Hacker News Reader

Comment by pfdietz | Hacker News Reader

What this shows is solar is increasingly threatening the electric utility business model. Even without net metering, demand destruction will cause the traditional model to stop working.

View on HN

Will it? I’m not sure how the utilities structure their prices wrt the actual cost, but they definitely separate the baseline connection cost from usage on bills (at least in the US), so they may not be killed by people using very little power as long as the connection fee actually covers things.

The hardest possible demand to meet is random, reasonal, and spikey demand spread diffusely over a large area. Which is more or less homes.

Conversely the easiest possible demand to meet is localized constant and high demand. Basically AI datacenters or industrial users. These guys are basically paying for the grid and residential have it as a subsidy.

The supermajority of the price of electricity is fixed costs related to installing and maintaining capacity. The marginal problem of increasing generation or utilization is cheap. I believe it's like under 20% even for gas power where you have to buy gas. For grid solar it would be even crazier because marginally its basically free they really don't care how much you use it even goes negative but the fixed costs are everything.

So what causes a lot of social problems is when wealthy people get their own private solar because the whole current pricing structure revolves around wealthy people using a lot of electricity and paying down the connection costs for poor people. If they have solar the poor people are fronting the maintainence cost which destabilizes everything.

Without load balancing, AI datacenters are very swingy (which is bad).

Unfortunately the connection fee does not cover all fixed cost. For a long time the model has been fairly "progressive" in this regard. Some of the fixed costs of the grid have been paid for by amortization over the per Kw cost, which had the effect of charging people who used more a larger chunk of these fixed costs. Now with the option to provide your own power if you have upfront capital for solar can build as big of a system as they want. As other comments in the thread have mentioned, net-metering is largely functioned as a subsidy to give money to people who are already doing fine financially. I want green energy, and I think that decentralization has definite benefits, but it's pretty hard to argue against maintaining the grid to allow re-balancing and covering supply shortfalls in specific areas. Here is a video discussing this problem - https://youtu.be/C4cNnVK412U?si=ZzZhoApFW3khqrdq&t=720

What you could do is bill per energy in e.g. 15 minute chunks, and separately bill for transformer/line capacity by e.g. the peak usage in any such chunk over the contract period, like they do in Germany for atypical load profile industrial users since decades ago.

Net metering is overall just entirely stupid as a concept; measure inbound and outbound flow separately if you can't just measure the 15 minute chunks; bill grid fees on the energy price on inbound and only pay energy price on outbound. Or even bill grid fees on outbound up to one of many available large substations, and thus handle the issue of demand across large distances making buildout of solar in a convenient but far away place not being disincentivized vs. more-demand-local buildout.

if the same solar also had enough battery capacity, sure. But they do not, they still need to buy at out of solar peak and that just causes problems for both sides.

I think grid should start moving into selling storage as a service. Just put a bunch of bulk storage at every transformer station and buy solar from consumers at solar peak, sell them back say 80% of it (or whatever margin is required to pay for it) off peak.

That way utility no longer have to haul megawatts all the way from the power plant all the time, any peak can be hauled from the batteries and let the other types of power plant more time to spool up, and the grid is more resilient to outages (assuming you were lucky and battery bank local to you still had some charge

LFP chemistries are approaching ~$50/kWh, and CATL's sodium chemistry is supposed to be ~$40/kWh (per CATL); soon it will be more expensive to ship the battery storage than the storage itself.

"Watershed moment:" Big battery storage prices hit record low in China auction - https://news.ycombinator.com/item?id=44504630 - July 2025 (4 comments)

IEA: The battery industry has entered a new phase - https://www.iea.org/commentaries/the-battery-industry-has-en... - March 5th, 2025

Naxtra Battery Breakthrough & Dual-Power Architecture: CATL Pioneers the Multi-Power Era - https://www.catl.com/en/news/6401.html - April 21st, 2025

China Already Makes as Many Batteries as the Entire World Wants - https://about.bnef.com/insights/clean-transport/china-alread... - April 19th, 2024

Currently "Sodium Battery Cell" is priced at US$50–70/kWh on the Shanghai Metals Market, which gives every appearance of being a real thing: https://www.metal.com/en/markets/43. They list LFP batteries currently at US$38–52/kWh: https://www.metal.com/en/markets/42

I think you may be dramatically overestimating how much container shipping costs.

