[thinkcontext]

Can't believe all the comments so far and no one feels the need to justify their opinion with numbers. Drives me nuts.

You are correct, according to National Renewable Energy Lab (NREL), rooftop solar is over twice as expensive as utility scale solar[0]. The NREL estimate is helpful in that it details the components of the cost. And, as you would expect, the hardware and labor costs are lower for utility scale because of, well, scale, but the dominating difference is soft costs (land, marketing, profit, overhead, etc). Its worth noting, soft costs for residential solar are significantly lower elsewhere in the developed world (DE, AU) mainly because permitting and marketing are much easier.

[0] https://www.nrel.gov/news/program/2018/costs-continue-to-dec...

[brownbat]

I'm a big fan of that study.

It suggests there may be an opportunity for a company to push rooftop solar for big strip malls and other large footprint single owner buildings - they aren't utility scale, but the install efficiencies make them pretty competitive.

The incentives would be tricky, since the owners are less likely to be paying for utilities than the leaseholders...

[floatrock]

https://www.wundercapital.com

The problem with commercial solar is the cost of the financing. Home solar is on par with a car loan, banks know how to do that with personal credit scores. Utility scale solar is the realm of private equity.

Something like a strip mall is a weird in-between that requires enough manual review that the cost to do all of the risk analysis that financing requires kinda outweighs the potential returns. Will that KMart or Costco keep that store open for the duration of the project? What happens when the tenancy changes?

I don't have any relation with Wunder, but from my understanding of them, they built some stuff to streamline all the financing and permitting required of commercial-scale to make those projects pencil out.

[kragen]

On the contrary, personal loans from banks routinely charge much higher interest rates than commercial paper. Think about the interest rate on a money market account, which consists of bonds. That's the rate PG&E is going to pay if they float a bond issue to build a solar power station. Large companies, particularly including electric generation utilities, can also get loans from banks at similarly low interest rates.

This is precisely the genius of SolarCity: they were able to use the low interest rates banks would charge them to install solar power on people's houses at a much lower cost of capital than the same banks would have charged the same homeowners without SolarCity's intermediation.

[thaumasiotes]

> This is precisely the genius of SolarCity: they were able to use the low interest rates banks would charge them to install solar power on people's houses at a much lower cost of capital than the same banks would have charged the same homeowners without SolarCity's intermediation.

That sounds like a pretty interesting pitch as a business model.

But if I recall correctly, SolarCity failed and had to be bailed out by, um, Tesla. Why? Was the free money just not enough?

[cogman10]

I'm going to push back a bit. Interest rates can be favorable to a company but not money market fund favorable. That would be a better interest rate than a US treasury note.

PG&E currently has 2.4% bonds out. Better than what you can get, but not lower than a mmf (usually less than a percent).

As a rule of thumb, rates won't sink lower than ten year tnote rates. Nobody will purchase a bond from a company that has a lower rate than a tnote.

[kragen]

Thank you for the concrete numbers and the correction/additional information!

[bwb]

Walmart is doing it in a lot of locations.

I've heard they also believe if electric cars get going they can offer free recharge with people buying xx dollars of groceries.

[guntars]

Yep, I’ve seen their solar roofs covering the parking spots in Flagstaff, AZ. It’s a win, win - they get power and it keeps the cars cooler.

[anderspitman]

I saw a random magazine 10 years ago that claimed Flagstaff was one of the top 3 "greenest" cities in the US. Prescott (where I'm from) was top 3 skinniest.

[BenFrantzDale]

Wouldn’t one approach be for property managers to lease rooftop space too, making a place for businesses focused on leasing rooftop space, installing solar, and selling power?

[diogenescynic]

Owners of the malls would still hold the rights to the roof, not the occupants of stores, so you’d just have to pay them enough that it’s worthwhile for the owners. However, I like the idea of requiring it for new developments over X square footage.

[Tempest1981]

FWIW, California is requiring solar on new homes, in 2020:

https://fortune.com/2018/12/06/california-solar-panels-new-h...

Maybe not as cheap as large-scale solar, but cheaper than natural gas?

