Hacker News Reader

Comment by S201 | Hacker News Reader

> If that trend keeps up in the 2020s, anything else will have a real hard time catching up.

That's the problem though: *IF* that trend holds. Past performance doesn't guarantee future performance.

Climate change presents an existential threat to civilization so relying on a single trend holding seems like a fairly risky bet to make if you ask me.

In that vein, I'd argue for throwing everything at the problem. Build it all, wind, solar, geothermal, and nuclear. If we overshoot and spend billions more than was necessary, great, that's a hell of a lot better than undershooting and losing trillions from entire economies shutting down.

View on HN

Solar power including storage is already the cheapest form of energy. Any further improvement is just gravy.

Regardless of how cheap it may be, the storage of solar power is still a big problem. If we're on the topic of cost trends and batteries, the price of lithium is up ~500%.[1] That's making batteries more expensive than in the past and given all of the demand for them I don't really see it going down any time soon, at least not until new lithium mining plants are built mainly in South America, which still isn't likely to drop prices again in the near future.[2]

1. https://tradingeconomics.com/commodity/lithium

2. https://www.economist.com/the-economist-explains/2022/06/09/...

Global production of lithium has doubled since 2015.

I think it's reasonable to expect energy storage to become cheaper purely by the mechanisms of capitalism. The market for energy storage scarcely existed 4 years ago -- there were a handful of pumped hydro storage plants and one compressed air energy storage plant and that was it. The US had 1.5 GW of battery capacity in 2020; we added 5.1 GW of capacity this year and currently scheduled projects will bring 30 GW of battery capacity online by 2025 (per the EIA).

It's clear from the current buildout and the current price of solar that there will be a large market for energy storage in the next 10-20 years. That means there is a large market opportunity to build grid energy storage and to increase production of the entire supply chain for grid energy storage. I would be surprised if it is not met.

There are alternatives for grid storage that are being deployed right now. Sodium or Flow batteries, CAES and underground pumped hydro. As lithium prices increase, alternative technologies will be deployed in scale and come down in cost.

If it's so cheap, batteries included, why are the most optimistic industry projections only putting it at 20% of the US grid by 2050? [1]

Why are they projecting that coal gas consumption will continue to grow, year over year for that same time period? [2]

[1] https://www.eia.gov/todayinenergy/detail.php?id=50357

[2] https://www.eia.gov/outlooks/aeo/production/sub-topic-03.php

The second link shows that natural gas increases steadily for exporting and for domestic industrial consumption. If you click the Excel link under the table graphics [1], it shows that natural gas consumption for generating electricity remains below 2020 demand levels until 2047. As for why they think that demand will pick up in the late 2040s after decades of stagnation, I'm not sure; maybe it's projections of population growth finally overcoming trends in fuel efficiency.

I will note that the EIA has badly underestimated renewable electricity growth in the past. Their projections are very restrained, far from "optimistic industry projections." You can read a retrospective on their past forecasting here:

https://www.eia.gov/outlooks/aeo/retrospective/

In particular look at the electricity section of the table, "Comparison of AEO Reference case projections with realized outcomes, 1994–2021." The EIA underestimated future usage of solar power 80% of the time and future usage of wind power 84% of the time. The magnitude of the underestimates was actually worse for solar power, though.

[1] https://www.eia.gov/outlooks/aeo/production/excel/subtopic3_...

It's also vastly underestimated growth in natural gas energy generation. By that reasoning, we can only assume that the 21st century will continue to be a century of fossil fuels.

The EIA almost always projects boring linear growth for everything. Their data is a terrific resource for what is currently happening, what has happened recently, and what is already scheduled on the pipeline. I would not put much stock in their projections beyond their Short Term Energy Outlook. I would certainly not describe them as the most optimistic industry projections.

They project that the solar growth rate will drop after 2023 when the ITC would have expired, but it was extended in the 2022 Inflation Reduction Act. From there, they expect the growth curve of solar to slowly decay. It's basically all a linear projection and scarcely any more insightful than that.

They expect that electricity generation from coal will be flat (except for already-announced early coal plant retirements). They expect nuclear to be flat. They expect generation from natural gas to increase (unless renewables are cheap, in which case they expect it to be flat). They fail to anticipate the replacement of coal by natural gas or renewables and they fail to anticipate the replacement of natural gas by renewables.

Finally, as a point of comparison, take a look at their projections for EVs and hybrids. The EIA projects that EVs and hybrids will still only make up about 20% of the market for new cars in the US in 2050. (See page 9 of [1]).

[1] https://www.eia.gov/outlooks/aeo/pdf/AEO2022_ChartLibrary_Tr...

The EIA is very consistently very wrong.

https://pv-magazine-usa.com/2020/07/12/has-the-international...

Your link talks about the IEA, which is the International Energy Agency. I was citing the EIA, which is the US Government Energy Information Administration.

They are not the same thing.

This is just Clayton Christensen/Geoffrey Moore/SteveBlank stuff. It's the same patterns of growth, price, adoption curves, irrational exuberance, etc down to the bad predictions by entrenched industry experts.

It's remarkable how cleanly it's playing out. There's probably some easy academic opportunities on writing about it if someone wants some easy first authorship opportunities while sitting in their pajamas, it's there for the taking.