("Shanghai Metals Market" is reportedly a consultancy which provides price data, not a market.)

Sodium batteries have significantly lower round-trip efficiency, so they have to be cheaper to win against LFP.

That's because of the huge range of voltages in the charge curve - resistive losses in wiring scale with the square of current. When the voltage drops, the current needs to increase for the same power.

If you can accept low power/current output (which EVs cant, but homes can), this isn't so bad.

EVs have giant batteries - they can be connected via their DC ports and charged/discharged via solar inverters - technically. The current spec for CCS and NACS doesn't allow for this (Chademo did, but they lost the 'format war'). Giant effing oversight on manufacturers' part if you ask me.

Some people have managed to trick their cars into reverse charging via solar hybrid inverters and some custom hardware and it works as advertised - which is no surprise since its a lithium battery charge controller charging/discharing a solar battery.

If you could use your 60kWh EV battery on top of the 10-20 kWh you have at home, it would be a game changer, most people could power their homes for a week on that sort of capacity.

Replacing all the road vehicles in the US with 70 kWh BEVs would mean their battery capacity would equal about 40 hours of the average US grid power consumption.

BTW, this will mean that EV charging is going to have to have variable rates, or else people will just ride over Dunkleflauten by charging up their EV at a charger, driving back, then using it to power the home.

I think variable consumption is a thing even now - we have peaker plants just for that purpose, so that could be handled as well.

In most places in the developed world utility-scale solar is much cheaper to build than rooftop solar. And there's value in having a stable grid to fall back on. I think the demand destruction story is overrated.

Balcony solar is even cheaper to build than utility-scale solar.

It's limited to 800W in Germany and 1200W in Utah. It isn't about to replace utility-scale generation anytime soon.

Those limits are just arbitrary regulations. It's easy to install 10x that on a residential roof.

And we come back to my original point. Residential roof installation is the most expensive way to install solar power. Utility scale solar is easily haf the cost of residential.

> Utility scale solar is easily half the cost of residential.

If it's only half, the problem it's not going to stop residential installs. By the time that utility power gets to them here in Australia it costs about 3 times as much, so they are going to install their rooftop systems anyway.

I can't speak for elsewhere, but here in Australia residential installs tend to be over provisioned. A small'ish install is 5kW. That generates about 20kWh per day. Typical household consumption is 1/2 that. Newer builds like mine tend to have far more - upwards of 20kW of panels. That's to cater for charging EV's. The result is grid solar installs are getting hammered by roof top solar: https://reneweconomy.com.au/wind-and-solar-hit-record-share-...

In the US it's typically about a quarter.

It shouldn't be that expensive. I had 10x 500w installed and wired on my 7-8 meter high roof in 1/2 day.

I only paid $50 where I ljve, but even with 10x higher labor costs in the US, it should be under $1000.

It shouldn't be, I agree. And eventually it won't be. But SEIA's report https://www.seia.org/research-resources/solar-market-insight... said that, in Q1 of 02024, utility-scale fixed-tilt installations in the USA averaged 98¢ per peak watt, of which 40¢ was the PV modules; and residential installations averaged 325¢ per peak watt, of which 20¢ was the PV modules, which is an atypically low share, historically speaking.

Where do you live?

I'm surprised at the 40¢ figure for utility scale. Are you sure that didn't include the mounting hardware? Utility-scale PV is typically with 1-axis trackers.

Another possibility would be that utility scale solar optimizes to a much large disparity between DC rating and AC rating than does domestic rooftop PV.

I was surprised at it too, but remember that the US basically doesn't allow Chinese solar panels to be imported.

Wait, so those were not US residential cost figures you quoted? Because I don't see how they'd allow Chinese imports for only residential installation.

Those were also US, yes. I don't know why they were so much cheaper. Possibilities include:

- Maybe the SEIA is wrong.

- Maybe utility-scale projects sign contracts to buy PV modules further in advance. The prices were falling rapidly.

- Maybe residential installers were buying low-cost panels with lower efficiency or with no warranty.

Balcony solar is installed on a unit-by-unit basis. It isn't 800W per building; it's potentially up to 800W for each unit in an apartment.

That's still not a lot. Less than a hairdryer uses.