[qserasera]

If a moderate sized location goes out of business would solar panels have decent resale value?

[hutzlibu]

Well, the prizes for new panels are still falling. And the cost of moving them might be too high..

But even used solar panels will be in demand.

[namibj]

Unless you need to demolish, you want to keep them in-place.

[bartimus]

Why would we not want to move them elsewhere?

[pjc50]

The panels stay generating even if the shop below has closed. The building remains owned by someone who has a financial interest in keeping them running.

[riffraff]

They would make the property more valuable when you sell it.

[harry8]

Great post, thanks for the numbers link. I've never heard of land being referred to as a "soft cost" before!

I'm having trouble making sense of them in terms of the cost differential between something like (say) a coal fired electricity plant (and/or nuclear) including (say) 20 year running costs, vs a solar & battery power plant of equivalent scale including the cost of purchasing the land required to build each. There may be better comparison assumptions, maybe lives of plants are different for example but I think that would be a good place to start looking at a comparison. Is per watt DC the right measure? (Probably it is but I'd like to see the justification, given the use of high voltages at scale - again I'm not the expert to know.)

One would assume from there for a power generating utility company the cost of sales, marketing, billing, etc would all be the same regardless of how you generate it. Entirely separate to that one could estimate the additional goodwill available for solar over coal or nuclear in advertising "we're clean, we're green, just look at our stunningly beautiful field of photovoltaics." Which must have /some/ value.

I wonder if the sun-chasing required in choosing the location makes getting the power from the plant to the grid more expensive? Numbers would help.

I'm also having trouble making sense of the listed "soft cost" numbers for residential. Land is already purchased so cost is zero, use what you generate so sales also zero, zero tax, zero net profit and nobody values the overhead of their time for these things, for which assumptions are critical if you do (but definitely a nice saving on the power bill you pay with /after tax/ dollars) - These assumptions I just listed are obviously wrong in that chart but why??? No explanation for numbers that don't add up when you apply obvious assumptions about what you do know makes us all suspicious, right? They really should fix that so it's obviously right, clear, honest and direct in a field I think we can all agree is marred by a veritable mountain of BS and propaganda. I'm absolutely not saying this is BS, nor am I saying it's wrong or propaganda - just I don't understand the numbers on the chart and /can't/ from what is actually there. That sucks for me trying to understand it but maybe I'm not supposed to understand these numbers because they're designed for someone else who has the necessary context. Is providing that context too hard?

[pjc50]

To dig deeper, go to the PDF: https://www.nrel.gov/docs/fy19osti/72399.pdf

> Land is already purchased so cost is zero, use what you generate so sales also zero, zero tax, zero net profit and nobody values the overhead of their time for these things

Their basis: "We model a 6.2-kW residential rooftop system using 60-cell, multicrystalline, 17.2%-efficient modules from a Tier 1 supplier and a standard flush mount, pitched-roof racking system"

So what they've done is effectively get quotes for installing their sample system and then divide by its power output. Sales tax, profit etc are added to the bill the customer pays the installer. Other tables give a more detailed breakdown.

[PuffinBlue]

that report puts a 76,800 square foot foundation (so perhaps triple that for total land usage) at $250,000. Is that anywhere remotely in the realms of reality for an area that might actually get permitted and connected at the other assumptions they've made?

It wouldn't be in my country (UK).

[Reason077]

UK solar farms generally don’t own the land they’re on, but pay a lease or license fee to the landowners.

A UK comparison would be the proposed Cleve Hill Solar Park in Kent, which will also have on-site battery storage. At 350MW it is almost as large as this LA project.

[cr0sh]

Not sure if it's still the largest, but this plant is a few hours drive away from where I live in Phoenix:

https://www.forbes.com/sites/uciliawang/2014/04/29/behold-th...

That's 2400 acres and it's up and running - so I assume permitted and connected to the grid. But I don't know the the total cost to compare things - I'd assume rather cheap, because Yuma is...not a "destination" (outside of maybe a stopover for the dunes).

[nachteilig]

Hey, could you please point me at some stats relating to the costs being lower, especially in Germany? Very interesting topic for me and I couldn’t find it in the study. Thanks!