This isn't just hopes and dreams though. There's an economy of scale argument called Swanson's law which is expected to continue for the forseable future: https://en.wikipedia.org/wiki/Swanson%27s_law

There's also quite a bit of research going on into different cell technology (https://en.wikipedia.org/wiki/Solar-cell_efficiency). The current physics wall puts it at about theoretical 4-fold increase in electricity per square meter compared to what we have today without consideration of the cost of manufacturing or longevity, which is expected to again, continue to improve.

Also as far as land goes, people are really bad thinking at scale. If you look here: https://www.seia.org/research-resources/major-solar-projects... you'll see all the installations and the ones planned. The sun doesn't shine every place, on every day, sure, but when you have that kind of coverage it's a different story.

There's a lot of arid desert out there and lots of potential for combining solar panels with existing food farms through something called agrivoltaics: https://en.wikipedia.org/wiki/Agrivoltaics ... many crops do as good or better with shade so this is just a straight increase in profit per acre for those crops with the dual use of energy generation.

This is all pretty new. The cost tipping points in favor of solar just happened a few years ago. We're seeing one of Clayton Christensen's graph of disruptive technology in real-time right now (see https://en.wikipedia.org/wiki/Levelized_cost_of_electricity where solar went from being the most to the least expensive in about a 5 year period)

Everything hits a physics wall, of course but not only does it look like there's still plenty of runway with solar before we get there but there's also lots of unexplored opportunities such as, just throwing this out there, a solar capturing wind turbine that can cleverly capture energy both ways. The point is this is still extremely greenfield

To be clear, I'm not arguing against solar here. We should be building as many forms of non-carbon-emitting energy generation as we possibly can.

A challenge of large solar plants like you're describing is getting the energy to where it's needed and when it's needed. A ton of solar panels in the desert does no good if there's no way to store that energy or get it to a city a few hundred miles away. All of that requires significantly more investment in storage mechanisms and interconnectors to move it around the grid. The latter actually being most concerning to me simply because of NIMBYs fighting projects to build more high voltage power lines because they're ugly and ruin views from their properties. It's hard to quantify that, but when you need to run so many power lines over such long distances that touch so many people's properties it becomes a permitting nightmare which dramatically slows down, or even cancels, projects.

I also touched on this in another comment below: In terms of storing solar energy with our current method of lithium-ion batteries, well, the cost of lithium is up ~500% in the past two years and is expected to stay that way for the foreseeable future: https://www.economist.com/the-economist-explains/2022/06/09/...

As to Swanson's Law, sure, I get where the trendline currently is, but as a counterexample: How's Moore's Law holding up these days?

I'm sure there's plenty of articles about different battery storage methods and all that being researched, but the fact is that we need proven, on-demand energy generation that drops into the existing grid NOW and nuclear provides that. But by all means, build all of it: nuclear, wind, solar, etc. Having too much carbon-free energy is a much better problem to have than too little.

First let me dispatch this "existing grid NOW and nuclear provides that" - it's a 7 year average lead time as of 2022; solar is 2. And if you granted all the accelerations you're going to propose to solar or wind as well, it'll still beat nuclear. It's just simpler and cheaper to deploy. It's also expandable and upgradable and you can start producing after the first station is done. Your 2022 nuclear reactor will still be based on 2022 technology in 2072. The solar farm, however, will be current with the times as the parts are swappable.

The predicted end to Moore's law had been anticipated for many years. This took nobody by surprised. The only surprise really is that it went on longer than expected.

From all indications we're in the solar version of the 1980s right now on the Moore's law timeline and in the future we'll see a solar PV equivalent that's as remarkable as say, the 2TB microsd card - imagine cheap, high-efficiency newspaper thin panels that you buy packaged like those plastic ponchos from a 7/11 that you just unfurl onto the ground, plug in and power whatever you feasibly need.

The future of solar will be like telling someone in the 1980s that we live in a time where we give away 32gb of storage for free with company logos stamped on it at trade shows and it is so normalized that nobody takes any notice.

> First let me dispatch this "existing grid NOW and nuclear provides that" - it's a 7 year average lead time as of 2022; solar is 2.

Nuclear can certainly be made quicker if we had the political will to do so but I'm not going to try to argue that it's quick. That said, if we would have invested in this 20 years ago like we should have it wouldn't have been a problem. The argument has been "wind & solar are going to ramp up so quickly in the near future so why build nuclear?" and then it takes longer than we thought and underwhelming progress is made.

Regardless of how long it takes:

> The best time to plant a tree was 20 years ago. The second best time is now.

> It's also expandable and upgradable and you can start producing after the first station is done. Your 2022 nuclear reactor will still be based on 2022 technology in 2072. The solar farm, however, will be current with the times as the parts are swappable.

This argument doesn't really make sense. You can't add reactors to a nuclear plant? Or retrofit them as needed? The primary form of generating all forms of electric other than photovoltaics uses steam to spin a turbine; that's the same now as it was 100+ years ago. Who cares if it's 2022 technology in 2072 if it's producing useful electricity?

I'm sure solar panels have some runway to the technology, but that's not the point here. Solar currently does not have the ability to drop into the existing grid with large scale, on-demand energy nor will it in the foreseeable future. It's a slice of the energy pie, but to say that it's going to improve so much doesn't mean it's a wise idea to put all your eggs in that basket.

Reply on news.ycombinator.com