A heavily used hairdryer is typically in use for 5 minutes per day, so a 1500-watt-peak hairdryer might average 5.2 watts. It needs 1500 watts of batteries (say, 3kWh of 0.5C batteries, or 0.4kWh of 4C batteries), not 1500 watts of solar panels.

That only applies if you connect it to the grid in each of those places, I believe.

I can't wrap my head around what off-grid balcony solar even means. Balcony solar is meant for apartment dwellers. Or detached home/townhome owners who don't want to spend on a rooftop install. You aren't installing balcony solar to go fully off-grid.

Sure, but you don't need to go fully off-grid to have an off-grid balcony solar panel. You can connect it to just your freezer, or just a storage heater, or to the batteries you run your homelab off of.

Balcony solar can still satisfy a significant portion of a household's demand.

Utility-scale solar needs land and has to deal with line losses. Rooftop solar does not have to deal with either.

Utility scale is 2x cheaper despite all that.

> Utility scale is 2x cheaper despite all that.

I think you mean cheaper for the utility. It is certainly not cheaper for me, the homeowner.

The only reason for that is that you still have access to the grid and are gaming the system via a rate structure that allows you to free ride.

Not necessarily - you have to pay for that infrastructure between the plant and your home.

Commercial and industrial use already makes up a large portion of demand. While the model will change to cater less to residential needs, overall demand for stable, high voltage generation is not going to go down.

Commercial properties often have enough roof area to meet most of their daytime demand on-site. And industrial consumption in Western countries has been flat or declining for years, so "stable, high-voltage generation" may face less demand than assumed.

If long term storage (like Standard Thermal) comes into play, I could see industrial users decide to just decouple from the grid too (or, use the threat to do so as a means to drive down what they are charged.)

We may see industrial users preposition themselves in locations with ample nearby PV potential. If I were building a factory in the US (or a data center) I would think twice before putting it in a higher population density area.

We may also see local microgrids develop. This would still have distribution costs, but not transmission.

I think in most countrys, you already pay one bill for the grid and one for the used electric power.

The bill you pay for the grid in such cases typically doesn't cover the fixed costs, and would have to be much higher if solar demand destruction becomes big.

There’s a business model where distributed solar production and storage is the norm and central grid based generation and delivery is the minority.

Such a model is extremely resistant and there’s less system infrastructure necessary. It’s quite feasible to redesign the system around a “distributed first” model.

Where do the massive upgrades to the distribution system required for this kind of setup come from?

We simultaneously hate utilities and want them to redesign and pay for a distribution system that was not intended for bidirectional load flow.

Our municipal distribution systems are barely adequate. Net metering produces essentially no revenue but imposes a huge load on that infra.

My understanding is there is less of a need for massive grid upgrades in this model due to the use of storage. Rather than having to be able to distribute peak loads from solar, requiring a larger connection, you can smooth out the supply and distribute an even amount throughout the day, using a smaller connection.

The section "1.1.3 Bringing large savings on grid expansions" [1] has a good explanation.

1. https://ember-energy.org/latest-insights/solar-electricity-e...

This is exactly wrong because transmission and distribution are different things. You can think of distribution as the 'last mile' infrastructure. It's not expensive in the way transmission is - individual services are relatively cheap. The reason it's expensive is that there are enormous amounts of it and municipal distribution systems can be wildly complex. Power flows in these systems are unidirectional. When you throw that assumption on its head you drastically change how these systems must be designed.

A major change like that would be astronomically expensive. The only reason it kinda works now is that very few people do grid connected generation after the meter. Once that becomes more widespread it's time for a huge investment, on the order of what the transmission system needs, but it all has to be paid for by local utilities, very few of which can get the capital together to do major projects like that.

> solar is increasingly threatening the electric utility business model

The writing is on the wall that the electric utility business model is a dying business like the career of bus or truck driver. Some countries will take a while to realize due to head in the sand , tariffs and corruption.

What will people do at night?

Shift usage to daytime and rely on battery storage.

Batteries are very expensive.

Couple used car batteries or usb batteries are cheap.

Enough to keep the lights on, a fan or charging a phone. Not to run an AC or dishwasher but enough for the basics.

They're cheap and getting cheaper. Your info is from 2021. Things change quickly in this industry.

Batteries are absolutely not cheap for what they store.

How much do you think they cost?

Installed costs for residential battery storage typically range from $800-1,200/kWh in the US market as of 2024-2025.

