hinge primarily on political will and regulation - not on physics or true construction time.

[Manuel_D]

Unfortunately 12 hours of storage is still going to be a shortfall, even with overproduction. Researchers analyze historical weather data and simulate how renewable grids would perform on that historical trends, measuring periods of underproduction. Even with 50% overproduction and 12 hours of storage, we're still looking at an unacceptably unreliable grid: https://www.nature.com/articles/s41467-021-26355-z

> However lets say that it is 12 hours/30TWh. In 2023, the world produced ~1.1 TWH of batteries. In 2014, the world produced 0.05 TWH of batteries (with steady growth year over year while prices fell by 10x). If you give grid scale batteries a 5 year lifespan (before recycling), that means we need 6TWh/year of grid scale battery production, which at current rates of increase in battery production, we are 5-7 years away from.

Even ignoring the fact that 12 hours is insufficient, you're making the following assumptions:

1. The production of batteries will sextuple in the next 5-7 years.

2. 100% (or close to 100%) of battery production will be dedicated to grid storage.

3. Electricity consumption will remain static.

The first one may or may not pan out. Battery production is already bottlenecked by resource extraction, and it's unclear if the rate of extraction can keep up. The nature of extraction is that once easily accessible deposits are exhausted, companies shift to the harder-to-access deposits. This is only economically viable if cost increases enough to incentivize that investment. The HN crowd tends to assume that everything adheres to Moore's law, but that doesn't work in reality. The price of steel, for instance, doesn't exponentially decline.

The second two are certainly not true. EV are predicted to make up the vast majority of battery sales. Redirecting batteries to grid storage would necessitate delaying EV adoption, ultimately increasing emissions. Stationary storage accounts for a small fraction of battery production (https://rmi.org/the-rise-of-batteries-in-six-charts-and-not-...). Electric vehicles only account for a bit under 20% of vehicle sales worldwide. With many countries slated to stop sales of ICE vehicles in the next 5-10 years, we're still looking at most future battery production going to satisfy EV demand even if it grows to 6TWh per year as per your assumptions.

And electricity use will certainly increase. Both as poorer countries develop and start deploying air conditioning and other electricity consumption. And as other sources of primary energy consumption is shifted to electricity. Remember, electricity generation only makes up ~40% of total energy consumption. The remainder will have to be converted to electricity as part of full decarbonization.

[adgjlsfhk1]

That nature paper is garbage. It's imagining a grid that is 100% solar+wind, which no one is proposing building. Changing that to a grid that has 20% nuclear/geothermal would completely change the figures (in that it would dramatically shift the wind/solar ratios).

Also my battery assumptions were missing the fact that the world already has ~5TWh of hydro which can be used as a battery (even when not pumped hydro by releasing only when you need power).

[Manuel_D]

Dams don't double as energy storage in the same way as batteries. Their rate of recharge is limited by precipitation. You can't run a dam turbine in reverse and fill the dam with excess power (you can with pumped hydro, but we have way less than 5 TWh of that). The best you can do is totally shut them off and let them refill. And in reality, dams have to constantly release a minimum amount of water to avoid creating a totally dry riverbed.