Comment by TheRealPomax | Hacker News Reader

Comment by TheRealPomax | Hacker News Reader

Parent Full thread

While this sounds like a lot, it's important to remember that the US consumed 4,000 terawatt hours in 2023, so even at 180GW now, and an extra 40GW a year, these numbers are a reason to get angry at why it's taking so long to get serious about solar instead of celebrating a victory. Until we're adding ten+ terawatts per year, numbers that says "it's going to take over ten thousands years for this to even be relevant" are not good numbers.

They're, of course, good numbers for the folks who had been hoping to switch and are now on solar, but they're not good numbers at any scale bigger than that.

And to clarify, because it looks like folks didn't realize this: the target isn't the current consumption rate. It's the rate we need to service everything that comes online "because we now have enough power", from EV-ing the entire US, to GPU mining (or whatever it's fad equivalent becomes), to industries we offshored or outsource because they needed too much electricity and weren't economically viable.

The target is one, and probably two, orders of magnitude more than current consumption.

GW and TWh are not the same unit. You need to multiply by the hours of effective sun in a year (and applying some reduction factor for the fact that there isn't 24h of sun every day, nor that the full production capacity is reached)

Even if that doesn't cover the energy demand, the number is not negligible for 438 GW of capacity. Assuming that the effective full sun equivalent in terms of energy production is only 10% of a day in average, we get: 24 h/d × 365 d × 0.1 × 0.438 TW = 384 TWh

I have no idea about the 10% factor, someone more knowledgeable can coreect me, but, we are not speaking thousands of years if the growth continues before the full energy production value is getting close to the energy demand.

0.1 is reasonable, allowing for cloud cover and dust films reducing output. Maybe a bit high because often panel nameplate output exceeds capacity in other parts of the chain (inverters, grid export points).

But the 438 number should be 180. Cleantechnica is perpetually over-optimistic. Today, we have 180.

Left implied in the original post, but let's make it explicit: we also want folks to switch from ICE to EVs, and the grid is going to have to increase capacity by an order of magnitude to service all of those. We can't yell "EVs are the future" without understanding that that also means we need vastly more electric power than there is today if we also want people to be able to rely on having power at all times. Because EVs don't charge "throughout the day", they get plugged in at night, all at the same time, and good luck with that on the current grid.

Even if 40GW per year gets us to the current coverage in a hundred years, that's not the target. The target needs to be 400GW a year _at the very least_ and realistically it needs to be a terawatt+ if we want to see it happen in our lifetime.

The target isn't today's grid capacity, nor today's usage. It's the one we can expect 50 years from now. Because public infrastructure gets updated at a glacial pace, and charged at "take what we quoted you, double it, then double it again because lol bribes that you can never prove" practices everywhere.

Americans drove 3.19 trillion miles last year. EVs can go ~3.1 miles/kWh, so if we swapped out all the vehicles for EVs, they'd have consumed ~1 PWh in electricity. That's 25% more electricity than we generated last year.

The total US nameplate generation capacity is just 1.3 TW - including mothballed generators, so I'm not sure why we'd need to add 400 GW of capacity per year let alone get to tens of TW of generation capacity.

Admittedly, solar can't generate electricity all the time, but even with a modest 20% capacity factor, it'd take about 600 GW of solar to generate 1 PWh of electricity per year.

Moreover, we rarely come close to using 100% of our existing generation capacity, outside of a few hours during hot summer days. EVs, being well, energy storage devices that are stationary the vast majority of the time, seem well suited to charging whenever we have excess supply - which is again, most of the time.

> I have no idea about the 10% factor

As a single data point in California, the factor for my solar installation was 15.4% for the last twelve months, so your guess seems pretty reasonable

In 2022 the US had ~143 TWh of utility scale solar PV [1]. Utility scale solar PV had a nameplate capacity of ~71 GW [2]. So, the average US capacity factor for utility installations is ~0.25.

Or, for simpler calculations, just multiply nameplate capacity by ~2,000 to get the expected annual generation in Wh.

[1] https://www.eia.gov/electricity/annual/html/epa_04_03.html

[2] https://www.eia.gov/electricity/annual/html/epa_04_03.html

4000TWh of energy per year comes out to about 456GW continuous power. So pretty much in the right ballpark?

you seem to be comparing energy units with power units. 1 GW of installed solar capacity should be give about 1 TW hour of energy per year (365 days * 3 hrs per day as a conservative estimate).

I don't think they're confusing anything. If we use your numbers, then 180GW of solar is less than 5% of energy needs, and that is increasing at less than half a percent annually.

At that rate, decarbonizing by replacing a meaningful amount of the other 95% with solar will not happen in our lifetimes.

Electrifying everything makes about 75% of that energy demand unnecessary as it was waste or avoidable inefficiencies of fossil fuels.

For an average of 456.3 GW per year. Even tripling that for peak loads, it's only 1368 GW.

Except we're not going to triple it. It needs to be an order of magnitude large to deal with the EV push, as well as bullshit like whatever-the-next-GPU-mining crazy is, as well as industry discovering that with cheaper electricity, manufacturing that got offshored because doing it domestically was too expensive suddenly starts coming back as a viable domestic industry. Like aluminum smelting.

And an order of magnitude is conservative given the effect of "finally having pentiful and affordable power" has on the industrial landscape. The target isn't "what we're using right now". That's how you build failing infrastructure before you even got started. The target is "heaven forbid we succeed, and everyone jumps on the band wagon, how much power are we going to need".

> the effect of "finally having pentiful and affordable power" has on the industrial landscape

The nice thing about induced demand from something getting cheap, is that it only happens if that thing gets cheap.

If we can't install enough solar to crater prices, then that extra demand simply won't happen. There won't be a massive spike in coal/gas use that makes the solar a waste of time.

> It needs to be an order of magnitude large to deal with the EV push

Please do the math. Switching every single US car to electric would be less than 150GW average, somewhere around a 25% increase in electrical demand. And even 50% electric is a lot of years away.

Reply on news.ycombinator.com