ROI for 24/7 solar+battery is negative in almost all residential cases using current technology and prices.

Let's assume the lower end of that range - $800/kWh and take the example of Arizona. https://www.aps.com/en/Residential/Service-Plans/Compare-Ser...

The difference between off-peak and on-peak rates there is about $0.21/kWh in the summer and $0.29/kWh in the winter, so assume an average delta of $0.25/kWh.

Assuming no solar panels, if you charge the battery on the grid at off-peak rates and discharge it completely at on-peak rates, you break even after using about 3200kWh. Assuming 2 kWh used every day during peak rates that's a breakeven period of 4.38 years.

Maybe my calculations are wrong somewhere, and I'd love to learn if that's the case.

I also acknowledge some big assumptions: namely, that you will always need to use all your stored energy day at the peak time (reasonable in the summer when the AC is running) and that you can use all your battery power for those loads instead of grid power. On the other hand electricity rates tend to go up over time and batteries last for years.

3.5 cents per KWh is exceptionally cheap and does seem to give a reasonable payback period for cheap enough batteries.

No I goofed a bit too, by assuming 2kWh used every day but $800 in installation cost (which is for 1kWh). Fixing that error increases the payback to 8.76 years/kWh of installed battery. But only if you assume rates don't change for all of those ~9 years.

But rates will likely increase. So will the delta between off-peak and on-peak rates. I think the true payback period is between 5 and 8 years.

For my experience a lot of installations really doesn't have much battery capacity cause batteries are pretty expensive at least here in Nigeria, but a lot of people are really happy with the system as long as they get electricity even if it's only during the day.

overprovision for their needs during the day and utilize battery power at night.

Solar panels are cheap but batteries are very expensive.

They can be depending on your needs. Lithium iron phosphate batteries are pretty cheap for their capacity. If you build your own power station with them you'd be surprised how far your money goes.

It is much more expensive that just buying electricity from the grid at night.

In Australia, if you have the space for rooftop solar, it's far cheaper in the long run to buy solar+battery. We did the math for our household and even if grid prices are stable (which they aren't, they're fast increasing) we're still going to make money back on the investment in less than 4 years.

Granted this includes a government rebate for the battery, but overall the prices have plummeted. Any government that isn't pushing for renewables and energy storage at this point is actively working against it's citizens.

Does that take into account the free electricity?

No, and once the Solar Sharer scheme kicks in it'll be very helpful in avoiding leaning too hard on the grid in the evening after rainy or overcast days.

It's a fantastic way to solve oversupply; give it to everyone, including those who have batteries in areas where the weather restricts solar output.

https://www.abc.net.au/news/2025-11-03/energy-retailers-offe...

The free electricity seems ideal to charge batteries with.

It's worth the price if you live in an area where grid stability is not a given.

Batteries are dirt cheap already and getting cheaper all the time. Pakistan would be buying them at the Chinese prices without a lot of tariffs or nonsense that might be misleading you into believing otherwise.

Think a bandwidth of 50-80$ per kwh cost levels for the manufacturer with a margin on top in a market where there's over production and prices are still trending down and margins are probably under quite a bit of pressure. That's the widely publicized cost levels for Chinese manufacturers that dominate the world supply currently. Some of the sodium ion batteries that are coming to market now are already at the lower end of that price bandwidth and could go to 10-20$/kwh over the next 5-10 years; maybe faster.

At those prices, anyone can afford plenty of battery to survive the sun not shining for days/weeks. Which in places like Pakistan would be redundant. It's far south and you can count the number of days that you shouldn't be wearing sunglasses outside per year on the fingers of one hand. Even when it's cloudy, there's plenty of light filtering through in that part of the world..

Prices you might be seeing in the US tell you more about the local politics there than the economics of batteries. The US has it self to blame for bad economics like that. Places like Pakistan aren't going to slow down because the US can't figure out all this new stuff. For them this is economic growth unlocked by vastly more energy than they've ever had access to. All they'll ever need basically.

"Batteries are dirt cheap already" they absolutely are not.

Multiple people have now explained you are in error here. I expect you will not repeat this falsehood going forward.

Installed costs for residential battery storage typically range from $800-1,200/kWh in the US market as of 2024-2025.

ROI for 24/7 solar+battery is negative in almost all residential cases using current technology and prices.

