> One problem is that they only produce when the sun is shining, or the wind is blowing, etc.

This is the irrelevant part when they're not the majority of the grid. We already have peaker plants to handle the differences in load throughout the day. They currently mostly run during the day and the early evening. You build a bunch of new solar and instead they'd mostly run at night, because the new solar takes the existing daytime demand plus the new daytime demand and the existing peaker plants take the new nighttime demand.

> They’re only cost-effective if you put them in the right places, and that’s usually far away from where people want to use that power.

This is not really true in the US. Solar doesn't work very well in Alaska because it's too far north, but it's fine basically anywhere south of the Canadian border, which is where essentially the entire US population actually lives. The average wind speed in every US state is enough to run a wind turbine.

There are some locations that are better than others, but better enough to build a thousand mile transmission line? That's almost a coin flip even before you consider what a liability it is to centralize power into a big exposed long-distance transmission line instead of just distributing the generation everywhere.

> Sure, you can truck in jet engines to use them as generators, but then you have to fuel them...

As a temporary measure, that's the trade off against short-term higher electric bills. As a permanent measure they're rubbish, but they're not needed as a permanent measure anyway.

> And datacenter builders want to externalize all their costs.

That's not what that means. You can't internalize demand for a commodity. All you can do is increase supply to meet the new demand.

Not true. We are already seeing duck curve problems where the price should go negative, unless the solar and wind farms are told to curtail.

They need a buffer to be able to fully and properly contribute to the grid. Without a buffer, solar and wind farms are already causing some problems. They can’t really grow much beyond where they are today, without adding a buffer. Sad, but true.

T. Boone Pickens is pushing wind as hard as he possibly can out in the Texas panhandle, and it’s getting harder and harder for him to do that, in part because much of that is done without a buffer. Granted, he’s also having to fight transmission line problems.

For solar to be effective from the perspective of the grid, it needs to be installed in a place that gets enough average sunshine. And that latitude line is way lower than the border between Canada and the US. You also have to take weather and climate patterns into account, because if you’re below that line but you get too much average cloud cover, then you’re still not cost-effective enough.

The only places where you can install solar panels without regards for those kinds of constraints, is where people don’t care how many panels they have to install, in order to get the amount of average power that they need. If you over-provision by enough, then you can install useful amounts of solar in otherwise considerably less than ideal sites. But then you’re also willing to pay a price and space premium that no grid operator would be willing to pay. Many homeowners might fit into this category.

And many places where you would want to install solar panels close to the consumers so that you can avoid long transmission lines, that land just isn’t available. Or not available at a price that any grid operator would be willing to pay.

> We are already seeing duck curve problems where the price should go negative, unless the solar and wind farms are told to curtail.

This happens with solar in the summer because it generates more in the summer, and then if you build enough for the other seasons you have "too much" on many days in the summer. But storage doesn't help much for that anyway because you'd need storage that lets you store power from July to use in January.

And "too much" isn't a problem when it only happens a minority of the time. You'll have critics complaining if it happens at all, but that doesn't mean it's a real problem.

The real problem is "not enough", which is the thing that doesn't get really bad until you're trying to run the majority of the grid on it. If it's only half then the other half is reliable power sources, some of those are still peaker plants and then you have the option to run them during the day on the days when it's cloudy, renewable generation is sometimes low but it's hardly ever actually zero, and by that point you're at something like 80% coverage on the lowest generation days and can get the remaining 20% through variable pricing.

> For solar to be effective from the perspective of the grid, it needs to be installed in a place that gets enough average sunshine.

A solar panel generates ~50% more in Arizona (on the border with Mexico) than New York (on the border with Canada). That isn't totally negligible but but it's not so large as to make it unworkable.

> Many homeowners might fit into this category.

Naturally, because it solves two problems at once. First, no land cost because you're using what would otherwise be an empty roof. Second, no transmission cost because the power is consumed where it's generated.

This is an easy win for homeowners because even at the less advantageous latitudes, they'd be paying ~$0.10/kWh for their own solar while the New York power company is asking 28.6.

And there are still plenty of not just homes but also commercial buildings that don't have solar on their roof. Give every Walmart and Home Depot an incentive to install a support scaffold full of panels over every parking lot. Why pay for a lot of land when we haven't even taken the easy wins yet?

> The real problem is "not enough", which is the thing that doesn't get really bad until you're trying to run the majority of the grid on it. If it's only half then the other half is reliable power sources, some of those are still peaker plants and then you have the option to run them during the day on the days when it's cloudy, renewable generation is sometimes low but it's hardly ever actually zero, and by that point you're at something like 80% coverage on the lowest generation days and can get the remaining 20% through variable pricing.

You're not thinking like a grid operator. They need to pay to keep those peaker plants operational, even if they're not being used. The more peaker plants that have to be kept offline for more of the time, the less attractive it is to buy the power from the solar or wind plants.

Until solar and wind can generate reliable amounts of power, day-in and day-out, 365 days out of the year, you will always need those peaker plants to take up the slack. But those renewable sources will never be able to do that. The best they can hope for is to overgenerate during the times that they are capable of doing so, and storing the excess somehow in another medium. That could be batteries, flywheels, pumped hydro, or whatever. But they need that buffer if they are going to be a larger part of the grid.

> A solar panel generates around two thirds as much in New York (on the border with Canada) as in Arizona (on the border with Mexico). That isn't totally negligible but it's not so large as to make it unworkable.

It might not be enough to keep homeowners from installing solar panels in those kinds of environments, but it's definitely enough to keep large-scale power generation facilities from building there.

> And there are still plenty of not just homes but also commercial buildings that don't have solar on their roof. Give every Walmart and Home Depot an incentive to install a support scaffold full of panels over every parking lot. Why pay for a lot of land when we haven't even taken the easy wins yet?

I agree that parking lots and roofs of buildings are targets that look easy. But the incentives just aren't there. It's still just way too expensive, for not nearly enough ROI to make it happen. And I just don't see incentives of that scale coming, not even with the New Democrats.

In theory, I don't disagree with you with regards to where we should be going and what we should be doing.

But in practice, there is a huge difference between theory and practice.

> They need to pay to keep those peaker plants operational, even if they're not being used.

But they are being used, every night. And then you can also use them during the day on the minority of days when it's cloudy.

> It might not be enough to keep homeowners from installing solar panels in those kinds of environments, but it's definitely enough to keep large-scale power generation facilities from building there.

Nope:

https://cleanview.co/power-projects/operating/solar-farms/us

You could easily confuse that with a population density map of the US. The main tell is that the dots are on the outskirts of the major cities rather than in them.

> It's still just way too expensive, for not nearly enough ROI to make it happen.

It kind of is happening. Nearly 100% of net new generation capacity in the US is renewables, with some new natural gas barely offsetting the gradual of coal to net to zero.