Unless you are proposing the laws of physics are different in the US, this is just a matter of market friction and blockages from regulations and tariffs. You know, the sort of things nuclear bros dismiss with a wave of their hand?

Batteries are cheap in developing world prices, cheap in rich country prices. They continue to become cheaper every year.

EDIT: I meant they're on the expensive side for the developing world, but cheap for rich countries. If you aren't stupid enough to tariff them.

You can order 2kWh of plug-in-ready battery for 400€ or so on Amazon. That's single digit cents per kWh over the lifetime of the battery. Bigger systems are cheaper.

Battery modules (not home scale, but utility scale) are around $50/kWh in China. If we assume a 20 year lifespan and 50% charge/discharge once a day, that adds (ignoring interest) $0.013/kWh to the energy cycled through the batteries (plus a small add from efficiency being not quite 100%).

This is quite cheap compared to (say) the fully loaded cost of energy from a nuclear power plant.

Smaller units will be more expensive per kWh, but not so enormously so as to render them impractical. And they will get cheaper quickly like all electronics do.

Installed costs for residential battery storage typically range from $800-1,200/kWh in the US market as of 2024-2025.

ROI for 24/7 solar+battery is negative in almost all residential cases using current technology and prices.

For $800 you can get a complete 2kWh system, including cells, inverter and panels shipped to your home in Europe, so if you're paying more than that for half the storage either your government or contractor is fleecing you.

I can confirm that prismatic 1kWh LFP cells cost ~$60 in single digit numbers.

What is the US doing differently that it is so expensive there?

I think I can answer that, though I'm not a Pakistani but as a Nigerian in a developing country, you might also have a petrol generator for night times. But for the majority of people just having your phone and power bank charged for the night is pretty ok, a plus if you can keep a handful of bulbs on also.

At night, what people will do is wait until the morning to run the washing machine.

Most people aren’t interested in being responsible for their own electrical generation. Especially with payback still being on the order of decades

Most people don’t have the capital to be responsible for their own electricity generation, except the rich.

It's really not expensive anymore. There's a Black Friday deal on amazon.de for an entry-level Anker Solix system with 4x500W panels and a 2.6 kWh battery. 1200 EUR.

For those that don't have the cash, financing is available.

Have you read this, about SunKing and SunCulture in Africa, recently posted on HN: https://climatedrift.substack.com/p/why-solarpunk-is-already...

Their smallest solar products are small lanterns. Simply having a pollution-free source of light is already a quality of life improvement for some people. One step up is to add a USB port to charge phones.

Oh I’m sure, Pakistan has alot of trade with China also so it’s probably cheaper than in the west. But it would still be expensive for the poorer Pakistanis, and would require some investment, they simply have less ability to do that than a richer middle class Pakistani, so the poor pay the poor tax because they can’t invest capital to bring their costs down.

Low Chinese prices are making it more and more possible though. I hope the future will be really different.

We're seeing solar with batteries going in in sub-Saharan Africa, so don't tell me something can't be made to work even with the poor in Pakistan.

A lot of people in my area are interested and it would be a net positive for them long term, but the area is poor so few can afford the initial costs despite knowing the money they would ultimately be saving.

My payback is 6-7 years in Canada with $0 invested.

“Responsible for my own power generation” = I do literally nothing. Nada.

I get $1000 a year for free.

Please show me someone who does not want $1000 per year for absolutely nothing.

Visit Australia, plenty of people are! When the real paybacks are generally 4-8 years (depending on what we're talking about) why wouldn't people be? We have 4.2 million solar systems (for reference there are almost 11 million dwellings). Just this year the Federal Government started giving out grants for home batteries and over 55,000 people have already taken that up and at least 90,000 home battery installations exist so far according to these stats: https://cer.gov.au/markets/reports-and-data/small-scale-inst...

Even if people don't go to the lengths I do (I like to watch the current generation and will slightly delay my use big loads like the washing machine, dishwasher, dryer etc. to try and use as much as my solar as possible), it's still very common for people to choose to do things like set the dishwasher timer in the day to use solar - which is great because it's also taking load off the grid.

#1 is correct. #2 is quite incorrect.

Most of the costs for residential solar are installation. Systems that you install yourself, e.g. balcony solar, have payback times below five years (even in less than ideal regions, like Germany). I would assume that labor in Pakistan is a bit cheaper.

Reply on news.ycombinator.com